A po 1 1 o- M oon * • >6 " % Bookbinders ( 103)9836 1800 \ j Journal of the Royal Society of Western Australia CONTENTS Recent Advances in Science in Western Australia The role of diet in determining water, energy and salt intake in the thorny devil Moloch horridus (Lacertilia: Agamidae) P C Withers and C R Dickman New records and further description of Macrothrix hardingi Petkovski (Cladocera) from granite pools in Western Australia N N Smirnov and I A E Bayly Preliminary observations on termite diversity in native Banksia woodland and exotic pine Pinus pinaster plantations M Abensperg-Traun and D H Perry Membership of The Royal Society of Western Australia 1994-1995 and Index of Interests Page 1 13 15 19 Al Volume 78 Part 1 March 1995 28858 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Vice-Presidents Joint Hon Secretaries Membership Secretary Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1994-1995 D I Walker W A Cowling S Hopper M G K Jones L N Thomas V Hobbs P Gardner R Froend P C Withers J E O'Shea M A Triffitt B Dell J Dodd D K Glassford D Gordon K McNamara V Semeniuk BSc (Hons) DPhil BAgricSc (Hons) PhD BSc (Hons) PhD MA PhD BSc MSc PGDip EIA BSc (Hons) PhD BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BSc (Hons) PhD BA MSc PhD BA (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encour¬ aged to attend meetings on the third Monday of every month (March-December) at 8 pm, Kings Park Board offices Kines Park, West Perth, WA 6005. Individual membership subscriptions for the 1994/1995 financial year are $40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1994 calendar year. For membership forms contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions and societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design: Mangle's kangaroo paw (Anigozanthos manglesii ) and the numbat ( Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia MUSEUM OF VICTORIA CONTENTS Recent Advances in Science in Western Australia The value of macroinvertebrate assemblages for determining priorities in wetland rehabilitation: A case study from Lake Toolibin, Western Australia R G Doupe and P Horwitz An Upper Cretaceous chert nodule, apparently marine ballast, from Princess Royal Harbour, Western Australia. J E Glover, R J Davey and C E Dortch Freshwater biogenic tufa dams in Madang Province, Papua New Guinea. W F Humphreys, S M Awramik and M H P Jebb Page 29 33 39 43 Volume 78 Part 2 June 1995 30086 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia COUNCIL 1994-1995 President D I Walker BSc (Hons) DPhil Immediate Past President W A Cowling BAgricSc (Hons) PhD Vice-Presidents S Hopper BSc (Hons) PhD M G K Jones MA PhD Joint Hon Secretaries L N Thomas BSc MSc PGDip El A V Hobbs BSc (Hons) PhD Membership Secretary P Gardner BEng (Hons) GDipCSci DipEd Hon Treasurer R Froend BSc (Hons) PhD Hon Editor P C Withers BSc (Hons) PhD Hon Journal Manager J E O'Shea BSc (Hons) PhD Hon Librarian M A Triffitt BA ALIA Members B Dell BSc (Hons) PhD J Dodd BA MSc PhD D K Glassford BA (Hons) PhD D Gordon BSc (Hons) PhD K McNamara BSc (Hons) PhD V Semeniuk BSc (Hons) PhD The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encour¬ aged to attend meetings on the third Monday of every month (March-December) at 8 pm. Kings Park Board offices, Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1995/1996 financial year are S40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1995 calendar year. For membership forms, contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. Ihe journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions and societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. Ihe Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design. Mangle's kangaroo paw ( Anigozcmthos nianglesii) and the numbat (Myrmecobius fascia tys) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia CONTENTS Page ISSN 0035-922X Recent Advances in Science in Western Australia 55 An early Triassic fossil flora from Culvida Soak, Canning Basin, 57 Western Australia G ] Retallack New Pleistocene and Holocene stratigraphic units in the 67 Yalgorup Plain area, southern Swan Coastal Plain V Semeniuk Seagrass communities in Exmouth Gulf, Western Australia: 81 A preliminary survey L J McCook, D W Klumpp & A D McKinnon Volume 78 Part 3 September 1995 The Royal Society of Western Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Vice-President Hon Secretaries Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1995-1996 S Hopper D I Walker M G K Jones P Gardner V Hobbs P Lavery R Froend P C Withers J E O'Shea M A Trifitt W A Cowling J Dodd D K Glassford D Gordon K Rosman V Semeniuk L N Thomas BSc (Hons) PhD BSc (Hons) DPhil MA PhD BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BAgricSci (Hons) PhD BA Msc PhD BA (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc MSc PGDipEIA The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encour¬ aged to attend meetings on the third Monday of every month (March-December) at 8 pm, Kings Park Board offices, Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1995/1996 financial year are $40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1995 calendar year. For membership forms, contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions and societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design : Mangle's kangaroo paw (Anigozanthos manglesii ) and the numbat ( Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia 1995 Medal Recipient: Professor A R Main 89 The Royal Society of Western Australia Medal Recipients 90 The study of nature - A seamless tapestry: Royal Society Medallist's Lecture for 1995 A R Main 91 Diurnal stratification of Lake Jandabup, a coloured wetland on the Swan Coastal Plain, Western Australia D S Ryder & P Horwitz 99 Cocoon formation by the treefrog Litoria alboguttata (Amphibia: Hylidae): A 'waterproof' taxonomic tool? P C Withers & S J Richards 103 Foraging patterns and behaviours, body postures and movement speed for goannas, Varanus gouldii (Reptilia: Varanidae), in a semi-urban environment G G Thompson 107 Constitution and Rules and Regulations 115 Volume 78 Part 4 December 1995 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Vice-President Hon Secretaries Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1995-1996 S Hopper D I Walker M G K Jones P Gardner V Hobbs P Lavery R Froend P C Withers J E O'Shea M A Trifitt W A Cowling J Dodd D K Glassford D Gordon K Rosman V Semeniuk L N Thomas BSc (Hons) PhD BSc (Hons) DPhil MA PhD BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BAgricSci (Hons) PhD BA Msc PhD BA (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc MSc PGDipEIA The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March -December) at 8 pm, Kings Park Board offices, Kings Park West Perth, WA 6005. Individual membership subscriptions for the 1995/1996 financial year are $40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1995 calendar year. For membership forms, contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions and societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design: Mangle's kangaroo paw ( Anigozanthos manglesii ) and the numbat ( Mymiecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia, 78:1-2, 1995 Recent Advances in Science in Western Australia Earth Sciences Twelve genera of chlorophycean algae in Triassic pa- lynological assemblages from north-western Australia are reviewed by W Bremmer (GEOMAR, Kiel) and C Foster (AGSO, Canberra). They suggest a possible evo¬ lutionary pathway for certain chlorococcalean coenobial families based on comparison with some living green algae (Orders Chlorococcales and Zygnematales) which develop resistant outer organic walls and/or cysts dur¬ ing their life cycle. Brenner W & Foster C B 1994 Chlorophycean algae from the Triassic of Australia. Review of Paleobotany and Palynology 80:209-234. Glengarry Group stratigraphy identified in the fold belt to the north can be applied to the southern foreland area on the Glengarry 1:250 000 map sheet. Four of the five major tectonic events that affected the Nabberu Province can be recognised on the foreland: listric fault¬ ing (Phase 1) relates to early basin subsidence by crustal extension; the succeeding convergent-styled Phases 2-4 (4 has no expression on the foreland) probably relate to the collision of the Capricorn Orogin; and the final tec¬ tonic act of normal faulting (Phase 5) may relate to the onset of crustal extension controlling the development of the younger Bangemall basin. Gee R D & Grey K 1994 Proterozoic rocks on the Glengarry 1:250 000 sheet — stratigraphy, structure and stromatolite bio¬ stratigraphy. Geological Survey of Western Australia, Report 41. Boudinage is shown by D Findlay (Consultant, South Perth) to be an important aspect of the structure of the goldfield, the lode distribution to be coincident with the principal necks, and the configuration of the lodes to match the characteristic fracture patterns of classical boudin necks. Boudinage is therefore interpreted to be an important control on the emplacement of mineraliza¬ tion and is proposed as a simpler alternative to the more complex shear-related models, and may be useful in ex¬ ploration for deposits of similar type. Findlay D 1994 Boudinage control on the emplacement of lodes of the Kalgoorlie goldfield. Australian Journal of Earth Sciences 41:105-113. A U/Pb isotopic age determined by researchers from the Key Centre for Strategic Mineral Deposit Research (University of Western Australia) and the Research School of Earth Sciences (Australian National Univer¬ sity), of 2991 ± 12 Ma, for zircons from the Mt Brown Rhyolite, in the Whim Creek Supersequence, supports previous evidence that the main periods of silicic volca- nism in the west Pilbara are much younger than in the East Pilbara. Available data now favour the westwards growth of the Pilbara Craton from ca. 3.5 to 2.9 Ga, rather than the single essentially tabular stratigraphy for the craton originally proposed. Barley M E, McNaughton N J, Williams I S & Compston W 1994 Geological note. Age of Archaean volcanism and sulphide mineralization in the Whim Creek Belt, west Pilbara. Australian Journal of Earth Sciences 41:175-177. Studies of Archaean lode-gold deposits from Western Australia, by local scientists (Key Centre for Strategic Mineral Deposit Research, University of Western Aus¬ tralia) indicate that the Pb isotopic data distribution for ore-associated galenas and pyrites is commonly linear, and determination of the initial Pb ratio relies on select¬ ing the least-radiogenic composition. Detailed studies of Victory Mine samples have led to the development of criteria for selecting target pyrite samples which best preserve the initial Pb ratio of the sulphide, and hence the ore fluid. The most important controls on the dis¬ placement of the measured Pb ratio of pyrite from the initial Pb ratio are the Pb content of pyrite, pyrite abun¬ dance, proximity of pyrite to ore-fluid channelway, and host rock. Ho S G, McNaughton N J & Groves D I 1994 Criteria for determining initial lead isotopic compositions of pyrite in Ar¬ chaean lode-gold deposits: a case study at Victory, Kambalda, Western Australia. Chemical Geology 111:57-84. Life Sciences Collaborative research, by scientists from James Cook University (Qld), Department of Conservation and Land Management (WA) and the Conservation Com¬ mission (NT), has examined the abundance and distri¬ bution of the dugong in Shark Bay, Western Australia. Aerial survey counts (corrected for perception and avail¬ ability biases) indicated 10146 ± se 1665 individuals, at a density of 0.71 ± se 0.12 dugongs km'2, with a high per¬ centage of calves (19%). Shark Bay is confirmed to be an internationally significant dugong habitat. Marsh H, Prince R I T, Saalfield W K & Shepherd R 1994 The distribution and abundance of the dugong in Shark Bay, West¬ ern Australia. Wildlife Research 21:149-161. A comparison by two South African scientists (De¬ partment of Botany, University of Cape Town) of plant traits for edaphically-matched sites at the Barrens, south¬ western Australia, and the Agulhas Plain, southwestern South Africa, indicated strong convergence in a wide range of traits related to morphology and function. Ex¬ amples of non-convergence are attributed to regional and historical processes rather than differences in the contemporary physical environments of the two areas. Cowling R M & Witkowski E T F 1994 Convergence and non¬ convergence of plant traits in climatically and edaphically matched sites in Mediterranean Australia and South Africa. Australian Journal of Ecology 19:220-232. Bird censuses at 20 mulga sites across Australia en¬ abled North American ornithologist, M Cody of the University of California at Los Angeles, to use null (bi¬ nomial) models to show that community composition is far more consistent among censuses than expected at random. Of the total of 81 bird species censused, 32% were interpreted to be '"core species" in "core niches" that accounted for almost 75% of the bird density in mulga communities. Regional variations in the composi¬ tion of some guilds illustrated various factors that con¬ tributed to a large species total in mulga, despite the predominance of "core species" Cody M L 1994 Mulga bird communities. I Species composi¬ tion and predictability across Australia. Australian Journal of Ecology 19:206-219. 1 Journal of the Royal Society of Western Australia, 78(1), March 1995 A study of the parrotfish family (Scaridae) by D Bellwood uses comparative morphology to identify spe¬ cies groups within the family, and determine their phy¬ logenetic relationships. Examination of 143 character states for 69 of the 180 species of Scaridae are used for analysis by the principal of maximum parsimony. A di¬ agnosis of supraspecific taxa, a key to genera, and a list of Recent species are also provided. Bellwood D R 1994 A phylogenetic study of the parrotfishes family Scaridae (Pisces: Labroidei), with a revision of genera. Records of the Australian Museum. Supplement 20. Researchers from the CSIRO Tropical Ecosystems Re¬ search Centre (NT) and Harvard University (USA) used an experimental study of the influence of abundance, high activity, and aggressiveness of Australian meat ants to demonstrate that their interference with other ant spe¬ cies has important implications for the sizes and densi¬ ties of colonies of the other species, and ultimately the overall ant community structure. Andersen A N & Patel A D 1994 Meat ants as dominant mem¬ bers of Australian ant communities: an experimental test of their influence on the foraging success and forager abundance of other species. Oecologia 98:15-24. The structure and fecundity of Banksia menziesii var¬ ies between its mesic range (Swan Coastal Plain) where the tree form has a relatively low fecundity, and its xeric range (Eneabba Plain) where the shrub form has a higher fecundity. Plants on road verges had a larger crown and a higher fecundity than plants more than 50 metres from the verge. Lament B B, Whitten V A, Witkowski E T F, Rees R G & Enright N J 1994 Regional and local (road verge) effects on size and fecundity in Banksia menzesii. Australian Journal of Ecology 19:197-205. Study of algal zonation on the intertidal limestone platforms off Perth by R Scheibling, of Dalhousie Uni¬ versity, Halifax, indicates a characteristic pattern of dense macroalgal beds nearshore with a barren zone on the seaward edge. Manipulative experiments indicated that the extent of algal cover was inversely related to the numbers of grazing molluscs, and that limpets and chi¬ tons accounted for 55 to 89% of the variation in algal cover whereas abalone accounted for <10%. The effect of the grazers was both through decreased space for colo¬ nization by algae due to the home scar area of the mol¬ luscs, and by grazing around the home scar area. Scheibling R E 1994 Molluscan grazing and macroalgal zona¬ tion on a rocky intertidal platform at Perth, Western Australia. Australian Journal of Ecology 19:141-149. Note from the Hon Editor: This column helps to link the various disciplines and inform others of the broad spectrum of achievements of WA scientists (or others writing about WA). Contributions to "Recent Advances in Science in Western Australia" are welcome, and may include papers that have caught your attention or that you believe may interest other scientists in Western Aus¬ tralia and abroad. They are usually papers in refereed journals, or books, chapters and reviews. Abstracts from conference proceedings will not be accepted. Please sub¬ mit either a reprint of the paper, or a short (2-3 sen¬ tences) summary of a recent paper together with a copy of the authors' names and addresses, to the Hon Editor or a member of the Publications Committee: Dr S D Hopper (Life Sciences), Dr A E Cockbain (Earth Sci¬ ences), and Assoc Prof G Hefter (Physical Sciences). Fi¬ nal choice of articles is at the discretion of the Hon Edi¬ tor. "Letters to the Editor" concerning scientific issues of relevance to this journal are also published, at the dis¬ cretion of the Hon Editor. Please submit a word process¬ ing disk with letters, and suggest potential reviewers or respondents to your letter. P C Withers , Hon Editor , Jour¬ nal of the Royal Society of Western Australia. 2 Journal of the Royal Society of Western Australia, 78:3-11, 1995 The role of diet in determining water, energy and salt intake in the thorny devil Moloch horridus (Lacertilia: Agamidae) P C Withers1 & C R Dickman2 department of Zoology, University of Western Australia, Nedlands WA 6907 2School of Biological Sciences, University of Sydney, Sydney NSW 2006 Manuscript received June 1994; accepted October 1994. Abstract We examine the nutritional and digestive constraints of obligate myrmecophagy to the Austra¬ lian agamid lizard, the thorny devil Moloch horridus. Observations of thorny devils feeding in their natural habitat, and examination of faeces collected from their habitat, confirm their essen¬ tially myrmecophagous diet. We identify two common types of ants that are eaten by thorny devils, Iridomyrmex sp from terrestrial trails, and Crematogaster sp from trails along the stems of currant bush (Leptomeria preissiana). Ants are consumed at a variable rate, from about 1 min'1 to 12 min1. The water, energy and solute content of ants is similar to other insects; the Iridomyrmex sp and Crematogaster sp were 62% water, and contained on a dry-mass basis 28-29 k] g1, 130-180 pmol Na+ g'1, and 220-240 pmol JO g1. The movements of thorny devils encompassed a variable area; lizards often remained within a restricted area over short periods, and did not exhibit any apparent signs of territoriality. Some lizards remained in restricted areas over considerable peri¬ ods (6 months to >1 year) whereas others dispersed widely. Thorny devils were observed to be active, and feeding, over a w’ide range of body temperatures, and their foraging did not appear to be related to, or restricted by, any particular body temperature. The digestive assimilation of thorny devils maintained in the laboratory was only 37% for dry matter, and 59% for metabolisable energy; the expected metabolisable energy efficiency is about 70% for a generalised insect diet, but would be expected to be lower for heavily sclerotised insects, such as ants. Our analyses of the nutritional value of ants, the measured metabolisable energy efficiency, and the assumption that thorny devils consume about 750 ants per day, indicate a field water turnover rate of 0.3 ml d and a field metabolic rate of 2.7 kj d'1. This calculated water turnover rate is considerably lower than expected for a 35 g lizard, but is similar to that measured by radio¬ isotopic turnover for thorny devils in the field (Withers & Bradshaw 1995). The calculated field metabolic rate is slightly less than that predicted for a 35 g lizard, and is similar to that measured for thorny devils in the field (Withers & Bradshaw 1995). We conclude that naturally-feeding thorny devils probably consume about 750 ants per day. Introduction The thorny devil, or mountain devil, (Moloch horridus) is an extremely cryptic, slow-moving agamid lizard that is wide-spread throughout much of semi-arid and arid Australia (Greer 1989; Cogger 1992). It is a sit-and-wait predator of small ants, primarily Iridomyrmex spp (Saville-Kent 1897; Davey 1923, 1944; White 1947; Sporn 1955; Paton 1965; Pianka & Pianka 1970). However, little is known of the natural history, ecology or physiology of the thorny devil (Bentley & Blumer 1962; Pianka & Pianka 1970; Gans et al. 1982; Sherbrooke, 1993; Withers 1993; Heatwole & Pianka 1993; Whitten 1993), in part because it is very cryptic and difficult to study in the field. Some other Australian lizards, particularly dragons, opportunistically consume ants but none are so exclu¬ sively ant-consumers as is the thorny devil (Pianka & Pianka 1970; Pianka 1986). Many consume termites © Royal Society of Western Australia 1995 opportunistically, and some geckos (Diplodactylus conspicillatus and Rhynchoedura ornata) and skinks (Ctenotus spp) are termite specialists (Pianka 1969). Food specialisation on social insects such as ants and termites may be advantageous because the prey are a clumped and concentrated food supply, but the nutritional and energetic consequences of a restricted diet such as ants or termites have been poorly addressed even for other taxa of ant-eaters such as mammals. For example, Redford & Dorea (1984) suggested that the low nutri¬ tional value of termites and especially of heavily sclerotised ants is a disadvantage (although alates have an exceptionally high nutritional value). The nitrogen levels of ants and termites are not particularly unusual, compared to other invertebrates, but Phelps et al (1975) reported a low digestibility by rats for termite protein. We are unaware of any studies which have determined for myrmecophages the digestive assimilation efficiency for either dry matter or energy. In this study, we document the composition of the diet for thorny devils at a study site located north of 3 Journal of the Royal Society of Western Australia, 78(1), March 1995 Southern Cross, Western Australia. We measure the wa¬ ter, energy, sodium and potassium contents for the ant fauna at this study site, and we determine the digestive efficiency of thorny devils for lridomyrmex sp in the labo¬ ratory. In addition, we examined thermoregulation and movements of thorny devils. These studies enable us to evaluate possible hygric, energetic, ionic, thermoregula¬ tory and locomotory constraints of myrmecophagy for free-living thorny devils, and to estimate values for, and ratios of, water, energy and solute turnovers. Methods Field Site Thorny devils were observed and studied in the field, in a sandplain habitat (latitude 30° 17'S, longitude 119° 44 'E; Fig 1) located 10-20 km north of Bungalbin, West¬ ern Australia (near fauna site 39 of Dell et al. 1985). The vegetation was spinifex ( Triodia scariosa and Plechtrachne sp) mixed with Eucalyptus leptopoda mallee and Banksia elderana, Acacia coolgardiensis and Callitris preissii tall shrubland (Dell et al 1985; Beard 1990). Currant bush (Leptomeria preissiana ; Fig 2A) was also present in the north-easterly parts of the study site. The study area is sub-desert, with warm winters (July, average monthly maximum 16-1 7°C) and hot summers (January, average monthly maximum 34-36°C); average annual rainfall is about 265 mm, with January- August being the wettest months (Dell et al. 1985). Thorny devils were routinely captured by pit-trap- ping, using 0.5 m long PVC pipe pit traps (160 mm dia) and aluminium flywire fences. Individuals were also lo¬ cated opportunistically by walking through the study area and examining the base of currant bushes where thorny devils were often seen feeding on arboreal ants. Their faeces were collected opportunistically in the field; the characteristic size and appearance, as well as their Figure 1. Location of field study site. almost exclusive ant content, readily identified Moloch faeces. Faeces were often located in aggregations of 5 to 20 individual pellets. Observations of Moloch feeding undisturbed at the study site enabled identification of their common ant prey as the numerous small terrestrial lridomyrmex sp, as well as an ant ( Crematogaster sp) common on currant bush. Individuals of these ant species, and numerous other species from the study site, nearby locations, and Figure 2. A (left). Typical habitat of thorny devils in the sandplain, showing a currant bush ( Leptomeria preissiana). B (right): Characteristic feeding posture of a thorny devil consuming ants from a trail on the trunk of a currant bush . 4 Journal of the Royal Society of Western Australia, 78(1), March 1995 Perth, were sampled for identification and determina¬ tion of water, energy and solute content. Ants were col¬ lected and placed in sealed plastic vials, and frozen for return to the laboratory for analysis. Movements of thorny devils The daily movements of thorny devils were moni¬ tored by thread spool and radiotelemetry, and longer term movements were monitored by recapture of marked (toe-clipped) individuals at the study site. A centre-unwinding nylon thread spool (Penguin Thread Company; 2 grams; « 270 m total length) was taped to the dorsal surface of the tail of some thorny devils. The centre-unwinding, free end of the thread was tied to a marker stake, and the lizard was able to walk unhindered by the thread trail. The length of thread that was unwound by the lizards movements was deter¬ mined by following the thread and rewinding it to the current location of the lizard, then measuring the length of retrieved thread in the laboratory. This technique de¬ termines the actual daily distance moved by the lizard. Small radiotelemeters (Biotrack SS-1 one-stage; ** 1 g) were taped to the dorsal surface of the tail of some thorny devils. These lizards were generally relocated daily using a Biotelemetry Systems radioreceiver and hand-held antenna. The actual location of the lizard was marked with a stake, and the distance and direction to the previous location were measured. Daily relocation allowed the determination of the daily point-to-point movement of lizards, and the average daily point-to- point movement over a period of 7 to 14 days. Thorny devils were marked on the abdomen using a felt-tip pen for a short-term identification, and many were marked permanently by toe-clipping. Composition of ants Ants were weighed individually, or in groups of 5 to 10 for smaller individuals, to ± 0.0001 g, then oven-dried at 60°C to determine water content. The dried bodies were then analysed either for energy or solute content. The energy content of 5 to 20 mg of dried ant bodies was determined using a microbomb calorimeter (Phillipson 1964) calibrated using oven-dried benzoic acid. It was not possible to determine the ash content of samples, so all energy values for ants are expressed as kj per gram total dry mass. The sodium and potassium contents of ant bodies were determined by digesting the oven-dried bodies in 0.5 to 1.0 ml of 10N nitric acid, then measuring the sodium and potassium ion contents of the digested supernatant with a Varian atomic absorption spectro¬ photometer, using caesium chloride as an internal standard. Feeding trials Thorny devils were returned to the laboratory in Perth, and fed daily with locally-collected lridomyrmex ants. The lizards were housed in the laboratory in glass aquaria, with a dry sandy substrate, and a heat lamp for thermoregulation during the day. It proved impossible to feed the lizards with known numbers of weights or live ants, and so lizards were weighed to ± 0.0001 g then placed in an aquarium containing freshly-collected ants, and allowed to feed for approximately 30 min, when they were removed and reweighed to determine the mass of ants consumed. The thorny devils were main¬ tained in the laboratory for 5 to 20 days in this fashion. Faeces were collected daily and oven-dried and stored for weighing and analysis of energy content. The dry matter and energy content of the ants was determined as described above. The energy content of 0.5 to 1.5 gram samples of faeces was determined using a Gallenkamp bomb calorimeter, calibrated using ben¬ zoic acid. The ash and sand content of the faeces was determined by weighing the sample crucible before and after bombing. Energy values for faeces are presented as kj per total dry weight (to calculate the total energy con¬ tent of collected faeces) and kj per ash-free dry weight (for comparison with the energy content of ants). The uric acid components of the faeces were included in en¬ ergy determination; the "faecal" mass and energy loss therefore included material/energy that was absorbed from the digestive tract but was subsequently excreted {e.g. urine), material /energy that was not absorbed from the digestive tract but eliminated as faeces, and mate¬ rial/energy added to the digestive tract e.g. secretions and sloughing of the gut lining. This was done because it was considered important to determine the overall material and energy balance of the diet, for comparison with a study of field metabolic rate of thorny devils (Withers & Bradshaw, 1995). Body Temperature When thorny devils were located in the field, their body temperature (Tb;°C) was measured immediately before disturbance whenever possible, using a Schultheis cloacal thermometer; air temperature (Ta; 1 m above ground in the shade) and substrate temperature (T ) at the site where the thorny devil was first observed were also measured. Thorny devils were routinely maintained in the labo¬ ratory in a large circular tank (dia = 3 m), with access to a heat lamp. Body temperature was determined for dev¬ ils at various times throughout the day, using a Schultheis cloacal thermometer. Results Diet of thorny devils Observation of thorny devils, feeding naturally in the field, indicated that lizards typically positioned them¬ selves over or just to the side of an ant trail, either on a sandy area or at the base of a currant bush (Fig 2B), and repetitively captured passing ants using their tongue. The ants that were observed to be eaten were invariably the common, small species; ants taken from the ground were lridomyrmex sp, and ants from the currant bush were Crematogaster sp. The lridomyrmex were readily dis¬ tinguishable from Crematogaster, which had a very spi¬ nous trunk, with two distinctive posteriorly-directed spines on the propodeum, and a "heart"-shaped abdo¬ men. Feeding rates observed for thorny devils in the field varied markedly from <1 to >10 ants min 1 (individual values were; 0.6, 2, 2, 2, 3.3, 7.7, and 6-12 min*1); mean 2.9 min*1. Over short periods, however, rates up to 1 sec*1 were recorded; this is the maximum possible rate 5 Journal of the Royal Society of Western Australia, 78(1), March 1995 because it took about 1 sec for a thorny devil to locate, capture and swallow an ant. The feeding rate appeared to vary with the abundance of ants and ambient tem¬ perature. Captive thorny devils were observed to feed at ant trails of Iridomyrmex spp in Perth at a rate of 0.5 to 5.8 min'1 (0.5, 1.3, 1.4, 2.6, 2.9, 3.1, 3.7, 4.3, 4.5, 4.7, 5.8 min"1); mean = 2.7 min1. The faeces of Moloch horridus are quite distinctive, be¬ ing large, ovoid and glossy in appearance; a urate pellet is frequently attached to one end of the faecal pellet (Fig 3A). The presence of essentially 100% ants in the diet is readily confirmed by visual inspection of crushed faecal pellets. Microscopic examination of faeces collected from the field confirmed the presence in the diet of small Iridomyrmex in localities without currant bush, and of mainly Crematogaster in localities writh currant bush (Fig 3B). In addition, pieces of charcoal (Fig 3C) and occa¬ sionally small pieces of vegetation (Fig 3D) were fre¬ quently found in the faeces. Presumably the latter items were either ingested through mistaken identity with small black ants (i.e. charcoal pieces) or perhaps as items being carried by ants (i.e. plant material). In addition to the common Iridomyrmex sp and Crematogaster sp consumed by thorny devils, a number of other ant species were collected from the ground and bushes at the study site, and other sites in the wheatbelt and in Perth. Ant composition The body mass (fresh) of ants varied considerably (Table 1), from less than 0.5 mg per individual, to over 100 mg per individual. The commonly-consumed Iridomyrmex (species 2) were about 0.45 mg, and the Crematogaster were slightly larger at 1.2 mg per indi¬ vidual. The water contents of the various ants ranged from less than 50% to over 80%. There were highly significant differences in water content between the various species of ants (for species with n>2 ; ANOVA; F1316g = 39; P<0.0005). The commonly-consumed Iridomyrmex 2 and Crematogaster sp (62% water) had an intermediate water content (Table 1). The energy content of the ants, ex¬ pressed per g total w'eight, ranged from less than 20 to more than 30 kj g1; the Iridomyrmex 2 and Crematogaster were intermediate, at 28 and 29 kj g"1 respectively. There were highly significant differences in energy contents of the various ant species (n>2; ANOVA; Fl357 = 4.8; P<0.0005). The sodium and potassium contents of the various ant species were quite variable, ranging from less than 100 to over 300 pmol g dry mass1, with highly significant differences observed between species for both sodium (n>2; ANOVA; F]041 = 4.8; P<0.0005) and potas¬ sium (n>2; ANOVA; F]03y = 8.5; P<0.0005). The com¬ monly-consumed Iridomyrmex 2 and Crematogaster sp had typical sodium contents (179 and 126 pmol g dry Table 1 Mean body mass and water, energy, sodium and potassium contents for various species of ants. The species observed to be eaten by thorny devils are indicated by an asterisk, and values are in bold face. Values are mean ± standard error, with the number of observa¬ tions in parentheses. Species Mass ^^Swet mass Water Content % Energy Content kJ 8d,ymJ Sodium Content Potassium Content umol g . ‘l not identified 10.6 (2) 72.2 (2) 27.7 (1) 219 (1) 268 (1) not identified 26.5 (2) 71.0 (2) 24.6 (1) 211 (4) 231 (1) Aphaenogaster 1 1.5 ±0.1 (4) 72.9 ± 0.6 (4) 32.0 (2) 163 (1) 195 (1) Camponotus sp 8.7 (1) 66.7 (1) Camponotus sp 14.2 (2) 67.8 (2) 18.6 (1) 233 (1) 181 (1) Camponotus sp 4.0 ± 0.2 (5) 68.0 ± 2.0 (5) 22.5 (2) 131 (2) 174 (2) Camponotus 1 6.9 ± 0.3 (10) 65.9 ± 0.9 (10) 22.7 ± 1.9 (5) 163 ± 17 (5) 255 ± 15 (5) Camponotus 2 5.35 ± 0.4 (10) 60.0 ± 0.9 (10) 24.5 ± 2.3 (7) ± 97 (2) 171 (2) Camponotus 3 19.3 ± 0.8 (3) 74.1 ± 1.6 (3) 18.1 (1) 328 (1) 344 (1) Crematogaster sp* 1.2 ± 0.1 (7) 62.0 ± 2.5 (7) 29.2 ± 1.2 (4) 126 (2) 216 (2) Iridomyrmex sp 2.45 (2) 68.5 (2) 39.2 (2) Iridomyrmex sp 1.8 ± 0.2 (5) 53.8 ± 1.7 (5) 17.8 ± 0.9 (3) 127 (2) 149 (2) Iridomyrmex 1 0.71 ± 0.03 (7) 63.2 ±1.4 (7) 26.7 ± 1.3 (3) 213 ± 31 (3) 240 ± 36 (3) Iridomyrmex 2* 0.45 ± 0.03 (13) 62.1 ± 1.4 (13) 28.1 ± 3.4 (4) 179 ± 41 (5) 241 ± 32 (5) Iridomyrmex 3 0.38 (2) 68.4 (2) 22.5 (1) 237 (1) 275 (1) Iridomyrmex 4 6.43 ± 0.3 (6) 70.5 ± 0.8 (6) 31.0 ± 1.0 (3) 252 ± 29 (3) 253 ± 13 (3) Iridomyrmex 5 0.9 ± 0.04 (14) 65.5 ± 1.0 (12) 30.8 ± 0.7 (6) 126 ± 12 (4) 160 ± 7 (4) Iridomyrmex agilis 0.77 ± 0.04 (3) 41 ± 3.2 (3) 31.5 (1) 56 (1) 40 (1) Iridomyrmex purpureus 10.5 ± 0.1 (20) 70.1 ± 0.5 (20) 22.9 ± 0.9 (7) 243 ± 39 (6) 227 ± 16 (6) Melophorus sp 29.1 ± 1.4 (4) 63.6 ± 2.9 (4) 22.4 (1) 101 (1) 162 (1) Meranoplus sp 1.3 ±0.1 (4) 64.4 (1) 27.1 (1) 34 (1) 43 (1) Pheidole sp 0.27 (1) 51.9 (1) 21.5 (1) Polyrachis sp 12.2 ± 0.8 (20) 70.1 ± 0.5 (19) 26.0 ± 0.8 (7) 120 ± 10 (6) 162 ± 15 (4) Rhytidoponera 14.9 (1) 58.4 (1) 164 (1) 167 (1) Rhylidoponera 1 29.1 ±2.1 (15) 58.5 ± 0.4 (15) 23.4 ± 0.7 (7) 115 ± 10 (6) 133 ± 13 (6) Rhytidoponera 2 6.4 ± 0.6 (17) 53.4 ± 0.6 (17) 28.7 ± 1.0 (7) 115 ± 12 (5) 149 ± 9 (5) Rhytidoponera 3 12.8 ± 0.5 (7) 54.9 ± 1.2 (7) 25.8 ± 0.7 (4) 108 ± 19 (3) 152 ± 9 (3) Rhytidoponera 4 36.6 ± 1.5 (20) 57.8 ± 0.9 (20) 25.3 ± 1.8 (4) 115 ± 13 (6) 127 ± 13 (6) 6 Journal of the Royal Society of Western Australia, 78(1), March 1995 mass'1 respectively) and potassium contents (241 and 216 pmol g dry mass'1 respectively). Movements of thorny devils The pattern of movements by thorny devils was vari¬ able; some lizards moved infrequently and for relatively Figure 3. A: The glossy, ovoid faecal pellet of a thorny devil, showing the attached uric acid aggregate, (scale bar = 1 cm) B: Typical view of the contents of the faecal pellet of a thorny devil, showing numerous pieces of ant exoskeleton and a charcoal granules, (scale bar = 5 mm) C, D: Parts of vegation from a faecal pellet, (scale bar = 1 mm) Figure 4. Pattern of movements of thorny devils. Axis tick marks are 20m intervals. Numbers next to location points (O) indicate the number of consecutive days that the lizard was located in that place (if >1). short distances (10-20 m day'1) whereas others moved more often and for greater distances (>100 m day1). A few individuals moved further than the length of a ny¬ lon spool (*» 270 m), or beyond radio-tracking range (*» 500 m) in a single day, and are necessarily excluded from numerical analysis here. The pattern of movement of individuals (erg. Fig 4) does not suggest any maintenance of exclusive home ranges, since the movements of many lizards overlapped those of others. Occasionally, two liz¬ ards were found together. The actual distance moved per day by thorny devils with spools was 77.9 m day'1 (± se 8.2; n=13). The direct distance moved per day by ra¬ dio-tracked thorny devils between daily relocation sites was 45.6 m point-to-point day'1 (± se 8.6; n=10). The net distance moved per day between the first and final relo¬ cation sites (determined over 7 to 14 days) was only 16.6 m day1 (± se 3.9; n=14). All of these values are signifi¬ cantly different (ANOVA, ¥2M = 20, P<0.001; Newman- KeuTs multiple comparison test). A number of thorny devils were recaptured between field trips, within about 100 m of the same location, after periods of 6 months (8 individuals), about 1.5 years (3), 2 years (1), and 3 or more years (3). One had moved 0.4 km after 4 months, five had moved about 1 km after 6 months to 1 year, two had moved about 2.5 km after 4 years, and one had moved about 8 km after 1.5 years. Digestion trials Ten thorny devils were studied in the laboratory (body mass = 28.6 ± se 3.7 g; range 15 to 52 g). The ants fed to thorny devils in the laboratory (Iridomyrmex 5) consisted of 65.5% water and contained 30.8 kj of energy per g total wet weight (Table 1). The faeces of thorny devils had an energy content of 7.80 ± 0.47 kj g'1 (n=24) expressed per total dry mass (including adherent sand particles) and 19.9 ± 0.62 kj g 1 (n=22) expressed per ash¬ free dry mass. The mean ash content of faeces was very high, 37.1 ± 1.5% (24), presumably because of their high sand content, both grains within and adherent to the outside of the faecal pellet. The thorny devils consumed ants at a rate of 0.16 ± 0.02 (n=10) g wet mass d'1 (or 0.057 ± 0.01 g dry mass d1), which is about 0.6% of their body mass per day. This is equivalent to a total energy intake of 1.8 kj day1. No thorny devils consumed sufficient ants per day to maintain a constant body mass; the observed weight change of -0.13 g d*1 (± se 0.02) was significantly differ¬ ent from 0 (ty = 6.5, PcO.001). However, this was signifi¬ cantly less (t,4 = 8.4, P<0.001) than the weight loss of non-feeding thorny devils of -0.30 g d'1 (± se 0.06, n=6). For five thorny devils, sufficient faeces were collected to enable a meaningful estimation of their steady-state energy assimilation. The total dry weight of faeces ex¬ creted over the duration of the feeding trials was 0.098 ± 0.022 g d 1 and the ash-free dry weight was 0.036 ± 0.008 g d The calculated metabolisable dry matter assimila¬ tion was 37%, but the food intake was not expressed as ash-free mass whereas the faecal energy was expressed as ash-free mass (to account for its high sand and other inorganic material content). The calculated metabolisable energy assimilation was 59%, which was considerably higher than the metabolisable dry matter assimilation. The metabolisable energy intake of the thorny devils was 7 Journal of the Royal Society of Western Australia, 78(1), March 1995 consequently 1.0 kj day1 (or 0.072 ml 02g_1 h1; assuming 20.1 Jslml 02). Body Temperature The mean body temperature of thorny devils able to thermoregulate behaviourally in the laboratory was 34.5 °C (± sd 2.43; ± se 0.48; n=26); the body temperature data were slightly negatively skewed, with a median value of 34.0 °C (Fig 5). The average for all field body temperatures recorded for thorny devils was 28.3 °C (± sd 6.6; ± se 0.7; n=93), but ranged widely from 14.5 °C to 38.7 °C. Mean air temperature was 25.6 (± sd 5.8; ± se 0.6; n=93), and ranged widely from 16.0 to 40.9 °C (Fig 5). Mean sub¬ strate temperature was 24.5 (± sd 6.5; ± se 0.8; n=93), and ranged widely from 14.0 to 41.7 °C. There was a significant linear regression relationship between all Tb and Tadata, and all Tband T. data; Tb = 4.1 (± se 1.8) + 0.95 (± se 0.07) Ta (n=92; r2 = 0.68; P<0.001) Tb = 5.9 (± se 1.7) + 0.88 (± se 0.07) T (n=67; r2 = 0.73; P<0.001) However, further analysis of the Tb and Ta data using two regression analysis to minimise the total residual sum of squares (Withers 1981; Yeager & Ultsch 1989; Fig 5) indicated separate relationships for Tb<32.5 °C and Tb>32.5 °C; for Tb<32.5 °C, Tb = -0.6 (± se 1.8) + 1.10 (± se 0.08) T (n=62; r2=0.77; P<0.001) and for T. >32.5 °C, D Tb = 29.5 (± se 1.7) + 0.20 (± se 0.06) T (n=27; r2=0.34; P<0.001) At the highest Ta's, three thorny devils were captured in or near burrows; their Tb was lower than the Ta (Fig Discussion Our observations of the diet of thorny devils are con¬ sistent with previous reports of their essentially being Air Temperature (°C) Figure 5. Relationship between body temperature and ambient air temperature for thorny devils in the field. Circles indicate data for surface active lizards; triangles are data for animals found in burrows. Closed circles and regression line are for liz¬ ards with Tb>32.5 °C (see text). entirely myrmecophagous (e.g. Saville-Kent 1897; Pianka & Pianka 1970). We have observed thorny devils to con¬ sume ants (Iridomyrmex sp) not only from terrestrial trails, but also Crematogaster sp from trails along the trunks of currant bushes, by standing next to or propped against the trunk with their forelegs (Fig 2B). Davey (1923, 1944) reported that thorny devils were selective for Iridomyrmex rufoniger over other Iridomyrmex, Ectatomma, Monomorium, Camponotus, Polyrhachis and Pheidole species. Sporn (1955) recorded that thorny dev¬ ils would eat six species of non-stinging ants. Pianka & Pianka (1970) reported the contents of thorny devil stomachs to consist mainly of small (2 to 5 mg) Iridomyrmex, and some sticks, stones, small flowers and insect eggs, presumably ingested coincidentally with ants or by mistake. The faeces of thorny devils are quite characteristic, being large, ovoid pellets often with an adhering uric acid aggregate (Pianka & Pianka 1970; Fig 3 A). The fae¬ ces are often found in groups of a few to over twenty pellets, varying from obviously fresh to much older ones. These aggregations of faeces in special "latrine" sites occur separately from basking sites (Pianka & Pianka 1970), and indicate either the activities of a single individual over a considerable time, the combined ac¬ tivities of a number of individuals, or both. The faeces of thorny devils from our study localities with currant bush contained numerous fragments of Crematogaster sp, whereas faeces from other nearby localities without cur¬ rant bush consisted of mainly fragments of Iridomyrmex sp. Thus, it appears that thorny devils will vary their feeding strategy and the species of ant that they feed on, depending on local conditions, but concentrate on rela¬ tively small and abundant ants which form trails. The other species of ants which we collected in the vicinity of the field site were not consumed, at least in signifi¬ cant numbers, presumably because of their unsuitable size, low abundance, or non-trailing behaviour. All previous reports of feeding rates have been for captive thorny devils. Saville-Kent (1897) reported that thorny devils consumed about 1000 to 1500 Iridomyrmex per meal, from ant trails. Davey (1923, 1944) reported that thorny devils consumed ants at a rate of about 45 min'1, in probably two meals per day lasting about 15 min each, which was sufficient to satiate the thorny dev¬ ils; consequently, the ingestion of ants was about 1350 individuals day*1. Similar rates of feeding were reported also by Duncan-Kemp (1933). In contrast, Sporn (1955) recorded thorny devils to consume 20 to 30 ants min1, with a meal lasting about 1 to \l/i hours i.e. about 1875 ants per meal. We have observed a wide range of feed¬ ing rates in the field, from <1 to >10 ants min*1. Pianka & Pianka (1970) observed the stomachs of thorny devils to contain about 1 to 2 cm3 (but up to 5 cm3) of ants; typi¬ cally there were about 2500 very small Iridomyrmex. It appears that the feeding rate, number of ants consumed per meal, length of meals, and number of meals per day might vary considerably, depending on factors such as ambient temperature, the size of the ants, and their abundance, and be higher for animals in captivity. There were considerable differences among various species of ants in their water, energy, sodium and potas¬ sium contents. The Iridomyrmex 2 and Crematogaster 8 Journal of the Royal Society of Western Australia, 78(1), March 1995 species which the thorny devils consumed were not un¬ usual in either water, energy, sodium or potassium con¬ tents, and we suggest that the species of ants which thorny devils consume are determined more by the size, abundance and behaviour of the ants than any nutri¬ tional aspect. Thorny devils move about 80 m day*1 (spool measure¬ ments), which is fairly similar to a distance of about 50 m day*1 determined by point-to-point daily; this is prob¬ ably because thorny devils appear to be fairly sedentary, and move in fairly straight lines rather than meander widely. Over a period of 7 to 14 days, the thorny devils appeared to remain generally in the same area, as the long-term average point-to-point movement was only about 20 m day'1 (see also Fig 4). Thus, thorny devils appear to rely on a food supply in a relatively small area, and their diet would reflect the local availability of ants ( lridomyrmex or Crematogaster spp). However, a few thorny devils were observed to move quickly over con¬ siderable distances (> 300 m d1), often essentially in a straight line; this contrasting movement pattern might reflect movements of individuals during the breeding season, or dispersing sub-adults or males. Movement records for longer periods are scarce, but the limited recapture data indicate that some individuals are ex¬ tremely sedentary, even over 6 months to 3 years, whereas others had moved a kilometre or more after 6 months to 4 years. Thorny devils were found to be active in the field over a wide range of body temperatures, as has been reported previously (Pianka & Pianka 1970; Pianka 1986). The mean of all Tb data recorded in this study was 28.3 °C (± sd 6.6; n=93), which is slightly lower than that reported by Pianka (1986) of 32.6 (± sd 4.1; n=190), at a slightly higher mean Ta of 27.1 °C (± sd 5.2). The lower Tb's and Ta's reported here probably reflect the consider¬ able data obtained for thorny devils located by spooling or radiotracking when inactive e.g. at night, and during cold and rainy conditions. The linear regression relation¬ ship of Tb = 4.6 + 0.93 Td was considerably different from that reported by Pianka (1986) of Tb = 18.1 + 0.54 Ta, and the regressions for Tb<32.5 and Tb>32.5 °C had a higher and lower slope, respectively, consistent with a non-linear relationship between Tb and Ta. Thorny devils are clearly not precise thermoregulators, and are not ap¬ parently restricted by Tb in being active, or feeding. In fact, one thorny devil was observed feeding in the rain, at a Tb and Ta of 16.5 °C! At extremely high Ta, thorny devils were seen to seek shelter in burrows, a behaviour not commonly observed otherwise. Energetics of myrmecophagy Thorny devils and other animals that specialise on eating ants (or termites) have a spatially clumped, but locally abundant prey. Their feeding strategy is usually an initial widely-foraging search and then a sit-and- wait feeding bout. This undoubtedly has marked impli¬ cations for aspects of their ecology e.g. anti-predator strategies including crypsis and defensive spines, slow movements, thermolability, extended activity period, and a stout body with a large stomach (Pianka & Pianka 1970; Pianka 1986). In addition, a highly specialised diet of ants (or termites) potentially has important digestive and nutritional implications, and determines water, energy and ion balance. The range of water content for the various ant species (40-80%; Table 1) is similar to that reported for other ants (75-77% for lridomyrmex sp; 64% for 'desert ants'; Bradshaw & Shoemaker 1967; Minnich & Shoemaker 1972), and termites and other in¬ sects (generally 50-80%; Redford & Dorea 1984). There are few values for the energy content of ants, but our values, which range from 20 to 40 kj gdry (Table 1), are typical for most biological samples (e.g. d'Oleire- Oltmanns 1977), including termites (32 kj gdr mass_1; Phelps et al. 1975). Hubert et al. (1981) reported a value of 17.3 kj g , *1 and 16.7 kj g , 1 for termites; val- ues determined here for ants range from 5 to 15 kj gwet *h The sodium and potassium contents of the ants mass r varied from about 50 to 250 pmol g, \ and the [sodium+potassium] content from about 100 to 500 umol g, These values are similar to those reported for desert ants of 140 pmol Na+ g*1 and 142 pmol K+ g*1 by Minnich & Shoemaker (1972), but are considerably lower than the values of 670 and 860 pmol g*1 reported by Bradshaw & Shoemaker (1967) for lridomyrmex sp consumed by the dragon Ctenophorus ornatus. The ash content of our ant species would be at least 1 to 2.5%, calculated from a composition of 750 pmol gdry ht_1 of Na* + K* + Cl*. The ash content of termites varies from <6 to >60%, generally being higher in workers than sol¬ diers, and higher in some species that consume soil rather than wood (Redford & Dorea 1984). Unfortu¬ nately, we were unable to determine the ash content for the ants which we studied, but presumably they are more like that of soldiers of wood-consuming species (6- 10%) than workers of some geophagous species(>60%). The nitrogen content of termites is generally about 1 to 10% of dry w eight (Redford & Dorea 1984). It is widely recognised that the digestion of ants and termites is difficult, in part due to their high chitin con¬ tent (Pianka & Pianka 1970; Redford & Dorea 1984; Pianka 1986). The chitin content of termites varies from 5.1% to 16.5% (Tihon 1946; Hubert et al. 1981), and ants are even more sclerotised than termites (Redford & Dorea 1984). The high ash content, particularly of soil¬ consuming worker termites, would also decrease the di¬ gestibility and energy content. Nevertheless, we are un¬ aware of any previous studies of the digestibility of ants or termites by myrmecophages, whether amphibian, reptilian, avian or mammalian. Our laboratory feeding trials with thorny devils, although only partly successful because none of the lizards consumed sufficient food to maintain a stable body mass, nevertheless indicate a low digestibility for ants. We estimate a metabolisable as¬ similation efficiency of about 59%, which is consider¬ ably less than that observed for lizards eating lightly- sclerotized insects (70-89% for mealworms and 70-90% for crickets), but is similar to that measured for heavily- sclerotized adult Tenebrio molitor (see Harwood 1979). Bell (1990) has estimated the dry matter assimilation of a "generalised" insect to be 78%, and the metabolisable energy assimilation to be 71%, assuming a total chitin content of 12.8% of dry mass, and excretion of nitrog¬ enous wastes and urate. Ants presumably are more sclerotised and have a higher chitin content than a "generalised" insect, and the metabolisable energy as¬ similation of 59% for thorny devils is expected to be lower than for a "generalised" insect. 9 Journal of the Royal Society of Western Australia, 78(1), March 1995 Field Turnovers Our studies indirectly enable the estimation of the field energy requirements of free-living thorny devils, and indicate the approximate values for water, energy, sodium and potassium turnover for free-ranging lizards (Table 2). If we accept a feeding rate of about 1500 ants per day, each weighing 0.5 mg wet weight and 0.2 mg dry weight, with an energy content of 30 kj gdr mass'1 and a metabolisable energy assimilation of 60%, then the cal¬ culated field metabolic rate is 5.4 kj day1. This is consid¬ erably more than the predicted field metabolic rate of 3.4 kj day1 for a 35g iguanid lizard (Nagy 1982a) and the actual field metabolic rate of 2.8 kj d 1 (Withers & Bradshaw 1995; Table 2). This discrepancy could be due to our over-estimating the feeding rate of thorny devils, the energy content of their food, or the digestibility of their food. The energy content of ants determined here is similar to expected values, and the metabolisable en¬ ergy assimilation of 59% does not seem unreasonably low (see above), and so the assumption of a food intake of 1500 ants d1 is the most probable error. A feeding rate of 750 ants d_1 yields the observed field energy turn¬ over rate. The feeding rate assumed above of 1500 ants day1 would confer a field water turnover rate (ignoring meta¬ bolic water production) of 0.45 ml day1, which is con¬ siderably lower than that predicted for a 30g semi-arid/ arid lizard (0.84 ml day1; Nagy 1982b) but greater than the measured field water turnover rate of 0.30 ml d'1 (Withers & Bradhsaw 1995; Table 2). Accounting for metabolic water production (estimated as 0.03 ml kj1; Withers 1992) increases the calculated field water turn¬ over rate to 0.61 ml day 5 A feeding rate of 750 ants d1 yields the observed field water turnover rate. We can also estimate the field sodium and potassium turnover rates for thorny devils. If the sodium and Table 2 Field water, energy, sodium and potassium turnover rates, and ratios of rates, calculated for thorny devils from their diet, as¬ suming a daily consumption of 1500 ants (wet mass 0.5 mg, dry weight 0.2 mg, 30 kj gdry mass\ 180 pmol Na+ gdry ^5 and 240 pmol K' gdry mass1, compared with the allometrically-predicted values (for a 35 g lizard) and observed field values. Calculated from Diet Predicted1 Measured2 Field Turnover Rates WTR (ml d1) 0.61 0.84 0.30 FMR (kj d1) 5.4 3.4 2.8 NaTR (pmol d1) 54 ~60 83 KTR (pmol d ') 72 - - Turnover Rate Ratios WTR/FMR (ml kj ’) 0.11 0.25 0.11 NaTR/WTR (pmol ml1) 89 ~71 277 KTR/WTR (pmol ml ') 118 - NaTR/FMR (pmol kj1) 10 ~18 30.2 KTR/ FMR (pmol kj ') 13 - . NaTR/KTR 1.3 - - 'Nagy (1982a, b); Bradshaw et al. (1987, 1991) 2Withers & Bradshaw (1995; data for trip 4) potassium contents of the ants are 180 and 240 pmol gdry m3ss_1 respectively, then the calculated turnover rates would be 54 pmol Na* day 1 and 72 pmol K' day1 (if all dietary sodium and potassium were assimilated into the body pool). Surprisingly, this sodium turnover is less than that measured of 83 pmol d \ despite our having apparently overestimated the feeding rate. Sodium turn¬ over has also been measured for other free-ranging liz¬ ards in order to estimate feeding rate. The field sodium turnover rate was about 1.0 to 1.2 pmol kg'1 d'1 for Lacerta viridis (Bradshaw et al. 1987), and 1.2 to 2.2 pmol kg'1 d*1 for Ctenophorus ( Amphibolurus ) nuchalis (Bradshaw et al 1991). These values are similar to the calculated value of 0.8 (750 ants d1) to 1.5 (1500 ants d1) pmol kg 1 d'1 for thorny devils. We are unaware of any studies which have determined the field potassium turn¬ over rate. Although we were unable to determine the actual field water, energy and ion turnovers of thorny devils in this study, our results predict ratios of field turnover rates (Table 2). The ratio of field water turnover (WTR) to energy turnover (FMR) calculated for thorny devils is 0.11 ml kj 1 (0.61-f5.4). This ratio depends on the energy and water content of the diet, and its metabolisable en¬ ergy assimilation (we assume that dietary water is 100% equilibrated with the lizard's body water pool), but not the absolute level of food consumption; it also assumes that the lizards are not able to drink in the field. Mea¬ surement of the ratio of WTR/FMR for thorny devils in the field as 0.11 ml kj'1 (Withers & Bradshaw 1995) sug¬ gests that we have correctly estimated the energy and water content of the diet, and its metabolisable energy assimilation, and that the thorny devils did not drink during the field study, although they clearly drink rain water and possibly dew when available (e.g. Bentley & Blumer 1962; Gans et al. 1982; Withers 1993; Sherbrooke 1993). The predicted ratio of WTR/FMR for a 30g semi- arid lizard is 0.25 (0.84-=-3.4), suggesting that these "pre¬ dicted" semi-arid lizards were drinking water. Acknowledgements: We are grateful to our numerous colleagues and stu¬ dents who assisted with these often arduous field studies of thorny devils. We sincerely thank Dr Jonathon Majer for identification of the various ants and John Dell for identification of the currant bush. We also thank Iain McLeod for assistance with ion analyses. All aspects of the study were undertaken with approval of the University of Western Australia Animal Welfare Committee and Radiation Protection Committee, the WA State Radiological Council, and the WA Department of Conservation and Land Management. A Roberts provided the photograph of the thorny devil. References Beard J S 1990 Plant Life of Western Australia. Kangaroo Press, Kenthurst, N. S. W. Bell G P 1990 Birds and mammals on an insect diet: A primer on diet composition analysis in relation to ecological energetics. Studies in Avian Biology 13:416-422. Bentley P J & Blumer W F C 1962 Uptake of water by the lizard Moloch horridus. Nature 194:699-700. Bradshaw S D & Shoemaker V H 1967 Aspects of water and electrolyte changes in a field population of Amphibolurus liz¬ ards. Comparative Biochemistry and Physiology 20:855-865. Bradshaw S D, Saint Girons H, Naulleau G & Nagy K A 1987 Material and energy balance of some captive and free-ranging reptiles in western France. Amphibia-Reptilia 8:129-142. 10 Journal of the Royal Society of Western Australia, 78(1), March 1995 Bradshaw S D, Saint Girons H & Bradshaw F J 1991 Patterns of breeding in two species of agamid lizards in the arid sub¬ tropical Pilbara region of Western Australia. General and Comparative Endocrinology 82:407-424. Cogger H G 1992 Reptiles and Amphibians of Australia. Reed Books, Chatswood. d'Oleire-Oltmanns W 1977 Combustion heat in ecological ener¬ getics: What sort of information can be obtained? In: Applica¬ tion of calorimetry in life sciences (ed I Lamprecht & B Schaarschmidt). Walter de Gruyter, Berlin, 315-324. Davey H W 1923 The moloch lizard, Moloch horridus. Victorian Naturalist 40:58-60. Davey H W 1944 Some lizards I have kept. Victorian Naturalist 61:82-84. Dell J, How R A, Newbey K R & Hnatiuk R J 1985 The Biological Survey of the Eastern Goldfields of Western Australia. Part 3. Jackson - Kalgoorlie Study Area. Western Australian Mu¬ seum, Perth. Duncan-Kemp A M 1933 Our Sandy Country. Angus & Robertson, Sydney. Gans C, Merlin R & Blumer W F C 1982 The water collecting mechanism of Moloch horridus reexamined. Amphibia-Reptilia 3:57-64 Greer A E 1989 The Biology and Evolution of Australian Lizards. Surrey Beatty & Sons, Chipping Norton. Harwood R H 1979 The effect of temperature on the digestive efficiency of three species of lizards, Cnemidophorus tigris, Gerrhonotus niulticarinatus and Sceloporus occidentalis. Com¬ parative Biochemistry and Physiology 63A:417-433. Heatwole H & Pianka E R 1993 Natural history of the Squamata. In: Fauna of Australia (ed C J Glasby, G J B Ross & P L Beesley). Australian Government Publishing Service, Canberra, 197-209. Hubert B, Gillon D & Adam F 1981 Cycle annuel du regime alimentaire des trois principals especes de rongeurs (Rodentia; Gerbilhdae et Muridae) de Bandia (Senegal). Mammalia 45:1-20. Minnich J E & Shoemaker V H 1972 Water and electrolyte turn¬ over in a field population of the lizard, Uma scoparia. Copeia 1972:650-659. Nagy K A 1982a Energy requirements of free-living iguanid liz¬ ards. In: Iguanas of the World: Their Behavior, Ecology, and Conservation (ed G M Burghardt & A S Rand). Noyes Publi¬ cations, Park Ridge, New Jersey, 49-59. Nagy K A 1982b Field studies of water relations. In: Biology of the Reptilia (ed C Gans & F H Pough). Academic Press, Lon¬ don, 483-501. Paton J 1965 Moloch horridus. South Australian Naturalist 40:25- 26. Phelphs R J, Struthers J K & Moyo S J L 1975 Investigations into the nutritive value of Macrotermes falciger (Isoptera: Termitidae). Zoologica Africana 10:123-132. Phillipson J 1964 A miniature bomb calorimeter for small bio¬ logical samples. Oikos 15:130-139. Pianka E R 1969 Sympatry of desert lizards Ctenotus in Western Australia. Ecology 50:1012-1030. Pianka E R 1986 Ecology and Natural History of Desert Lizards. Analyses of the Ecological Niche and Community Structure. Princeton University Press, Princeton. Pianka E R & Pianka H D 1970 The ecology of Moloch horridus (Lacertilia: Agamidae) in Western Australia. Copeia 1970:90- 103. Redford K H & Dorea J G 1984 The nutritional value of inverte¬ brates with emphasis on ants and termites as food for mam¬ mals. Journal of Zoology 203:385-395. Saville-Kent W 1897 The Naturalist in Australia. Chapman & Hall, London. Sherbrooke W C 1993 Rain-drinking behaviors of the Australian thorny devil (Moloch horridus: Agamidae). Journal of Herpe¬ tology 27:270-275. Sporn C C 1955 The breeding of the mountain devil in captivity. Western Australian Naturalist 5:1-5. Tihon L 1946 A propos des termites au point de vue alimentaire. Bulletin agricole do Congo beige 37:865-868. White S R 1947 Observations on the mountain devil ( Moloch horridus). Western Australian Naturalist 1:78-81. Whitten G J 1993 Family Agamidae. In: Fauna of Australia. Vol. 2A Amphibia & Reptilia (ed C J Glasby, G J B Ross & P L Beesley)- Australian Government Publishing Service, Canberra, 242-254. Withers P C 1981 The effects of ambient air pressure on oxygen consumption of resting and hovering honey bees. Journal of Comparative Physiology 141: 433-437. Withers P C 1992 Comparative Animal Physiology. Saunders College Publishing, Philadelphia. Withers P C 1993 Cutaneous water acquisition by the thorny devil ( Moloch horridus : Agamidae). Journal of Herpetology 27: 265-269. Withers P C & Bradshaw S D 1995 Water and energy balance of the thorny devil Moloch horridus : Is the devil a sloth? Am¬ phibia-Reptilia 16:47-54. Yeager D P & Ultsch G R 1989 Physiological regulation and con¬ formation: a BASIC program for the determination of critical points. Physiological Zoology 62:888-907. 11 Journal of the Royal Society of Western Australia, 78:13-14, 1995 New records and further description of Macrothrix hardingi Petkovski (Cladocera) from granite pools in Western Australia N N Smirnov1 & I A E Bayly2 1Institute for Animal Evolutionary Morphology and Ecology, Russian Academy of Sciences, Leninsky Prospekt 33, Moscow 117071, Russia department of Ecology and Evolutionary Biology, Monash University, Clayton VIC 3168 Manuscript received June 1994; accepted September 1994 Abstract New geographic records, taxonomic data and descriptive figures are presented for a cladoce- ran, Macrothrix hardingi Petkovski, which is recorded here from 8 new localities, all of which are shallow pools on granite rocks in Western Australia. This cladoceran species, which shows some similarities to M. longiseta Smirnov, is characterized inter alia by its dark black coloration and distinct occipital notch in the region of the head pore. Introduction Extensive collections of invertebrates from temporary pools on granite inselbergs in southern Western Austra¬ lia have been made by one of us (IAEB) during the win¬ ters of 1977, 1990, 1992 and 1993. Some results from these samplings have already been published (Bayly 1982; Frey 1991), but most of it remains unpublished. Despite this, it is already clear that these small ephem¬ eral aquatic habitats harbour a rich fauna (over 80 differ¬ ent metazoan taxa have thus far been recognised) with several species endemic to Western Australia. Frey (1991), for example, described a new genus of chydorid cladoceran ( Plurispina ) containing two species, both presently endemic to Western Australia, from the 1977 collections. Bayly (1992), in a paper dealing with some general aspects of granite rock-pools in Western Austra¬ lia, listed several additional taxa including the cladoce¬ ran Daphnia jollyi Petkovski, that are known only from these distinctive habitats. The cladoceran, Macrothrix hardingi, which was originally described by Petkovski (1973) from two localities near Merredin, should now be added to this list of species from granite rock-pools. Collections made in 1990 and 1993 contained an abun¬ dance of this conspicuous, black cladoceran. The cladoceran genus Macrothrix is well represented in Australia; there were 17 Australian species (including M. hardingi) out of a total generic complement of 34 species in a recent world review of the Macrothricidae (Smirnov 1992). Although Macrothrix hardingi was well figured by Petkovski (1973), this original description is not readily accessible to Australian workers. Conse¬ quently several new figures, including some of the fea¬ tures not included by Petkovski, are presented below. Distribution and Taxonomy Macrothrix hardingi Petkovski (Figs 1-12) Specimens examined Specimens were examined from the following locali¬ ties in Western Australia. Pool on south-west corner of © Royal Society of Western Australia 1995 Elachbutting Rock; 30°36'S, 118°37'E; 15.viii.1990; 49 fe¬ males. Pool on Sanford Rocks; 31°14'S, 118°46'E; 15.viii.1990; 9 females. Pool at southern edge of Mount Bailey [or Bayly]; 31°47'S, 119°07'E; 19.vi.1993; numer¬ ous females. Pool close to summit of Mount Bailey [or Bayly]; 31°47'S, 119°07'E; 19. vi. 1993; numerous females. Pool near base of Geeraning Rock; 30°32'S, 118°36'E; 20.vi.1993; numerous females. Pool just below summit of Baladjie Rock; 30°57'S, 118°52'E; 22.vi.1993; numer¬ ous females. Pool on Chiddarcooping Hill; 30°54'S, 118°41'E; 22.vi.1993; numerous females. Pool at western end of Weowanie Rock; 31°08'S, 119°45'E; 23. vi. 1993; numerous females. Designation of neotype Since it has been established that there is now no original type material of M. hardingi, a neotype (female) has been deposited in the Australian Museum (Sydney) [Registered number P42691]. Reference slides have also been deposited in the Zoological Museum of Moscow University [Registered numbers 3672-3679.] Description of female Body outline oval in lateral view (Fig 1), with a dis¬ tinct dorso-posterior angle. Dorsal outline arched, with occipital notch in region of head pore (Fig 2), slight hump posterior to occipital notch. Antennules (AI, Fig 3) rod-like, with groups of setae on distal part, with short sensory papillae. Antennae (All, Fig 4) with setae 0-0-1 -3/ 1-1-3, spines 0-1-0-1/0-0-1, spines much shorter than length of segments, largest seta of All (proximal seta of 3-segmented branch) with very short unilateral setules along entire length. Labral lamella (Fig 5) short, cuneiform, with blunt apex. Postabdomen (Figs 6, 7) with minute setules on anal margin, preanal margin bare, with weak serrations. Setae natatoriae (Fig 8) with very long distal segments (1.5 times as long as proximal segments); total length of the setae natatoriae 0.75 times that of body length. Inner distal lobe of thoracic limb I (Fig 9) with 3 setae of different length, bearing unilateral short setules; Fryer's forks present on thoracic limb 1 (Fig 10). Exopodites of thoracic limbs III (Fig 11) with 4 setae. Expodites of thoracic limbs IV (Fig 12) with 3 setae. Up to 10 parthenogenetic eggs in brood pouch. Length ca 1.5 mm. [The male is unknown.] 13 Journal of the Royal Society of Western Australia, 78(1), March 1995 8 Figures 1-8. Macrothrix hardingi Petkovski, drawn from material collected from Sanford Rocks, Western Australia. Scale bars in mm. 1, left lateral aspect of whole animal; 2, magnification of dorsal occipital notch and head pore; 3, AI (antennules); 4, All (antennae); 5, labrum; 6, postabdomen; 7, proximal portion of postabdomen; 8, seta nataloria. Differential diagnosis Macrothrix hardingi, in addition to its black colour, differs from other Macrothrix species (see Smirnov 1992) in AI not dilating distally and in having no strong spines. The largest antennal seta has short setules on one side only. The dorsal outline has a characteristic depression in the region of head pore. The postabdomen is not bilobed and has very short setules. Macrothrix hardingi is somewhat similar to M. longiseta Smirnov, differing from it in the absence of long setules along the distal segment of the outer distal lobe of tho¬ racic limb I and in the black colouration. It is important to note that Smirnov & Timms (1983) separated Echinisca Lievin from Macrothrix Baird on the key feature that the antennules are dilated distally in the latter but not in the former. On this basis, hardingi would have to be referred to Echinisca. However, Smirnov (1992) synonymised Echinisca with Macrothrix. Ecology Macrothrix hardingi lives in pools that are commonly less than 10cm deep in the middle, and is often abun¬ dant in the very shallow (less than 1cm deep) water at the extreme edges of the pools. It is very conspicuous to the naked eye by virtue of its jet-black colour against the light-coloured granite. In this respect it resembles Daph- nia jollyi, another cladoceran that is restricted to granite rock-pools in Western Australia. The pools inhabited by Macrothrix hardingi and Daphnia jollyi contain very clear water and it is physically impossible for these species to avoid strong light because of the extreme shallowness of the pools. It is likely that the heavy black pigmentation is photo-protective, preventing damage from potentially injurious ultraviolet light. The calanoid copepod, Boeckella opaqua Fairbridge, another species confined to the same type of habitat as these two cladocerans, is also invariably strongly pigmented but is bright red from the accumulation of carotenoids. Figures 9-12. Macrothrix hardingi Petkovski. 9, outer distal lobe and inner distal lobes of thoracic limb I; 10, Fryer's forks of thoracic limb I; 11, exopodite of thoracic limb III; 12, thoracic limb IV. Acknowledgments: This study was partly supported by a grant of the G Soros International Science Foundation to one of us (NNS). References Bayly I A E 1982 Invertebrate fauna and ecology of temporary pools on granite outcrops in southern Western Australia. Australian Journal of Marine and Freshwater Research 33:599- 606. Bayly I A E 1992 Freshwater havens. Landscope 7(4):49-53. Frey D G 1991 The species of Pleuroxus and of three related genera (Anomopoda, Chydoridae) in southern Australia and New Zealand. Records of the Australian Museum 43:291-372. Petkovski T K 1973 Zur Cladoceran-fauna Australiens II. Sididae und Macrothricidae. Acta Musei Macedonici Scientiarum Naturalium 13:161-192. Smirnov N N 1992 The Macrothricidae of the World. SPB Aca¬ demic Publishing, The Hague. Smirnov N N & Timms 1983 A revision of the Australian Cladocera (Crustacea). Records of the Australian Museum Supplement 1:1-132. 14 Journal of the Royal Society of Western Australia, 78:15-17, 1995 Preliminary observations on termite diversity in native Banksia woodland and exotic pine Pinus pinaster plantations M Abensperg-Traun1 & D H Perry2 1 CSIRO, Division of Wildlife and Ecology, LMB No 4, Midland WA 6056 2 26 Egham Road, Victoria Park WA 6100 Manuscript received September 1994 , accepted February 1995 Abstract Preliminary observations on termite diversity recorded 16 termite species from 10 genera for native Banksia woodland at Gnangara on the Swan Coastal Plain and 7 species from 5 genera for adjacent areas of 40 - 70 year-old exotic maritime pine Pinus pinaster plantations. Only 3 species were recorded eating pine; the remaining 4 species in pine plantations survived by foraging on remnant stumps, logs and roots of Eucalyptus only. We hypothesize that low termite diversity in pine may in part be due to replacement of palatable native woody plants with unpalatable exotic pine and changes in soil microclimatic conditions. Following plantation establishment, reinvasion by termites appears not to occur as readily as it does by other soil and litter arthropods. Introduction The diet of termites consists principally of cellulose obtained from living, dead but sound, or decomposed vegetation, humus or soil, or various combinations of the above (Wood 1978). Although the wood of many species of plant may be eaten, there is generally consid¬ erable partitioning of the available food resource within any particular habitat. Species of Coptotermes and Heterotermes, for instance, are able to eat relatively undecayed wood with high levels of plant chemical de¬ fences, while others such as species of Amitermes eat more decayed wood, including humus (Wood 1978; Miller 1991). Certain native and exotic tree species may be preferred or avoided by the termites (French et al. 1981; Postle & Abbott 1991; Abensperg-Traun 1993; Abensperg-Traun et al 1993). In Australia, the exotic pines Pinus pinaster and P. radiata have not been readily accepted by many termite species; notable exceptions are species of Coptotermes and Heterotermes , and the tropical Mastotermes danviniensis ( e.g . Gay 1957; Greaves 1959; Spragg & Paton 1980; Abensperg-Traun 1993). Replace¬ ment of native vegetation with pine plantations may thus be associated with a marked decline in termite di¬ versity. No previous Australian studies have compared termite communities in native woodland with commu¬ nities inhabiting pine plantations. The present study re¬ ports preliminary results of surveys on termite diversity in native Banksia woodland on the Bassendean sands of the Swan Coastal Plain, and on adjacent stands of 40 - 70 year-old maritime pine. Methods Study area The study was conducted at the Gnangara Pine Plan¬ tation (31°45'S, 115°48'E), about 20 km north of Perth, © Royal Society of Western Australia 1995 Western Australia, where pine plantations abutt onto native Banksia woodland. The soils are Bassendean sands, described in detail by McArthur & Bettenay (1960). The area has a mediterranean climate with cool wet winters and hot dry summers, with a mean annual rainfall of 866 mm (Bureau of Meteorology, Perth). Native vegetation on the Bassendean sands at Gnangara is mostly open Banksia woodland. Melaleuca raphiophylla (paperbark), Nuyisia floribunda (christmas tree), and species of Allocasuarina (she-oak) are also com¬ mon. Eucalypts such as Eucalyptus calophylla (marri), E. marginata (jarrah), £. todtiana (coastal blackbutt) and E. rudis (flooded gum) are largely restricted to the swales between coastal dune systems. The shrubby understorey is diverse and consists predominantly of species of Acacia, Xanthorrhoea, Hakea , jacksonia, Melaleuca and Adenanthos (Havel 1976; Heddle 1980). For the establishment of pine plantations at Gnangara in the late 1920's and early 1930's, all banksias and eucaiypts were clear-felled. The area was then burnt. Only logs of large trees survived the burn, and these were left on the ground. The area was then ploughed and pine seedlings were planted. Controlled burns in pine plantations at Gnangara are part of management strategy and usually follow thinning operations (C Sand¬ ers, CALM, pers comm). Termite sampling Termites were sampled opportunistically, using a spade and axe, by D H Perry from 1953 to 1966. The collections were made over an area of about 500 ha of Banksia woodland, and about 200 ha of two approxi¬ mately 30 year-old pine stands (Wetherell Block). The pine stands replaced a Banksia woodland containing large eucalypts, and remnants of these were still present at sampling time. Termites were sampled in the soil, mounds (termitaria) and timber, and a list of all col¬ lected species was recorded. The pine plantations origi- 15 Journal of the Royal Society of Western Australia, 78(1), March 1995 nally sampled for termites by D H Perry were again sampled in August and November 1994, allocating a to¬ tal of approximately 4 man-hours to each of two study areas. Approximately 2 ha were sampled in each area. Specimens were identified to species following Perry et al. (1985). Banksia woodland was not sampled in 1994. We therefore assume that termite diversity in the Banksia woodland at the present time resembles that when sampled about 35 years ago, and that the early surveys remain a valid control to the 1994 data. The first pine stand (Area 1; Compartment 15) was planted in 1930 and contained no large eucalypt trees prior to plantation establishment. It was clear-felled and replanted in 1962 following a fire. It had not been Thinned' when sampled in 1994, with a dense stand of 15 - 20 m tall pines at 2 - 3 m intervals. Canopy cover was close to 100 %, and a 15 to 20 cm deep layer of pine needles covered the soil surface suggesting that the stand had not been burnt since it was planted. There was no understorey of native (or exotic) plants with the exception of a small number of M. rhaphiophylla. No remnants of Banksia stumps or roots were located, so termite sampling was restricted to the soil, and to stand¬ ing and fallen dead pine. The immediately adjacent pine stand (Area 2; Com¬ partment 4) was planted in 1927. It supported Banksia as well as large jarrah and marri trees prior to the estab¬ lishment of pines, and carried its first crop of approxi¬ mately 30 m tall pines when sampled in 1994. The stand had been thinned to a density of approximately one tree at about 15 m intervals, giving a canopy cover of < 25 %. Although both areas had been under pine for over 60 years, they differed markedly as habitat for termites. In contrast to Area 1, Area 2 contained numerous remnants of eucalypt logs and stumps, and also regenerating balga ( Xanthorrhoea preissii), paperbark (M. raphiophylla), jar¬ rah (E. marginata) and marri (E. calophylla), which might provide foraging and nesting sites for termites. Low tree density and pine needle cover would facilitate higher levels of penetration of sun light, more closely resem¬ bling pre-plantation conditions. Results and Discussion The survey of Banksia woodland, about 35 years ago, identified 16 termite species from 10 genera (Table 1). Sampling within adjacent pine plantations identified only four species, three of which survived on a diet of pine ( Coptotermes acinaciformis raffrayi, C. michaelseni and H. platycephalus ); Amitermes medians was recorded forag¬ ing on jarrah logs only. In 1994, we recorded six species of termite in pine plantations, bringing the total termite diversity for pine plantations to 7 species (Table 1). In the unthinned stand, Area 1 which lacked remnants of eucalypt logs and stumps, we recorded only Coptotermes michaelseni. In the open, older stand, Area 2, we recorded two species of Coptotermes and one species of Heterotermes eating pine. Hesperotermes infrequens, Amitermes conformis and Xylochomitermes occidualis were recorded on remnant jar¬ rah logs and stumps; A. conformis was also sampled within mounds of C. a. raffrayi. Stumps and fallen tim¬ ber of M. raphiophylla did not support any termites. No Table 1 Termite species recorded from native Banksia woodland at Gnangara on the Bassendean sands of the Swan Coastal Plain, and species recorded where Banksia had been removed and re¬ placed with maritime pine ( Pinus pinaster). Banksia Termite species woodland Pine plantation Sampled Sampled Sampled in 1994 1953 1953 Without With to 1966 to 1966 Eucalyptus Eucalyptus Kalotermitidae Kalotermes hilli + Rhinotermitidae Coptotermes a. raffrayi + + + Coptotermes michaelseni + + + + Heterotermes occiduus + Heterotermes platycephalus + + + Termitidae Amitermes conformis + + Amitermes heterognathus + Amitermes modicus + + Amitermes sp. + Hesperotermes infrequens + + Microcerotermes newmani + Nasutitermes exitiosus + Occasitermes occasus + Pa racap ri termes kraepel in ii + Xylochomi termes occid ualis + + Xylochomitermes tomentosus + Total number of species 16 4 1 6 termites were found inhabiting the top layer of the soil, despite the availability of a rich humus layer. The gen¬ era recorded in Banksia woodland but not in pine plant¬ ations are Kalotermes, Microcerotermes, Nasutitermes, Occasitermes and Paracapritermes (Table 1). Termite diversity in pine plantations may have been low compared to Banksia woodland because the area sampled for termites, and sampling effort, was greater for Banksia than for pine areas. However, we hypoth¬ esize that two factors may have contributed to the lower termite diversity of pine plantations; (1) replacement of palatable native woody plants with exotic, unpalatable species, and (2) modification of soil microclimatic condi¬ tions. Chemical composition of the timber, and hence its palatability, is often of critical importance to foraging termites (Wood 1978; Wood & Johnson 1986). For ex¬ ample, Nasutitermes exitiosus does not appear to eat pine and this may be because essential oils in the timber con¬ tain chemicals which are an important constituent of the termites' alarm pheromone (Moore 1969). Consistent with our observations, termite attacks on pine baits in the central wheatbelt of Western Australia were also re¬ stricted to Coptotermes (1 species) and Heterotermes (1 species) despite the presence of a species-rich commu¬ nity of wood-eating termites (Abensperg-Traun 1993). Studies in eastern Australia show a progressive de¬ cline of Nasutitermes exitiosus colonies under pine plan¬ tations. This has been attributed to lower soil tempera¬ tures due to excessive shade rather than lack of food because the durable logs and stumps of eucalypt timber 16 Journal of the Royal Society of Western Australia, 78(1), March 1995 remained for many years (Ratcliffe et al. 1952; Gay 1957; Lee & Wood 1971). Our observations are consistent with these other studies. Several studies have shown that the diversity of other soil and litter arthropods is markedly lower in pine com¬ pared to relatively undisturbed native vegetation on similar soils (Springett 1971, 1976; Majer 1978, 1985; Rossbach & Majer 1983). Some of these studies (e.g. Springett 1976) also indicate that with increasing age of the plantation, arthropod diversity also increases. Suc¬ cessful reinvasion by other arthropods may be influ¬ enced by a progressive decline in canopy and pine needle cover through logging and fire. Our observations suggest that the rate of reinvasion by termites may be slower compared to many other arthropods, and that the return of native trees for food may be an important factor influencing reinvasion by termites. The observations reported here are clearly prelimi¬ nary. However, given that about 70 000 ha are under pine plantations in south-west Western Australia (Abbott 1993), an experimental study into the long-term effects of pine plantations on termite diversity is war¬ ranted. Acknowledgements: We thank C Sanders (CALM) for permission to collect termites at Gnangara. 1 Abbott, D Ewart and G Smith commented con¬ structively on the manuscript, for which we are grateful. References Abbott I 1993 Review of the ecology and control of the intro¬ duced bark beetle Ips grandicollis (Eichhoff) (Coleoptera: Scolytidae) in Western Australia, 1952 - 1990. CALMScience 1:35-46. Abensperg-Traun M 1993 A comparison of two methods for sam¬ pling assemblages of subterranean, wood-eating termites (Isoptera). Australian Journal of Ecology 18:317-324. Abensperg-Traun M, Black R & Bunn S 1993 Selection of woody food by termites (Isoptera) at the Harry Waring Marsupial Reserve near Jandakot, Western Australia. Western Austra¬ lian Naturalist 19:247-254. French J R J, Robinson P J & Bartlett N R 1981 A rapid and selective field assessment of termite wood feeding preferences of the subterranean termite Heterotermes ferox (Frogg) using toilet roll and small wood-block baits. Sociobiology 6:135-151. Gay F J 1957 Termite attack on radiata pine timber. Australian Forestry 21:86-91. Greaves T 1959 Termites as forest pests. Australian Forestry 23:114-120. Havel J J 1976 The potential of the northern Swan Coastal Plain for Pinus pinaster Ait. plantations. Forests Department of Western Australia, Perth, Bulletin No. 76. Heddle E M 1980 Effects of changes in soil moisture on the na¬ tive vegetation of the northern Swan Coastal Plain, Western Australia. Forests Department of Western Australia, Perth, Bulletin No. 92. Lee K E & Wood T G 1971 Termites and Soils. Academic Press, New York. Majer J D 1978 Preliminary survey of the epigeic invertebrate fauna with particular reference to ants, in areas of different land use at Dwellingup, Western Australia. Forest Ecology and Management 1:321-334. Majer J D 1985 Invertebrate studies in disturbed and pristine habitats of Dryandra State Forest. Forests Department of Western Australia, Perth, Research Paper No. 80. McArthur W M & Bettenay E 1960 The development and distri¬ bution of soils on the Swan Coastal Plain. CSIRO Soil Publica¬ tion, Melbourne, No. 16. Miller L R 1991 A revision of the Terrnes-Capritermes branch of the Termitinae in Australia (Isoptera: Termitidae). Inverte¬ brate Taxonomy 4:1147-1282. Moore B P 1969 Biochemical studies in termites. In: Biology of Termites (ed K Krishna & F M Weesner). Academic Press, New York, 407-432. Perry D H, Watson J A L, Bunn S E & Black R 1985 Guide to the termites (Isoptera) from the extreme south-west of Western Australia. Journal of the Royal Society of Western Australia 67:66-78. Postle A & Abbott I 1991 Termites of economic significance in suburban Perth, Western Australia: a preliminary study of their distribution and association with types of wood (Isoptera). Journal of the Australian Entomological Society 30:183-186. Ratcliffe F N, Gay F J & Greaves T 1952 Australian Termites. The Biology, Recognition and Economic Importance of the Com¬ mon Species. CSIRO, Melbourne. Rossbach M H & Majer J D 1983 A preliminary survey of the ant fauna of the Darling Plateau and Swan Coastal Plain near Perth, Western Australia. Journal of the Royal Society of West¬ ern Australia 66:85-90. Spragg W T & Paton R 1980 Tracing, trophollaxis and population measurement of colonies of subterranean termites (Isoptera) using a radioactive tracer. Annals of the Entomological Soci¬ ety of America 73:708-714. Springett J A 1971 The effects of fire on litter decomposition and on the soil fauna in a Pinus pinaster plantation. Annales de Zoologie 17:529-535. Springett J A 1976 The effect of planting Pinus pinaster Ait. on populations of soil microarthropods and on litter decomposi¬ tion at Gnangara, Western Australia. Australian Journal of Ecology 1:83-87. Wood T G 1978 Food and feeding habits of termites. In: Produc¬ tion Ecology of Ants and Termites (ed M V Brian). Cambridge University Press, Cambridge, 55-80. Wood T G & Johnson R A 1986 The biology, physiology, and ecology of termites. In: Economic Impact and Control of Social Insects (ed S B Vinson). Praeger, USA, 1-68. 17 Journal of the Royal Society of Western Australia, 78:19-24, 1995 Membership of The Royal Society of Western Australia 1994-1995 Notes — indicated for each member is membership status (ORD = Ordinary, HON = Honorary, AS = Associate; HAS = Honorary Associate, LIF = Life Member) and the date of joining; indicated in brackets is the date of subsequent changes in membership status, and when awarded the Royal Society of Western Australia medal, (b) is business address, (h) is home address. Abbott, I J ORD December 1976 (b) Department of Conservation & Land Management, Australia II Drive, Crawley WA 6009 ph:442-0309 fax:386-6399 interests: ecology, forests, islands, soil biology, science policy Abensperg-Traun, M ORD March 1991 (b) Division of Wildlife and Ecology, CSIRO, LMB No 4, Midland WA 6056 ph: 290-8132 fax: 290-8134 interests: ecology: invertebrates, myrmecophagy, conservation biology Ahmat, A L ORD May 1970 Kalgoorlie WA Allen, A D ORD July 1976 East Perth WA Arbuckle, D J F ORD October 1983 Farramere, South Africa Armstrong, C J ORD April 1991 Armadale WA Atkins, R P ORD November 1976 (b) Waterways Commission, 216 St Georges Terrace, Perth WA 6000 ph: 327-9738 fax: 327-9770 (h) 21 Hartung Street, Mundaring WA 6073 ph: 295-2954 interests: management of waterways Atkinson, D M M ORD July 1976 Claremont WA Atkinson, P R ORD May 1970 (b) 28 Ruskin Street, Parnell Auckland New Zealand ph: (09) 303-1893 fax: (09) 303-1612 Backhouse, J ORD July 1976 (b) Geological Survey of WA, 100 Plain Street, East Perth WA 6004 ph: 222-3159 interests: palaeontology, palynology, stratigraphy, geology, Palaeozoic and Mesozoic sedementary rocks in WA Bailey, J M ORD May 1988 Murdoch WA Baird, A M HON July 1928 (1973] Peppermint Grove WA Baker, SK ORD April 1987 Nedlands WA Balia, S A ORD December 1992 (b) Environmental Management, Water Authority of WA, Leederville WA 6007 ph: 420-3242 fax:420-3176 (h) 1 Wongan Avenue, Hilton WA 6163 ph: 335-5238 interests: ecology, wetlands Balme, B E ORD November 1958 Claremont WA Balme, H E HAS July 1980 Claremont WA Balme, J M ORD July 1976 Armidale NSW Bamford, M J ORD March 1988 (b) MJ & AR Bamford Consulting Ecologists, 23 Plover Way, Kingsley WA 6026 ph: 309-3671 interests: ecology, small vertebrates [frogs, reptiles, birds, mammals], fire, mine site rehabilitation Bannister, J L ORD June 1967 Perth WA Barley, ME ORD April 1987 Nedlands WA Bathgate, J A ORD (h) 14 Langler Street, East Victoria Park WA 6101 ph: 361-3416 interests: plant pathology, diseases of native plants Bayly, I A E ORD March 1992 (b) Department of Ecology & Evolutionary Biology, Monash University, Clayton VIC 3168 ph: (03) 905-5667 fax: (03) 905-5613 E-mail: IAN.BAYLY@SCI.MONASH.EDU.AU (h) 114 Belgrave-Hallam Road, Belgrave South VIC 3160 interests: limnology, animals, invertebrate zoology, ecology, WA granites Baynes, A ORD August 1967 Perth WA Beard, J S P HON June 1962 [1990] [Medallist 1983] Applecross WA Beggs, BJ ORD July 1971 Booragoon WA Bekle, H ORD July 1982 Minora WA © Royal Society of Western Australia 1995 Bell, D T ORD August 1977 (b) Botany Department, University of Western Australia, Nedlands WA 6907 ph: 380-215 fax: 380-1001 E-mail: dbell@uniwa.uwa.edu.au (h) ph: 450-3569 interests: plants, ecology Bennett, E M ORD September 1963 Gooseberry Hill WA Berry, PF ORD July 1980 Perth WA Bevan, A W R ORD Perth WA Birch, P V ORD May 1987 (b) Perth Observatory, Walnut Road, Bickley WA 6076 ph: 293-8255 fax: 293-8138 E-mail: pvbirch@earwax.pd.uwa.edu.au interests: astronomy Black, W R ORD April 1987 Nedlands WA Blyth, J D ORD April 1987 Wanneroo WA Boardman, W J ORD December 1984 Salter Point WA Borowitzka, M A ORD August 1986 (b) Biological & Environmental Sciences, Murdoch University, Murdoch WA 6150 ph: 360-2333 fax:310-3505 E-mail: borowitz@possum.Murdoch.edu.au interests: algae [biotechnology, primary production, toxins, sponge symbioses, blooms, ecology], coral reef, bioactive molecules, marine biology, seagrass epiphytes Bottomley, G A ORD July 1971 Claremont WA Bougher, A R ORD June 1987 Mount Hawthorn WA Bourgault du Coudray, P L STU September 1991 (b) Division of Environmental Science, Murdoch Univer¬ sity, Murdoch WA 6150 ph: 360-6396 fax: 310-4997 E-mail: bourgau@csuvaxl .Murdoch.edu.au (h) Lot 5296 Hud man Road, Boya WA 6056 ph: 299-8176 interests: soils [solute leaching, macropores, salinity control], salinity [solute leaching and movements, control], drylands Bowdler, S ORD July 1983 Nedlands WA Bowers, C L ORD June 1987 Kalamunda WA Bradley, JS ORD October 1993 Murdoch WA Bradshaw, S D ORD December 1965 (b) Zoology Department, University of Western Australia, Nedlands WA 6907 ph: 380-3769 fax:380-1029 E-mail: bradshaw@uniwa.uwa.edu.au (h) 156 Hensman Road, Subiaco WA 6008 ph: 381-5010 interests: comparative endocrinology, ecophysiology of desert animals, arid-zone, adaptation, philosophy of science, physiology Brearley, A STU October 1993 Nedlands WA Bridge, P J ORD March 1966 (b) Hesperian Press, 65 Oats Street, Carlisle WA 6101 ph: 362-5955 fax: 361-2333 (h) 61 Washington Street, Victoria Park WA 6100 interests: mineralogy, history Bridgewater, P ORD July 1976 Canberra ACT Briggs, I M ORD December 1983 (b) Agricultural and Resource Economics, Faculty of Agriculture, University of Western Australia, Nedlands WA 6907 ph: 380-2105 fax: 380-1098 E-mail: ibriggs@uniwa.uwa.edu.au (h) 37 Meriwa Street, Nedlands WA 6009 interests: natural resource management, water resources, regional resource planning Brooker, M G ORD April 1987 Midland WA Brown, R G ORD April 1987 (b) Geology and Geophysics Department, University of Western Australia, Nedlands WA 6907 ph: 380-2679 fax: 380-1037 E-mail: rbrown@uniwa.uwa.edu.au (h) 2 Rowan Place, Woodlands WA 6018 ph: 446-2171 interests: geology, sedimentology, quaternary history 19 Journal of the Royal Society of Western Australia, 78(1), March 1995 Bunn, E R ORD December 1985 West Perth WA Bunn, S E ORD October 1985 Nathan QLD Burbidge, A A ORD July 1983 Wanneroo WA Burbidge, A H ORD June 1978 (b) WA Wildlife Research Centre, Department of Conservation & Land Management, PO Box 51, Wanneroo WA 6065 ph: 405-5100 fax: 306-1641 interests: biogeography, birds, conservation, threatened birds Burman, R R ORD March 1988 Nedlands WA Bume, R V ORD May 1991 (b) Australian Geological Survey Organisation, PO Box 378, Canberra ACT 2601 ph: (06) 249-9291 fax: (06) 249-9970 E-mail: rbume@agso.gov.au interests: coastal geoscience, stromatolites, microbialites, coastal geoscience Burrows, N D ORD March 1994 (b) Department of Conservation and Land Management, 50 Hayman Road, Como WA 6152 ph: 334-0300 fax: 334-0327 (h) 21 Sandra Way, Rossmoyne WA 6148 ph: 354-2637 interests: forestry [science and management], fire [management, behaviour, ecology], ecology, arid-zones Butler, R J T ORD August 1965 South Guildford WA Calver, M C ORD July 1989 Bateman WA Cannon, J R ORD April 1987 (b) Department of Chemistry, University of Western Australia, Nedlands WA 6907 ph: 380-3152 fax: 380-1005 interests: chemistry, natural products, development in South East Asia Carey, S W LLF July 1933 Dynnyrne TAS Carstairs, S ORD Como WA Chalmer, P N ORD May 1981 Mount Barker WA Christensen, P E S ORD July 1978 Balingup WA Churchill, D M ORD May 1956 Apollo Bay VIC Cleverly, W H HON July 1938 [1982] Kalgoorlie WA Coates, D ORD Como WA Cockbain, A E ORD July 1976 South Perth WA Coleman, PJ ORD October 1963 Nedlands WA Collins, M T ORD March 1990 Melville WA Collins P ORD June 1987 Woodlands WA Colquhoun, I J ORD February 1983 Applecross WA Conacher, A J ORD October 1972 (b) Geography Department, University of Western Australia, Nedlands WA 6907 ph: 380-2705 fax: 380-1054 interests: soil, soil/slope relationships, rural environment, land degradation and management Congdon, R A ORD July 1975 (b) Department of Botany and Tropical Agriculture, James Cook University, Townsville QLD 4814 ph: (077) 81-4731 fax: (0 77) 25-1570 E-mail: robert.congdon@jcu.edu.au interests: plants [ecology, productivity, nutrient cycling, rainforest, wetlands], ecology, agroforestry, forestry Connell, S W ORD March 1988 Kallaroo WA Considine, J A ORD September 1987 (b) Plant Sciences, Faculty of Agriculture, University of Western Australia, Nedlands WA 6907 ph: 380-1783 fax: 380-1108 E-mail: consid@uniwa.uwa.edu.au (h) 201 Millbank Drive, Bidaminna WA 6503 ph: (096) 553009 fax: (096) 553009 interests: seeds [physiology, dormancy, germination], plants [development, productivity, improvement], floriculture, tropical fruit [growth, development] Cowling, W A ORD July 1983 (b) Department of Agriculture, Baron-Hay Court, South Perth WA 6151 ph: 368-3528 fax: 474-2840 (h) 81 Arlington Avenue, South Perth WA 6151 interests: plants, genetic resources Crawford, I M ORD November 1961 (h) 5 Kirway Street, Floreat Park WA 6014 ph: 387-3019 interests: archaeology, aboriginal studies Creed, K E ORD July 1987 Murdoch WA Cresswell, I D ORD July 1987 Belconnen ACT Crombie, D S ORD October 1985 Dwellingup WA Curry, S J HON December 1966 Swanbourne WA Davies, S J J ORD September 1977 Mount Helena WA Davis, C E S ORD March 1938 Floreat Park WA Davison, E M ORD April 1982 (b) Department of Conservation and Land Management, Brain Street, Manjimup WA 6258 ph: (097) 71-1988 fax: (097) 77-1183 (h) 148 Bateman Road, Mt Pleasant WA 6153 ph: 364-3816 interests: plant pathology, mycology, forestry, botany Davy, R ORD December 1966 (b) Geochemistry Section, Geological Survey of WA, 100 Plain Street, East Perth WA 6004 ph: 222-3153 fax: 222-3622 (h) 13-B Todd Avenue, Como WA 6152 ph: 367-3248 interests: geology, exploration, geochemistry, economic geology, geochemistry De Laeter, J R ORD June 1972 (b) School of Physical Sciences, Curtin University of Technology, GPO Box U 1987, Perth WA 6000 ph: 351-3045 fax: 351-3048 (h) 4 The Parapet, Burrendah WA 6155 ph: 457-2578 interests: astrophysics, science education, physics, mass spectrometry del Marco, A P ORD November 1992 North Perth WA Dell, B ORD June 1975 Murdoch WA Dickman, C R ORD April 1986 Sydney NSW Dixon, K W ORD June 1984 (b) Kings Park and Botanic Gardens, Kings Park Road, West Perth WA 6005 ph: 480-3629 fax: 322-5064 (h) 36 Branksome Gardens, Citv Beach WA 6015 ph: 385-7969 interests: plants [conservation, seed germination, micropropagation, germplasm storage, breeding], habitat restoration, taxonomy Dodd, J ORD April 1978 (b) Department of Agriculture, Baron-Hay Court, South Perth WA 6151 ph: 368-3679 fax: 474-3814 interests: weeds, biological control, plants, ecology, Rottnest Island, terrestrial vegetation, banksia woodland Dodds, F S ORD April 1972 Wembley WA Doupe, R G STU August 1993 West Perth WA Dyson, S E ORD October 1993 Mount Pleasant WA Edinger-Reeve, R ORD May 1988 (b) Department of Employment, Education, Training, WA Area South Office, 13th Floor QV1, 250 St Georges Terrace Perth WA 6000 ph: 429-3609 fax: 429-3707 Eliot, I G ORD August 1981 Nedlands WA Elkington, C R ORD September 1958 Kenwick WA Evans, C A ORD May 1990 Floreat WA Farrington, P ORD March 1981 (b) Division of Water Resources, CSIRO, Private Bag, Wembley WA 6014 ph: 387-0395 fax: 387-8211 E-mail: Peter@Per.DWR.CSIRO.au (h) 4 Thomas Way, Karrinyup WA 6018 ph: 341-1055 interests: plants, water use, salinity, agroforestry, hydrology, catchments, forestry, agroforestry Ferguson, W C ORD March 1978 St Kilda VIC Finucane, S J ORD (b) Dames and Moore, South Shore Centre, 85 The Esplanade, South Perth WA 6151 ph: 367-8055 fax: 367-6780 (h) 06/21 Angelo Street, South Perth WA 6151 ph: 367-8949 interests: ecology, arid zone biology, wetlands, environ¬ mental impact assessment and management Firman, J B ORD August 1988 (h) 14 Giddens Court, North Lake WA 6163 ph: 331-2742 Foulds, W ORD November 1981 (h) 86 Lakelands Drive, Gnangara WA 6065 ph: 306-3354 interests: plant ecology Fox, JED LIF July 1975 Bentley WA Friend, G R ORD April 1987 Wanneroo WA Friend, J A ORD April 1987 Wanneroo WA Froend, R H ORD September 1988 (b) Surface Water Branch, Water Authority of WA, P O Box 100, Leederville WA 6007 ph: 420-3118 20 Journal of the Royal Society of Western Australia, 78(1), March 1995 fax: 420-3176 E-mail: froend@essunl.Murdoch.edu.au (h) 60 Harris Road, Bicton WA 6157 ph: 339-2804 interests: plants [ecology, aquatic, terrestrial, water relations], wetlands, ecology, management, Gardner, P E ORD July 1992 North Fremantle WA Garnett, P J ORD (b) Edith Cowan University, Joondalup Drive, Joondalup WA 6027 ph: 405-5665 fax: 405-5717 E-mail: P.Garnett@cowan.edu.au (h) 17 Radstock Street, Karrinyup WA 6018 ph: 447-7724 interests: environmental chemistry, chemistry, education, environment Gazey, P STU November 1993 Maida Vale WA Geary, J K ORD September 1943 Karrinyup WA Gentilli, J HON September 1993 [1993] (b) Department of Geography, University of Western Australia, Nedlands WA 6907 ph: 380-2664 fax: 380-1054 (h) 65 Bruce Street, Nedlands WA 6009 ph: 386-2492 interests: climatology, biogeography, migrations George, A S ORD June 1961 (b) Tour Gables', 18 Barclay Road, Kardinya WA 6163 ph: 337-1655 fax: 337-9404 (h) as above interests: plants, systematics of Australian flora, natural history, editing texts George, P R ORD March 1987 Karrinyup WA Ghisalberti, E L ORD April 1989 Nedlands WA Gibson, N ORD April 1991 Wanneroo WA Gladstones, J S ORD March 1957 (b) 27 Pandora Drive, City Beach WA 6015 ph: 385-9436 (h) as above interests: agronomy, plant breeding, viticulture, ecology Glassford, D K ORD December 1984 Bullcreek WA Glover, J J E HON July 1956 [1982] Nedlands WA Goble-Garratt, D ORE) April 1989 (b) 117 Tower Street, West Leederville WA 6007 ph: 382-2383 fax: 382-2383 (h) as above interests: natural science, resource utilisation, urban environments Gole, MJ ORD December 1984 Gooseberry Hill WA Gordon, D M ORD December 1979 (b) LeProvost, Dames & Moore ph: 474-1933 fax: 368-2294 interests: ecology and physiology of aquatic plants, ecophysiology of mangroves, photosynthesis of marine and freshwater plants Gozzard, J R ORD April 1989 East Perth WA Graham, J ORD February 1973 Wembley WA Gray, N M ORD April 1948 (h) 7/9 Sutherland Street, Cremorne NSW 2090 ph: (02) 908-4443 interests: engineering geology, geophysics, regional geology Grayling, P M ORD October 1984 Fremantle WA Grieve, B J HON July 1948 [1975] [Medallist 1979] (b)PO Box 613, Nedlands WA 6009 (h) Unit 142, St Louis Estate Retirement Village, 3 Dean Street, Claremont WA 6010 ph: 384-4819 Griffin, E A ORD March 1978 Victoria Park WA Groves, D I ORD March 1971 Nedlands WA Guppy, M ORD (b) Biochemistry Department, University of Western Australia, Nedlands WA 6907 ph: 380-3331 fax: 380-1148 E-mail: mguppy@uniwa.uwa.edu.au interests: biochemistry, metabolism, metabolic depression Haig, D W ORD October 1985 Nedlands WA Hall, G P ORD June 1990 Como WA Hallam, S J ORD June 1967 (h) 2 Pool Street, York WA 6302 interests: archaeology (Australia), prehistoric settlement, ecology Hallberg, J A ORD August 1970 Neerabup WA Halse, S A ORD April 1992 Wanneroo WA Hamilton, R ORD March 1991 Como WA Hardy, G E St J ORD April 1994 Murdoch WA Hare, R ORD April 1950 San Lorenzo, Makati Metro Manila, Phillipines Harris, PG ORD July 1976 Nedlands WA Hart, R P ORD April 1982 (b) Hart, Simpson & Associates, 324 Onslow Road, Shenton Park WA 6008 ph: 388-3972 fax: 382-1395 (h) 21 Rankin Road, Shenton Park WA 6008 ph: 382-1086 interests: environmental impact assessment, ecology, wildlife, plants [disease, dieback], environmental management Harvey, M S ORD March 1992 (b) Curator of Arachnology, WA Museum, Francis Street, Perth WA 6000 ph: 427-2737 fax: 328-8686 E-mail: harveym@muswa.dialix.oz.au interests: arachnology, taxonomv [arachnids] Hatch, A B HON July 1958 [1982] Murdoch WA Hatcher, B G ORD April 1987 (b) CFRAMP-Resourcement Unit, Tyrell Street, Kingstown, St Vincent WEST INDIES ph: (809) 457-1904 fax: (809) 457-2414 (h) CFRAMP-Resource Assessment Unit, Tyrell Street, Kingstown WEST INDIES ph: (809) 456-9621 interests: benthic ecology, oceanography, conservation biology, fisheries, marine geology Hefter, G T ORD September 1993 Murdoch WA Hesp, P A ORD April 1992 Singapore Hilliard, R W ORD October 1993 (b) LeProvost Dames & Moore, Level 5 South Shore Plaza, 85 The Esplanade, South Perth WA 6151 ph: 474-1933 fax: 367-6780 (h) Lot 8 Burton Road, Darlington WA 6070 interests: environmental impact assessment, environmetnal management [marine, coastal, riverine], monitoring studies Hillman, R M ORD April 1987 Dalkeith WA Hnatiuk, R J ORD July 1976 Cook ACT Hobbs, A A ORD November 1992 Nedlands WA Hobbs, RJ ORD September 1984 Midland WA Hobbs, VJ ORD September 1 984 (b) Mathematical and Physical Sciences, Murdoch University, Murdoch WA 6150 ph: 360-2817 Hobday, JD ORD March 1974 Mosman Park WA Hodgkin, E P HON July 1946 [1975] (h) 86 Adelma Road, Dalkeith WA 6009 ph: 386-7895 interests: ecology, estuaries Hodgson, E C ORD July 1946 Daglish WA Hopkins, E R ORD November 1959 Manning WA Hopper, S D ORD December 1976 West Perth WA (b) Kings Park and Botanic Garden, West Perth WA 6005 ph: 480-3600 fax: 322-5064 interests: plant conservation biology, evolution, systemat¬ ics, botanic gardens Hopwood, J M ORD August 1989 (b) Mathematics Department, University of Western Australia, Nedlands WA 6907 ph: 380-3356 fax: 380-1028 E-mail: HOPWOOD@MATHS.UWA.EDU.AU (h) 20 Shakespeare Street, Leederville WA 6007 ph: 444-4804 interests: climate [modelling, dynamics of tropical cyclones, palaeoclimatology], soils, wind erosion, palaeoclimatology, monsoons, surface temperature history, inversion of borehole temperature Horwitz, P H J ORD July 1987 (b) Department of Environmental Management, Edith Cowan University, Joondalup Drive, Joondalup WA 6027 E-mail: P.Horwitz@cowan.edu.au interests: inland aquatic ecology, invertebrate biogeogra- phv, invertebrate conservation Hos, D PC ORD April 1974 (b) Unit 7, 21 McCabe Street, North Fremantle WA 6159 ph: 430-8460 fax: 430-8465 (h) 3 May Close, Mosman Park WA 6012 ph: 385-3505 interests: palynology [Australia, South East Asia], diatoms Hosken, D STU March 1994 Nedlands WA 21 Journal of the Royal Society of Western Australia, 78(1), March 1995 Howden, P R ORD October 1963 Attadale WA Ingram, B S ORD April 1966 (b) 19 Rennington Street, Dianella WA 6062 ph: 276-2775 fax: 276-2710 (h) as above interests: stratigraphic palynology Isted, W T C ORD December 1976 Mount Pleasant WA Jablonski, NG ORD October 1991 Nedlands WA James, S A STU November 1993 Woodlands WA Jasinska, EJ ORD December 1992 (b) Department of Zoology, University of Western Australia, Nedlands WA 6907 ph: 380-3970 fax: 380-1029 E-mail: edvtajj@uniwa.uwa.edu.au (h) 42 Majella Road, Mirrabooka WA 6061 ph: 344-3931 interests: animals [groundwater, cave, wetlands), tree root-fauna associations, aquatic microinvertebrates, freshwater springs, developmental plasticity Jenkins, CFH HON July 1929 [1973] [Medallist 1966) Claremont WA Jenkins, E A HON July 1933 [1965] Claremont WA Johnson, G I ORD May 1990 (b) Metex Resources NC, P O Box 204, South Perth WA 6151 ph: 474-2939 fax: 474-2937 interests: earth sciences Johnston, D B ORD March 1987 Attadale WA Johnstone, M H ORD September 1949 Carlingford NSW Jones, E ASO (h) 6 Gairloch Street, Applecross WA 6153 ph: 364-9702 interests: plants, native flora of WA Jones, M G K ORD April 1992 (b) WA State Agricultural Biotechnology Centre, School of Biological & Environmental Sciences, Murdoch University, Perth WA 6150 ph: 360-2424 fax: 310-3505 E-mail: mgkjones@Murdoch.edu.au (h) ph: 316-1369 interests: plants [genetic engineering, biotechnology, molecular biology, tissue culture, transformation] Jones, RAC ORD April 1992 South Perth WA Keeley, G J ORD July 1984 Palmyra WA Keighery, G J ORD April 1976 Wanneroo WA Kendrick, G W ORD December 1965 Perth WA Kenneally, K F ORD March 1969 Como WA Kinnear, AC ORD November 1978 Wembley Downs WA Kinnear, J E ORD November 1978 Wembley Downs WA Klemm, V V ORD June 1987 Halls Head WA Koch, L E HON March 1958 [1989] Salters Point WA Ladd, P ORD April 1987 (b) School of Biological & Environmental Sciences, Murdoch University, Murdoch WA 6150 ph: 360-2219 fax:310-4997 E-mail: ladd@esunl.Murdoch.edu.au interests: plants, ecology, palaeoecology Lamont, B B ORD March 1972 (b) School of Environmental Biology, Curtin University of Technology, PO Box U1987, Perth WA 6102 ph: 351-7368 fax: 351-2495 interests: ecology, conservation and ecophysiology of native plants Lawrence, M E ORD June 1987 Forrestfield WA Leary, D E STU March 1994 Flagstaff Hill SA Lemson, K L STU Nedlands WA Lenanton, R J ORD August 1974 North Beach WA LeProvost, M I ORD June 1987 South Perth WA Lintem, M J ORD May 1987 Mount Hawthorn WA Littlejohn, MJ ORD March 1956 Parkville VIC Loneragan, J F ORD November 1961 Murdoch WA Loneragan, N R ORD February 1981 Cleveland QLD Loneragan, W A ORD September 1963 (b) Department of Botany, University of Western Australia, Stirling Highway, Nedlands WA 6907 ph: 380-2216 fax: 380-1001' (h) 1 Desford Close, Shelley WA 6148 ph: 457-3238 interests: plants ecology, conservation, taxonomy, dendrochronology Longnecker, N ORD September 1993 Nedlands WA Lord, J H ORD November 1942 (h) 25 Hillway Street, Nedlands WA 6009 ph: 386-5795 interests: geology, mineral exploration Lund, M A ORD November 1993 Joondalup WA Luscombe, A F ORD May 1990 Carine WA Lynch, D A STU September 1990 Nedlands WA Lyons, TJ ORD July 1988 (b) Biological & Environmental Sciences, Murdoch University, Murdoch WA 6150 ph: 360-2925 fax: 310-3505 E-mail: lyons@atmos.Murdoch.edu.au interests: mesoscale meterology, boundary layer studies, air pollution meterology Macey, D J ORD September 1990 (b) Biological & Environmental Sciences, Murdoch University, Murdoch WA 6150 ph: 360-2363 fax: 310-3505 E-mail: macey@possum.Murdoch.edu.au interests: animals physiology, chitons, iron metabolism, thalassemia, lampreys Macfarlane, T D ORD December 1973 Manjimup WA Main, A R HON July 1951 [1982] Nedlands WA Main, B Y ORD September 1952 (b) Zoology Department, University of Western Australia, Nedlands WA 6907 ph: 380-2239" fax: 380-1029 interests: arachnology, spiders, taxonomy, biogeography, life history strategies, environmental and social history of WA wheatbelt Majer, J D ORD November 1974 (b) School of Environmental Biology, Curtin University of Technology, P O Box U1987, Perth WA 6001 ph: 351-7964 fax: 351-2495 E-mail: jmajerj@info.curtin.edu.au (h) 18 Catherine Street, Subiaco WA 6008 ph: 381-9281 interests: entomology, ants, conservation, environment, minesite restoration Manning, C R ORD December 1991 Applecross WA Marchant, N G ORD September 1963 Como WA Marsh, L M ORD November 1955 (b) WA Museum, Francis Street, Perth WA 6000 ph: 427-2751 fax: 328-8686 (h) 6 Lillian Street, Cottesloe WA 6011 ph: 383-2742 interests: taxonomy, echinoderms of Australia and Indo- Pacific, scleractinian corals, zoogeography Marshall, J K ORD June 1974 Wembley WA Masters, B K ORD June 1970 (b) PO Box 315, Capel WA 6271 ph: (097) 27-2474 fax: (097) 27-2670 (h) as above interests: geology, mineral sands, environmental manage¬ ment [natural areas, dairy waste, catchments wetlands], environmental impact assessment, post-mining rehabilita¬ tion McArthur, W M ORD June 1951 Alfred Cove WA McCaw, W L ORD June 1986 (b) Science and Information Division, Department of Conservation & Land Management, Brain Street, Manjimup WA 6258 ph: (097) 71-1988 fax: (097) 77-1183 interests: fire behaviour, fire ecology, forest management, semi-arid woodlands McComb, A J ORD May 1963 (b) School of Biological & Environmental Science, Murdoch University, Murdoch WA 6150 ph: 360-2191 fax: 310-4997 interests: water quality inestuaries and near-shore marine systems McDonald, D STU Innaloo WA McGrath, JF ORD June 1984 Busselton WA McMillan, R P ORD May 1956 (h) 64 St Ives, Northshore Dampier Avenue, Kallaroo WA 6025 ph: 307-8973 interests: insects [buprestidae, formicidae, antiquilines], photography, environmental entomology McNamara, K J ORD September 1979 (b) WA Museum, Francis Street, Perth WA 6000 ph: 328-4411 fax:328-8686 E-mail: mcnamk@muswa.dialix.oz.au interests: palaeontology, evolutionary theory Mees, G F ORD May 1959 Busselton WA Merrifield, H E ORD November 1969 Rockingham WA 22 Journal of the Royal Society of Western Australia, 78(1), March 1995 Merrilees, D HON July 1959 [1979] (h) RMB 256, Manjimup WA 6258 ph: (097) 72-3587 interests: geology, quaternary studies, archaeozoology Millington, A J ORD August 1952 Kununurra WA Mills, C H ORD October 1986 Bassendean WA Milne, V A ORD April 1987 Albany WA Moore, L S ORD March 1990 (b) Groundwater and Environment Branch, Water Authority, P O Box 100, Leederville WA 6902 ph: 420-2308 fax: 420-3176 (h) 4 Burrendah Boulevard, Willeton WA 6155 interests: stromatolites, environmental management, lake ecology, cyanobacteria Moore, R D ORD March 1994 Nedlands WA Morgan, B STU (h) 140 Townshend Road, Subiaco WA 6008 ph: 382-3548 interests: plant pathology Morgan, K H ORD August 1961 (b) K H Morgan and Associates, Bentley Business Centre, Unit 10, 4 Queen Street, Bentley WA 6102 ph: 384-3689/ 356-5455 fax: 351-9853 (h) 3 Jennifer Way, Rossmoyne WA 6148 ph: 354-3689 interests: groundwater hydrogeology, geological explora¬ tion and valuation, industrial water supply [particularly mining], environmental hydrogeology Morgan, W R ORD Como WA Morrison, D A ORD April 1987 Scarborough WA Morrison, P F ORD October 1993 South Perth WA Moulds, M S LIF July 1965 (b) Entomology Department, Australian Museum, P O Box A285, Sydney South NSW 2000 ph: (02) 339-8221 fax: (02) 360-4350 E-mail: maxm@amsg.Austmus.oz.au (h) 16 Park Avenue, Waitara NSW 2077 ph: (02) 487-2792 interests: entomology [cicadas, moths, butterflies] Muir, B G ORD August 1967 (b) Muir Environmental, 1400 Coulston Road, Boya WA 6056 ph: 299-6804 fax: 299-8302 (h) as above interests: environmental impact assessment and rehabili¬ tation [especially of mines and quarries], environmental policy, statutory compliance auditing, environmental toxicology, artificial wetlands, ecology of tropical environments Murphy, D P ORD May 1993 (b) Environmental Officer, Western Mining Corporation, LMK Operations, P O Box 22, Leinster WA 6437 ph: (090) 37-9929 fax: (090) 37-9090 (h) P O Box 194, Leinster WA 6437 ph: (090) 37-9929 interests: environmental management, avian biology, ecophysiology of arid-zone fauna, arid-zone ecology Myers, J S ORD September 1987 East Perth WA Nash, D W ORD August 1972 (b) D W Nash & Associates, 86 Marlow Street, Wembley WA 6014 ph: 387-4202 fax:387-1212 (h) 86 Marlow Street, Wembley WA 6014 ph: 387-4202 fax: 387-1212 interests: information technology, scientific computing [earth sciences], general computing Noel, D ORD August 1989 (b) PO Box 27, Subiaco WA 6008 ph: 385-3400 fax: 385-1612 interests: nut and tree crops, earth expansion, human society, ore genesis O'Shea, J E ORD November 1992 (b) Zoology Department, University of Western Australia, Nedlands WA 6907 ph: 380-2242 fax: 380-1029 E-mail: joshea@uniwa.uwa.edu.au interests: vertebrate morphology, reptiles, bats Oldham, J A ORD June 1987 (b) 44 Berkeley Crescent, Floreat Park WA 6014 ph: 387-8355 fax: 387-8355 interests: environmental management, natural resources management Oliver, K R ORD April 1994 Wilson WA Orsini, J P ORD June 1992 (b) National Threatened Species Network, 78 Stirling Highway, Perth WA 6000 ph: 384-3756 fax: 220-0652 (h) 15 Hooley Street, Swanboume WA 6010 ph: 384-3756 fax: 220-0652 interests: wildlife conservation, endangered species, biodiversity, malleeefowl, whales, flora, environmental policy and leglisation, environmental education Osborne, J M ORD September 1985 Perth WA Osman, A H ORD August 1962 (h) 24 Stanhope Road, Killara NSW 2071 ph: (02) 498-3615 fax: (02) 418-2132 interests: geology. South East Asia, reading Oxnard, C ORD October 1993 Nedlands WA Palmer, L ORD (b) University Department of Paediatrics, GPO Box D184, Perth WA 6001 ph: 340-8176 fax:388-2097 E-mail: lyle@ichr.uwa.edu.au (h) Unit 2, 165 Roberts Road, Subiaco WA 6008 ph: 382-3205 interests: genetic epidemiology of asthma, genetic epidemiology of cerebral palsy, paediatric epidemiology Parker, C A ORD August 1959 [Medallist 1986] Nedlands WA Parker, I N ORD August 1988 South Fremantle WA Pate, J S ORD October 1977 Nedlands WA Pearce, A F ORD December 1981 (b) Division of Oceanography, P O Box 20, North Beach WA 6020 ph: 246-8288 fax: 246-8233 E-mail: alan.pearce@per.ml.csiro.au (h) 8 Currajong Road, Duncraig WA 6023 ph: 246-2910 interests: oceanography (general), ocean currents, Leeuwin current, satellite remote sensing of the oceans Pearce, R H ORD March 1955 Kalamunda WA Pearman, G I ORD September 1963 (b) Chief, Division of Atmospheric Research, CSIRO, PMB #1, Mordialloc VIC 3195 ph: (03) 586-7650 fax: (03) 586-7553 E-mail: chief@dar.csiro.au (h) 67 Larnook Crescent, Aspendale VIC 3195 ph: (03) 580-5960 interests: global composition of the atmosphere, global carbon cycle,climate variability and change Peet, L J ORD August 1965 Dalkeith WA Perry, G ORD June 1966 Como WA Perry, M W ORD November 1971 South Perth WA Play ford, P E ORD August 1952 (b) Geological Survey of WA, 100 Plain Street, East Perth WA 6004 ph: 222-3157 fax:222-3633 E-mail: p.playford@dme.wa.gov.au (h) 102 Thomas Street, Nedlands WA 6009 ph: 386-4169 interests: geology, history, anthropology, heritage Podger, F ORD October 1993 Como WA Poole, W E ORD April 1951 Dickson ACT Potter, I C ORD July 1981 Salters Point WA Prendergast, W F LIF April 1985 Claremont WA Price, L K ORD March 1987 Belmont WA Price, M STU (b) Alan Tingay and Associates, 35 Labouchere Road, South Perth WA 6151 ph: 474-1300 fax:474-3394 (h) 47 Regent Avenue, Mt Pleasant WA 6153 ph: 364-3149 interests: zoology, environmental science Prider, R T HON July 1932 [1976] [Medallist 1970] Nedlands WA Purcell, P G ORD April 1989 (b) 141 Hastings Street, Scarborough WA 6019 ph: 245-2155 fax: 341-8679 (h) as above interests: petroleum exploration, conservation and development issues, aboriginal land rights issues Quilty, PG ORD June 1962 (b) Australian Antarctic Division, Channel Highway, Kingston TAS 7150 ph: (002) 32-3205 fax: (002) 32-3351 E-mail: pat_qui@antdiv.gov.au (h) 211 Nelson Road, Mt Nelson TAS 7007 ph: (002) 25-3217 interests: palaeontology, antarctica, early maritime exploration of Australia, music Ravine, D ORD June 1983 Cyncoed, Cardiff, Wales 23 Journal of the Royal Society of Western Australia, 78(1), March 1995 Rice, G E ORD September 1980 Ringwood East VIC Rich, PJ ORD April 1991 (b) University of the Americas, Cholula, Puebla 72820 MEXICO (h) Hoover Institution, Stanford University, Stanford California USA interests: computers in education Ride, W D L ORD September 1958 Hughes ACT Ridsdill-Smith, T J ORD December 1980 Wembley WA Roberts, A STU September 1994 Nedlands WA Roberts, J D ORD September 1993 (b) Department of Zoology, University of Western Australia, Nedlands WA 9607 ph: 380-2224 fax: 380-1029 E-mail: droberts@uniwa.uwa.edu.au (h) 106 Holland Street, Wembley WA 6014 ph: 387-2471 interests: amphibians, conservation, biogeography, evolution [particularly frogs) Roe, R ORD July 1969 Kensington WA Rohl, L ORD Wembley WA Rook, M K ORD April 1986 North Perth WA Rooke, I J ORD July 1982 Margaret River WA Sanders, C C ORD June 1968 Perth WA Sappal, K K ORD March 1991 Perth WA Saunders, D A ORD April 1983 Guildford WA Scott, J ORD February 1991 Maida Vale WA Scott, J K ORD September 1978 (b) CSIRO Division of Entomology, Private Bag, PO Wembley WA 6014 ph: 387-0644 fax: 387-0646 E-mail: JOHNS@CCMAR.CSIRO.AU interests: biological control of weeds Scott, W D ORD September 1991 Murdoch WA Searle, D J ORD June 1987 Dalkeith WA Seddon, G ORD June 1962 (b) Centre Studies Aust. Literature, Department of English, University of Western Australia, Nedlands WA 6907 ph: 380-2255 fax: 380-1030 (h) 'Lemarville', 186 High Street, Fremantle WA 6160 ph: 335-9005 fax: 335-9005 interests: environmental planning and design, cultural history, history of science Semeniuk, C A ORD October 1986 Warwick WA Semeniuk, V ORD March 1979 Warwick WA Shaw, S E ORD July 1959 North Ryde NSW Shearer, B L ORD May 1968 Como WA Shivas, R G ORD July 1988 Doubleview WA Shugg, H B ORD November 1959 (h) 5 McCallum Crescent, Ardross WA 6153 ph: 364-1226 interests: population control, maintenance of biodiversity, public administration Smith, E B J ORD October 1969 Attadale WA Smith, G G HON July 1951 [1979] Claremont WA Smith, L A ORD May 1972 Perth WA Sproul, A N ORD July 1978 (h) 38 Coogee Road, Ardross WA 6153 ph: 364-4349 interests: general entomology [particularly control of insects of agricultural importance] Stace, H ORD May 1993 (b) Botany Department, University of Western Australia, Nedlands WA 6907 ph: 380-1862 fax: 380-1001 E-mail: hmstace@uniwa.uwa.edu.au (h) 18/370 Barker Road, Subiaco WA 6008 ph: 382-4385 interests: plant genetics and evolution Start, A N ORD May 1979 Roleystone WA Steer, BT ORD July 1987 Como WA Stephens, LJ ORD August 1985 Roleystone WA Stephenson, N C ORD June 1969 (b) Department of Geology & Geophysics, University of New England, Armidale NSW 2351 ph: (067) 72-2860 fax: (067) 71-2898 interests: metamorphic petrology Stoneman, G ORD October 1988 (b) Department of Conservation & Land Management, Hackett Drive, Crawley WA 6009 ph: 442-0321 fax: 389-8603 (h) 1 Lowanna Way, City Beach WA 6015 ph: 385-7108 interests: forest science, forest ecology, forest hydrology, silviculture Sweetingham, M W ORD April 1990 East Fremantle WA Tassell, C B ORD July 1976 Launceston TAS Taylor, J C ORD May 1970 (h) 8 Circe Circle, Dalkeith WA 6009 ph: 386-1633 Teichert, C HON July 1938 [1975] (h) 5505 North 10th Street, Arlington Virginia 22205 USA interests: geology, palaeontology Terrill, S E HON July 1928 [1973] WA interests: family history Thomas, L N ORD April 1987 East Fremantle WA Tingay, A ORD May 1991 Darlington WA Tommerup, I C ORD April 1994 (b) CCMAR, CSIRO, Private Bag, Post Office, Wembley WA 6014 interests: plant/ fungal molecular biology, ecosystem sustainability Trendall, A F ORD May 1963 Applecross WA Triffitt, M A ORD November 1987 Nedlands WA Underwood, R INS Como WA Unkovich, MJ ORD October 1994 Nedlands WA Utting, E P LIF May 1963 Mosman Park WA Walker, C J STU March 1994 South Perth WA Walker, D I ORD August 1989 (b) Botany Department, University of Western Australia, Nedlands WA 6907 ph: 380-2089 fax: 380-1001 E-mail: diwalker@uniwa.uwa.edu.au interests: marine botany, education Wallace, G I ORD April 1980 Geraldton WA Wardell-Johnson, G W ORD November 1987 Bridgetown WA Warren, S (h) 24 Richardson Avenue, Dynnyrne TAS 7005 interests: geotectonics Watts, J STU Kelmscott WA Whitaker, D ORD November 1993 Kingsley WA Wilkinson, D S ORD April 1992 Mt Hawthorn WA Willmott, S P ORD June 1962 Forrestfield WA Wills, R T ORD April 1990 (b) W. A. Herbarium, P O Box 104, Como WA 6152 ph: 334-0500 fax: 334-0515 E-mail: rwills@wa.erin.gov.au (h) 56 Inverness Crescent, Menora WA 6050 ph: 271-8901 interests: ecology, conservation, plant disease, fire ecology, mycology, climate change, computing, geographic information systems Withers, P C ORD April 1987 (b) Zoology Department, University of Western Australia, Nedlands WA 6907 ph: 380-2235 fax: 380-1029 E-mail: pwithers@uniwa.uwa.edu.au (h) 2 Lookout Road, Kalamunda WA 6076 ph: 293-3115 interests: arid zone physiology and ecology of animals, natural history [frogs, lizards, birds, mammals], comput¬ ing, ecophysiology Woodall, R ORD September 1955 (b) Western Mining Corporation, P O Box 409, Unley SA 5061 ph: (08) 372-7220 fax: (08) 372-7250 (h) Wonnaminta House, P O Box 9A, Crafers SA 5152 ph: (08) 339-4550 fax: (08) 370-9461 interests: geology, geoscience, exploration management, mineral resources, petroleum resources Woods, PJ ORD June 1987 Cottesloe WA Wooler, S ORD May 1990 Rossmoyne WA Wooller, R D ORD October 1993 Murdoch WA Wurm, PAS ORD June 1987 Casuarina NT Wycherley, P R ORD March 1972 (h) 160 Nicholson Road, Subiaco WA 6008 ph: 381-8407 interests: economic and amenity botany, arboriculture, conservation, parks and recreation Youngson, W K ORD June 1987 Keysbrook WA Yu, B ORD April 1993 Sunnybank QLD 24 Journal of the Royal Society of Western Australia, 78: 25-27, 1995 Index of Interests aboriginal studies Crawford, I M agricultural science Jones, F G W agriculture: entomological Sproul, A N agroforestry Congdon, R A; Farrington, P agronomy Gladstones, J S air pollution Lyons, T J algae: Borowitzka, M A animal morphology O'Shea, J E animals: ants Majer, J D animals: archaeozoology Merrilees, D animals: bats O'Shea, J E animals: birds Withers, P C animals: cave Jasinska, E J animals: chitons Macey, D J animals: corals (scleractinian) Marsh, L M animals: echinoderms (Australia and Indo-Pacific) Marsh, L M animals: frogs Roberts, J D; Withers, P C animals: groundwater Jasinska, E J animals: groundwater, cave, wetlands Jasinska, E J animals: insects Moulds, M S animals: invertebrates Horwitz, P H J; Bayly, I A E animals: lampreys Macey, D J animals: lizards Withers, P C animals: malleeefowl Orsini, J P animals: pests Jones, F G W animals: physiology Macey, D J; Withers, P C animals: whales Orsini, J P Antarctica Quilty, P G anthropology Playford, P E aquatic microinvertebrates Jasinska, E J arachnology Harvey, M S; Main, B Y arboriculture Wycherley, P R archaeology Crawford, I M, Hallam S archaeozoology Merrilees, D arid-zone: ecology Burrows, N D; Finucane, S J arid-zone: ecophysiology Bradshaw, S D; Murphy, D; Withers, P C astronomy Birch, P V astrophysics De Laeter, J R atmosphere: boundary layers, air pollution Lyons, T J atmosphere: global composition, carbon cycle Pearman, G I bats O'Shea, J E bioactive molecules Borowitzka, M A biochemistry Guppy, M biodiversity Orsini, J P; Shugg, H B biogeography Gentilli, J; Roberts, J D biogeography: birds Burbidge, A H biogeography: invertebrates Horwitz, P H J biogeography: spiders Main, B Y biological control: weeds Scott, J K biology: arid zone Finucane, S J birds: biology Burbidge, A H; Murphy, D birds: conservation Burbidge, A H botany: Davison, E botanic gardens: Hopper, S chemistry: development in South East Asia Cannon, J R chemistry: education, environmental Garnett, P J chemistry: natural products Cannon, J R climatology Gentilli, J; Wills, R T climatology: modelling, dynamics of tropical cyclones, palaeodimatology Hop wood, J M climatology: variability, change Pearman, G I coastal geoscience Burne,R V computing Nash, D W; Rich, P; Wills, R T; Withers, P C conservation Abensperg-Traun, M; Dixon, K W; Hatcher, B G; Loneragan, W; Purcell, P; Roberts, J D; Wills, R T; Wycherley, P R conservation: insects Majer, J D conservation: invertebrates Horwitz, P H J conservation: plants Hopper, S; Lamont, B B conservation: threatened birds Burbidge, A H conservation: wildlife Orsini, J P corals Borowitzka, M A; Marsh, L M cultural history Seddon, G cyanobacteria Moore, L S cyclones Hopwood, J M dendrochronology Loneragan, W development issues Purcell, P developmental plasticity Jasinska, E J diatoms Hos, D P C disease: plants Bathgate, J A; Wills, R T drylands: solute leaching, movements Bourgault du Coudray, P L earth expansion Noel, D earth science Johnson, G earth science: computing Nash, D W ecology Abbott: I J; Hallam, S; Hart, R P; Wills, R T ecology: aquatic Froend, R H ecology: arid zone Burrows, N D; Finucane, S J; Murphy, D; Withers, P C ecology: benthic Hatcher, B G ecology: coral reef Borowitzka, M A ecology: ecosystem sustainability Tommerup, I C ecology: estuaries Hodgkin, E P ecology: fire McCaw, W L; Wills, R T ecology: forests Stoneman, G ecology: inland Horwitz, P H J ecology: invertebrates Abensperg-Traun, M ecology: lakes Moore, L S ecology: management Froend, R H ecology: palaeoecology Ladd, P ecology: plants Bell, D T; Congdon, R A; Foulds, W; Froend, R H; Gordon, D; Ladd, P; Lamont, B B; Loneragan, W ecology: Rottnest Island Dodd, J ecology: terrestrial Bamford, M J; Froend, R H; Withers, P C ecology: tropical Muir, B G ecology: viticulture Gladstones, J S ecology: WA granites Bayly I A E ecology: wetlands Balia, S A ; Froend, R H ecophysiology Bradshaw, S D; Gordon, D; Lamont, B B; Murphy, D; Withers, P C education: chemistry, environment Garnett, P J education: environment Orsini, J P education: science De Laeter, J R; Rich, P; Walker, D I endangered species Burbridge, A H; Orsini, J P endocrinology: comparative Bradshaw’, S D engineering: geology Gray, N M entomology: agricultural control Sproul, A N entomology: ants, conservation Majer, J D entomology: cicadas, moths, butterflies Moulds, M S entomology: environmental McMillan, R P environment: policy, leglisation, education Orsini, J P environment: rehabilitation, mine sites Bamford, M J; Majer, J D environment: rural land degradation and management Conacher, A ] environment: urban Goble-Garratt, D environmental chemistry Garnett, P J environmental entomology McMillan, R P environmental history: WA wheatbelt Main, B Y environmental hydrogeology Morgan, K H environmental impact assessment Finucane, S J; Hart, R P; Hilliard, R W; Masters, B K; Muir, B G environmental management Hart, R P; Hilliard, R W; Masters, B K; Moore, L S; Murphy, D; Oldham, J A environmental planning, design Seddon, G environmental policy Muir, B G environmental science Price, M epidemiology: asthma, cerebral palsy, paediatrics Palmer, L estuaries: ecology Hodgkin, E P estuaries: water quality McComb, A J evolution Hopper, S; McNamara, K J; Roberts, J D; Stace, H exploration: Lord, J H; Morgan, K H; Purcell, P; Woodall, R family history Terrill, S E fire: ecology Bamford, M J; Burrows, N D; McCaw, W L; Wills, R T fire: management Burrows, N D 25 Journal of the Royal Society of Western Australia, 78(1), March 1995 fire: small mammals Bamford, M J fisheries Hatcher, B G flora: native Jones, E; Jones, F G W floriculture Considine, J A forestry Abbott, I J; Burrows, N D; Davison, E; Stoneman, G forestry: agroforestry Congdon, R A; Farrington, P forestry: ecology Stoneman, G forestry: hydrology Stoneman, G forestry: management Burrows, N D; McCaw, W L forestry: silviculture Stoneman, G freshwater springs Jasinska, E J fungi: molecular biology Tommerup, I C genetic engineering: plants Jones, MGK genetics: epidemiology of asthma, cerebral palsy Palmer L genetics: plants Cowling, W A; Stace, H geochemistry Davy, R geographic information systems Wills, R T geology Backhouse, J; Davy, R; Merrilees, D; Noel, D; Osman, A H; Playford, P E; Stephenson, N C; Teichert, C; Woodall, R geology: economic Davy, R geology: engineering Gray, N M geology: exploration Davy, R; Lord, J H; Morgan, K H; Woodall, R geology: marine Hatcher, B G geology: mineral sands Masters, B K geology: Palaeozoic and Mesozoic sedementary rocks in WA Backhouse, J geology: quaternary Brown, R G; Merrilees, D geology: regional Gray, N M geology: resources Woodall, R geology: sedimentology Brown, R G geology: South East Asia Osman, A H geology: valuation Morgan, K H geophysics Gray, N M geoscience: coastal Bume, R V geo tectonics Warren, S habitat restoration Dixon, K W heritage Playford, P E history' Bridge, P J; Playford, P E history: environmental and social, WA wheatbelt Main, B Y history: family Terrill, S E history: science Seddon, G human society Noel, D hydrology: catchments Farrington, P hydrology: forests Stoneman, G hydrology: groundwater Morgan, K H hydrology: hydrogeology Morgan, K H information technology Nash, D W insects: buprestidae, formicidae, antiquilines, photography McMillan, R P iron metabolism Macey, D J islands Abbott, I J lakes: ecology Moore, L S land management Conacher, A J limnology Bayly, I A E management: natural resources, environment Oldham, J A management: waterways Atkins, R P management: wetlands Froend, R H mangroves: Gordon, D marine biology: seagrass epiphytes Borowitzka, M A marine botany Walker, D I marine geology Hatcher, B G marine systems: water quality, near-shore McComb, A J marine: early maritime exploration Quilty, P G metabolism: biochemistry Guppy, M; Withers, P C meterology: air pollution, mesoscale Lyons, T J microbialites Bume, R V migrations Gentilli, J mineral exploration Lord, J H mineral resources Woodall, R mineral sands Masters, B K mineralogy Bridge, P J mining: industrial water supply Morgan, K H molecular biology: fungi, Tommerup, I C molecular biology: plants Jones, MGK; Tommerup, I C monsoons Hopwood, J M music Quilty, P G mycology Davison, E; Wills, R T myrmecophagy Abensperg-Traun, M; Withers, P C natural history: editing texts George, A S natural history: terrestrial vertebrates Withers, P C natural products Cannon, J R natural resource management Briggs, I M; Oldham, J A natural science Goble-Garratt, D nematodes parasitic on plants Jones, F G W oceanography Hatcher, B G; Pearce, A F ore genesis Noel, D paediatrics: epidemiology Palmer, L palaeoclimatology: monsoons, surface temperature history, inversion of borehole temperature: climatology Hopwood, J M palaeoecology Ladd, P palaeontology McNamara, K J; Quilty, P G; Teichert, C palaeontology; Palaeozoic and Mesozoic sedimentary rocks Backhouse, J palynology: Australia, South East Asia, diatoms Hos, D P C palynology: Palaeozoic and Mesozoic sedimentary rocks Backhouse, J palynology: stratigraphic Ingram, B S parks and recreation Wycherley, P R pathology: plants Bathgate, J A; Davison, E.; Morgan, B petroleum exploration Purcell, P petroleum resources Woodall, R petrology: metamorphic Stephenson, N C philosophy of science Bradshaw, S D photosynthesis: Gordon, D physics: mass spectrometry De Laeter, J R physiology: animals Bradshaw, S D; Macey, D J; Withers, P C physiology: plants Gordon, D plants: aquatic Froend, R H plants: banksia woodland Dodd, J plants: biological control Scott, J K plants: biotechnology Jones, MGK plants: breeding Considine, J A; Dixon, K W; Gladstones, J S plants: conservation Dixon, K W; Hopper, S plants: conservation, ecophysiology Lamont, B B plants: development, productivity, improvement Considine, J A plants: disease Bathgate, J A; Hart, R P; Wills, R T plants: ecology Bell, D T; Congdon, R A; Foulds, W; Froend, R H; Ladd, P; Lamont, B B; Loneragan, W plants: economic and amenity botany Wycherley, P R plants: ecophysiology Lamont, B B plants: flora Jones, E; Jones, F G W; Orsini, J P plants: genetic engineering Jones, MGK plants: genetics, evolution Cowling, W A; Jones, MGK; Stace, H plants: germplasm storage Dixon, K W plants: marine botany Walker, D 1 plants: micropropagation Dixon, K W plants: molecular biology Jones, MGK; Tommerup, I C plants: native flora of WA Jones, E; Jones, F G W plants: nut and tree crops Noel, D plants: nutrient cycling Congdon, R A plants: parasitic nematodes Jones, F G W plants: pathology Bathgate, J A; Davison, E; Morgan, B plants: productivity Congdon, R A plants: rainforest Congdon, R A plants: seed germination Dixon, K W plants: systematics of Australian flora George, A S plants: terrestrial Dodd, J; Froend, R H plants: tissue culture Jones, MGK plants: transformation Jones, MGK plants: water relations Froend, R H plants: water use Farrington, P plants: weed biological control Dodd, J plants: weeds Scott, J K plants: wetlands Congdon, R A population control Shugg, H B prehistoric settlement: Hallam, S public administration Shugg, H B 26 Journal of the Royal Society of Western Australia, 78(1), March 1995 rainforest Congdon R A regional geology Gray N M regional resource planning Briggs I M rehabilitation: mine sites Bamford, M J; Masters, B K resources: mineral, petroleum Woodall, R resources: utilisation Goble-Garratt, D Rottnest Island: ecology Dodd, J salinity Bourgault du Coudray, P L; Farrington, P satellite remote sensing Pearce, A F science: education De Laeter, J R science: history Seddon, G science: policy Abbott, I J sedimentology Brown, R G seeds: germination Considine, J A; Dixon, K W seeds: physiology, dormancy Considine, J A silviculture Stoneman, G social history: WA Wheatbelt Main, B Y soils: biology Abbott, I J soils: solute leaching, macropores, salinity control, drylands Bourgault du Coudray, P L soils: wind erosion Flopwood, J M soils:soil/slope relationships Conacher, A J sponges Borowitzka, M A stratigraphy: Palaeozoic and Mesozoic sedimentary rocks Backhouse, J stromatolites Burne, R V; Moore, L S taxonomy Dixon, K W; Loneragan, W taxonomy: arachnids Harvey, M S taxonomy: echinoderms Marsh, L M taxonomy: plants George, A S; Hopper, S taxonomy: spiders Main, B Y thalassemia Macey, D J toxicology Borowitzka, M A tree root-fauna associations Jasinska, E J tropical fruit: growth, development Considine, J A urban environments Goble-Garratt, D viticulture Gladstones, J S water quality: estuaries, near-shore marine systems McComb, A } water resources Briggs, I M water: hydrology Morgan, K H water: industrial water supply, mining Morgan, K H waterways: management Atkins, R P weeds: biological control Dodd, J wetlands Congdon, R A; Finucane, S J; Masters, B K wetlands: artificial Muir, B G wetlands: ecology Balia, S A; Froend, R H wetlands: fauna Jasinska, E J wetlands: management Froend, R H wildlife Hart, R P woodlands: semi-arid McCaw, W L zoogeography: scleractinian corals Marsh, L M zoology Bayly, 1 A E; Price, M; Withers, P C 27 Journal of the Royal Society of Western Australia, 78:29-32, 1995 Recent Advances in Science in Western Australia Earth Sciences This study by J Clarke, of Western Mining Corpora¬ tion, confirms the great age of the landscape of the Yilgam Craton. The relict saprolite dates back to the Late Permian-Middle Jurassic. Stripping of this regolith occurred in the Middle Jurassic-Early Eocene during the formation of the palaeodrainage system. About 400 m of material was denuded from the craton to infill adjoining sedimentary basins, and further deep weathering occurred concurrently with erosion. Increasing aridity led to disorganization of the drainage to form a chain of lakes. Aridity increased in the Kambalda area in the Pliocene, leading to the establishment of the present-day semi-arid geomorphic regime. However, aridity has persisted for only 2% of the history of the Kambalda landscape, and arid geomorphic models are only appro¬ priate for understanding the most recent evolution of the regolith. Clarke J D A 1994 Geomorphology of the Kambalda region, Western Australia. Australian Journal of Earth Sciences 41:229-239. This thematic issue, edited by N Exon (AGSO, Canberra) documents various studies of the Mesozoic rocks of the North West Shelf, based on samples and seismic profiles acquired on two cruises of the AGSO research vessel Rig Seismic. Six papers examine palaeontology, one examines sedimentary petrology, one examines igneous petrology, three examine seismic profiling, and one examines the Wallaby profile. There is also an overview of the shelf and a synthesis of the results of the various papers. This volume provides an important contribution to our understanding of a major petroleum province. Exon N F (Editor) 1994 Thematic issue: Geology of the outer North West Shelf, Australia. AGSO Journal of Australian Geology & Geophysics 15:1-190. D Lowe, of Stanford University (USA), considers that three well-documented occurrences of stromatolites older than 3.2 Ga are abiotic. Small conical structures in the Strelley Pool chert (Warrawoona Group, WA, 3.5 - 3.2 Ga) probably formed through evaporitic precipita¬ tion. A domal structure from the North Pole chert (also Warrawoona Group) formed by soft-sediment deforma¬ tion. Domal and pseudocolumnar structures in lami¬ nated chert (Onverwacht Group, South Africa, 3.5 - 3.3 Ga) probably formed through inorganic precipitation. Lowe D R 1994 Abiological origin of described stromatolites older than 3.2 Ga. Geology 22:387-390. Most Neoproterozoic and Phanerozoic basins in Western Australia can be classified as Neoproterozoic, Palaeozoic or Mesozoic-Cainozoic based on their age of dominant fill and tectonic activity. This paper by R Hocking (GSWA, Perth) rationalizes the basin sub¬ divisions using a common set of descriptive terms. The Savory, Amadeus, Officer and possibly the Yeneena and Karara Basins are Neoproterozoic. The Gunbarrel, Southern Bonaparte, Ord, Canning and Southern Carnarvon Basins are Palaeozoic. The Northern Bonaparte, Browse, Roebuck and Northern Carnarvon Basins (together comprising the Westralian Superbasin) are Mesozoic-Cainozoic. The Perth Basin (with its outlier, the Collie Basin) contains both Palaeozoic and Mesozoic elements, while the Eucla and Bremer Basins are primarily Cainozoic depocentres. Hocking R M 1994 Subdivisions of Western Australian Neoprotozeroic and Phanerozoic sedimentary basins. Geologi¬ cal Survey of Western Australia Record 1994/4. G Le Blanc Smith (GSWA, Perth) describes the coal resources of the Collie Basin, a fault-bounded, post- depositional pull-apart basin preserved as a consequence of right-lateral shear in a transitional setting. The 226 km2 basin contains about 1200 m of Permian siliciclastics overlain by a veneer of Cretaceous rocks. The oldest Permian rocks were laid down in a glaciofluvial and glaciolacustrine setting, with fluvial to upper delta-plain alluvial coal deposition constituting the balance of the succession. The coal resources total 2400 Mt and are currently mined in four opencut and three underground mines; almost 80% is used for power generation. Over the part 100 years, about 100 Mt of coal has been produced. Le Blanc Smith G 1993 Geology and Permian coal resources of the Collie Basin, Western Australia. Geological Survey of Western Australia. Report 38. A remarkably well-preserved Frasnian radiolarian fauna is described by J C Aitchison (University of Sydney), from carbonate concretions in the Gogo Formation. It is the best preserved and most diverse Frasnian assemblage yet documented, with 57 species (41 new) assigned to 14 genera. All elements of the fauna are common, but it is dominated by ceratoikiscids of which 11 are new species. Aitchison J C 1993 Devonian (Frasnian) radiolarians from the Gogo Formation, Canning Basin, Western Australia. Palaeontographica Abt A 228:105-128. The incisoscutid arthrodire Gogosteus sarahae gen. et sp. nov is described and illustrated by J Long (Western Australian Museum). This taxon, like many of the superbly preserved arthrodires from the Gogo Forma¬ tion, is known from only a single specimen. It is repre¬ sented by most of the headshield, trunkshield and dentition; the bones are uncrushed and preserved in three dimensions. As a result of the new material, the family Incisoscutidae Denison 1984 is redefined and the new superfamily Incisoscutoidea is defined to include Incisoscutidae and Camuropiscidae. Long J A 1994 A second incisoscutid arthrodire (Pisces, Placodermi) from the Late Devonian Gogo Formation, West¬ ern Australia. Alcheringa 18:59-69. Deep seismic reflection and magnetic data were collected along a 75 km traverse across the Darling Fault Zone by researchers from Curtin University and the Uni¬ versity of Western Australia; 50 km of the traverse was over the Archaean Yilgam Craton and 25 km was over the Perth Basin. The seismic data suggest that the crust beneath the western Yilgarn Craton is divided vertically into three structural zones. From 0 to 7-9 km, there is a zone of thin-skinned compressional tectonism expressed 29 Journal of the Royal Society of Western Australia, 78(2), June 1995 as a series of thrusts, which links to a detachment surface at about 7-9 km depth. At 7-9 to 25 km, there is a zone with several minor detachments and a large intru¬ sive igneous body. From 25 to 40 km, there is a zone of continuous seismic events that dip east at about 25°. The seismic data provide a good fault plane image of the Phanerozoic Darling Fault, which is distinct from the Precambrian proto-Darling Fault that is interpreted to coincide with a wide, non-reflecting zone extending west beneath the Perth Basin. Middleton M F, Wilde S A, Evans B J, Long A, Dentith M & Morawa M 1995 Implications of a geoscientific traverse over the Darling Fault Zone, Western Australia. Australian Journal of Earth Sciences 42:83-93. Correlation of rocks in the Savory Basin with those of adjacent basins is impeded by poor outcrop and lack of subsurface information, but two new correlations by collaborators from Macquarie University and GSWA (Perth) clarify the age and relationships of the Savory basin. First, the Skates Hill Formation contains distinc¬ tive stromatolites (Acaciella australica) previously recorded from the Bitter Springs Formation of the Amadeus basin. Second, the intergrading sandstone- diamictite of the Boondawari Formation is very similar to the intergrading Pioneer Sandstone-Olympic Forma¬ tion of the Amadeus Basin; this correlation is also supported by carbon isotope chemostratigraphy. Walter M R, Grey K, Williams I R & Calver C R 1994 Stratigra¬ phy of the Neoprotoerozoic to Early Palaeozoic Savory Basin, Western Australia, and correlation with the Amadeus and Officer Basins. Australian Journal of Earth Sciences 41:533-546. The Archaean supracrustal sequence at Forrestania contains at least five komatiitic belts, which can be traced along strike for over 30 km. Lithologies range from olivine cumulates, which crystallized at the inter¬ face between substrate and flowing lava, to spinifex- textured rocks which underwent rapid crystallization within ponded lobes. Characteristic associations of different rock types are interpreted in terms of varia¬ tions in eruption rate, which presumably reflect proxim¬ ity to the vent or major lava river. The channel-facies rocks (olivine cumulates) are flanked by sequences of spinifex-textured flow-units that formed from the episodic overflow of lava from the distributary chan¬ nels. Perring C S, Barnes S J & Hill R T 1995 The physical volcanol¬ ogy of Archaean komatiite sequences from Forrestania, South¬ ern Cross Province, Western Australia. Lithos 34:189-207. The geochemistry of the Perseverance and neighbouring Rocky's Reward nickel deposits, which are associated with metamorphosed komatiite flows erupted onto a substrate of felsic crystal tufts, is described by an international collaboration of researchers from CSIRO (Perth), the University of Alabama, and the University of Melbourne. The deposits are overlain by the Persever¬ ance Ultramafic Complex, which consists of a central dunite lens flanked by olivine orthocumulates and spinifex-textured komatiites. Samples from the A-zones of komatiite flows fall into two distinct geochemical groups. Samples from flows flanking the central dunite show well-constrained linear trends of major and trace elements typical of komatiite suites related by simple olivine fractionation. Samples from the Perseverance mine, Rocky's Reward and from unmineralized flows at the base of the Persevereance Ultramafic Complex all show evidence for light REE enrichment relative to the other group. Geochemical computer modelling suggests that the mineralization is due to wholesale assimilation of floor rocks close to the site of sulphide accumulation. Barnes S J, Lesher C M & Keays R R 1995 Geochemistry of mineralised and barren komatiites from the Perseverance nickel deposit. Western Australia. Lithos 34:209-234. Rock magentic properties and palaeomagnetism of weakly metamorphosed banded iron-formation (BIF) from the Palaeoproterozoic Hamersley Group and Protoerozoic BIF-derived iron ores were investigated by P Schmidt and D Clark (CSIRO, North Ryde). Palaeomagnetic pole positions were calculated for BIFs at Paraburdoo (40.9°S, 218.9°E) and Wittenoom (36.4°S, 218.9°E) and for Mt Tom Price iron ore (37.4°S, 220. 3°E) and Paraburdoo iron ore (36.4°S, 209. 9°E). These poles are indistinguishable from each other, and from the Mt Jope Volcanics overprint pole. The magnetization of the BIFs was probably acquired during burial metamor¬ phism of the Hamersley Group. Schmidt P W & Clark D A 1994 Palaeomagnetism and mag¬ netic anisotropy of Protozoic banded-iron formations and iron ores of the hamersley basin, Western Australia. Precambrian Research 69:133-135. Solid bitumen envelopes in Permian sandstones from the Kennedy Group are shown by B Rasmussen and J Glover (University of Western Australia) to be useful in subdividing the diagenetic sequence into three intervals, thus establishing the order of appearance of diagenetic minerals with greater precision than was previously pos¬ sible. The envelopes develop from the polymerization of hydrocarbons around the radioactive detritals monazite, xenotime and zircon. Their paper reports the presence of authigenic florencite (Ce, La, A1 phosphate) and xenotime (YP04), and bitumen-filled fission tracks in monazite for the first time in Australian sedimentary rocks. Rasmussen B & Glover J E 1994 Diagenesis of low-mobility elements (Ti, REEs, Th) and solid bitumen envelopes in Permian Kennedy Group sandstone. Western Australia. Jour¬ nal of Sedimentary Research A64:572-583. Life Sciences This most recent book published by John Long, of the Western Australian Museum, entitled 'The Rise of Fishes', provides a superbly illustrated history of the evolution of the fishes. It integrates aspects of geology, evolution, anatomy, and global environmental 30 Journal of the Royal Society of Western Australia, 78(2), June 1995 complete acccount of the spectacular fossil history of the fishes over their 500 million year record, and provides a thorough discussion of the latest evidence for the evolu¬ tion of the first fishes from invertebrates. Included in the book are hundreds of colour photographs and dramatic full-colour reconstructions of extinct fishes. Long J A 1995 The Rise of Fishes. 500 Million Years of Evolu¬ tion. Johns Hopkins University Press, Baltimore/University of New South Wales Press, Sydney. This monograph by D Bickel examines all 253 species (with 208 newly described) of dipteran flies of the sub¬ family Sciapodinae. It describes aspects of morphology, fossil history, systematics, natural history, biogeography and accidental introductions, and systematics. It reviews all zoogeographic regions although it emphasises Australia and the Orient. Nine new genera are erected, and there are nomenclatural changes for taxa from all biogeographic regions. Bickel D J 1995 The Australian Sciapodinae (Diptera: Dolichopodidae), with a review of the Oriental and Australasian faunas, and a world conspectus of the subfamily. Records of the Australian Museum. Supplement 21 Patterns of morphological diversity are examined by J Clements (University of Western Australia) and W Cowling (Department of Agriculture, WA) for 157 accessions of wild lupins Lupinus angustifolius from the Aegean region. Two groups with desirable agronomic characteristics originated in the Dhodhekanisos Islands; these had rapid and tall growth, prolific podding on the main stem, pods high off the ground, many upper lateral branches, large leaves, pods and seeds, and high seed yield. Clements J C & Cowling W A 1994 Patterns of morphological diversity in relation to geographical origins of wild Lupinus angustifolius from the Aegean region. Genetic Resources and Crop Education 41:109-122. Researchers from Murdoch University and the De¬ partment of Lands, Parks and Wildlife (Hobart) demon¬ strate a substantial correlation between the breeding ages of short-tailed shearwaters in pair bonds. This is largely explained by the prolonged pair-bond between mates and the predominant availability of unpaired, in¬ experienced birds. The probability of breeding success depends on the breeding ages of the partners and especially the length of their pair bond. Bradley J S, Wooller R D & Skira I J 1995 The relationship of pair-bond formation and duration to reproductive success in short-tailed shearwaters Puffinus tenuirostris. Journal of Ani¬ mal Ecology 64:31-38. A survey of the Cenozoic fossil fauna of Barrow Island in nothwestern Australia, by K McNamara and G Kendrick (Western Australian Museum), has yielded a rich fauna of Middle Miocene and Late Pleistocene molluscs, and Middle Miocene echinoid echinoderms. These faunas have little in common with the contempo¬ raneous faunas from the Nullarbor Limestone in the Eucla Basin; the echinoid fauna has more in common with Miocene faunas of Java and India. The mollusc and echinoid faunas show strong affinities with modern communities. McNamara K J & Kendrick G W 1994 Cenozoic molluscs and echinoids of Barrow Island, Western Australia. Records of the Western Australian Museum. Supplement 51. The traditional and current uses of the genus Eremophila , Australian desert shrubs, have been reviewed by G Richmond (Curtin University) and E Ghisalberti (University of Western Australia). This ecologically-important genus is represented by over 200 species, of which 75% occur in Western Australia. Species of Eremophila have long been of cultural importance to the Aboriginal people. They are useful in revegetation programmes and have poten¬ tial in horticulture and as sources of phytochemicals. Richmond G S & Ghisalberti E L 1994 The Australian desert shrub Eremophila (Myoporaceae): medicinal, cultural, horticul¬ tural and phytochemical uses. Economic Botany 48:35-59. Exposure of dormant seed to cold smoke, derived from burnt native vegetation, has a positive influence on germination in 45 of 94 species of native vegetation examined by researchers from Kings Park and Botanic Garden, and the University of Western Australia. Some species showed earlier germination if smoke-treated, whereas other species continued to germinate for a longer period of time if smoke-treated. Dixon K W, Roche S & Pate J S 1995 The promotive effect of smoke derived from burnt native vegetation on seed germina¬ tion of Western Australian plants. Oecologia 101:185-192 Researchers from the University of Western Australia have shown that the lesser long-eared bat Nyctophilus geoffroyi was attracted to both calling tettigoniid bushcrickets and synthesised tettigoniid calls broadcast through loud-speakers. More bats were attracted to long-duration than short-duration calls, whereas high calling rate or high calling intensity had no special at¬ traction. Two-speaker choruses appeared to be more at¬ tractive than single-speaker choruses. These results are interpreted in the context of predation being a selective force on calling strategies of tettigoniid bushcrickets. Hosken D J, Bailey W J, O'Shea J E & Roberts J R 1994 Localisation of insect calls by the bat Nyctophilus geoffroyi (Chiroptera: Vespertilionidae): a laboratory study. Australian Journal of Zoology 42:177-184. B Lamont and E Witkowski provide a stage-by-stage protocol for identifying simple or biased lottery, or non¬ lottery, patterns of seedling recruitment. For four co¬ occurring species of Banksias after two experimental fires, seedlings still alive after 3 years of two ( Banksia speciosa and B. Baxteri ) conformed to a biased lottery whereas two (B. coccinea and B. pulchella) has no math¬ ematical structure. Lamont B B & Witkowski E T F 1995 A test for lottery recruit¬ ment among four Banksia species based on their demography and biological attributes. Oecologia 101:299-308. Physical Sciences A group of researchers from the Chemistry Depart¬ ment at the University of WA and the Research School of Chemistry at the ANU in Canberra have made a detailed study of the synthesis and structure of bicyclic hexamine ligands, which can act as cages for metal ions. These ligands are very strong bases, accepting up to five protons with pK^'s ranging from 0 to 12. Bottomley G A, Clark 1 J, Creaser 1 1, Engelhardt L M, Geve R J, Hagen Ks, Harrowfield J M, Lawrance G A, Lay P A, Sargeson A M, See A J, Skelton B W, White A H & Wilner F R 1994 The synthesis and structure of encapsulating ligands: 31 Journal of the Royal Society of Western Australia, 78(2), June 1995 properties of bicyclic hexamines. Australian Journal of Chem¬ istry 47:143-179. Electrical conductivities of simple electrolytes dissolved in 2-cyanopyridine have been measured by GT Hefter (Murdoch University) and M Salomon (US Army Power Sources Laboratory). Only weak association occurs in dilute solution because of the very high dielectric constant. A conductivity maximum is observed in the very concentrated solutions typical of nonaqueous batteries. Hefter GT & Salomon M 1994 Conductivities of 1:1 salts in 2- cyanopyridine. Journal of Solution Chemistry 23:579-593. Researchers at Murdoch University have made a quantum mechanical study of the effect of excess charge on bond strengths in silane molecules. The calculations show that both positive and negative charges reduce the bond strengths so that they can be broken even by infrared light. This effect may explain the well-known photo¬ degradation of amorphous silicon solar cells. Clare BW, Talukder G, Jennings PJ, Cornish JCL & Hefter GT 1994 Effect of charge on bond strength in hydrogenated amor¬ phous silicon. Journal of Computational Chemistry 15:644-652. Chemists at Murdoch University have been able to synthesize asymmetric di-acetate derivatives of benzo [c] pyrans using titanium tetraisopropoxide. These com¬ pounds are related to aphid pigments. Giles RGF & Joll CA 1995 An asymmetric synthesis of benzo [c] pyrans related to the aphid pigments. Tetrahedron Letters 36:1125-1126. Note from the Hon Editor: This column helps to link the various disciplines and inform pthers of the broad spectrum of achievements of WA scientists (or others writing about WA). Contributions to "Recent Advances in Science in Western Australia" are welcome, and may include papers that have caught your attention or that you believe may interest other scientists in Western Aus¬ tralia and abroad. They are usually papers in refereed journals, or books, chapters and reviews. Abstracts from conference proceedings will not be accepted. Please submit either a reprint of the paper, or a short (2-3 sentences) summary of a recent paper together with a copy of the authors' names and addresses, to the Hon Editor or a member of the Publications Committee: Dr S D Hopper (Life Sciences), Dr A E Cockbain (Earth Sciences), and Assoc Prof G Hefter (Physical Sciences). Final choice of articles is at the discretion of the Hon Editor. "Letters to the Editor" concerning scientific issues of relevance to this journal are also published, at the dis¬ cretion of the Hon Editor. Please submit a word process¬ ing disk with letters, and suggest potential reviewers or respondents to your letter. P C Withers, Hon Editor , Journal of the Royal Society of Western Australia. 32 Journal of the Royal Society of Western Australia, 78:33-38, 1995 The value of macroinvertebrate assemblages for determining priorities in wetland rehabilitation: A case study from Lake Toolibin, Western Australia. R G Doupe1 & P Horwitz Department of Environmental Management, Edith Cowan University, Joondalup Drive, Joondalup WA 6027 Present Address: PO Box 101, Kununurra WA 6743 Received June 1994 , accepted April 1995 Abstract The use of macroinvertebrates in environmental assessment is well known, but is often per¬ ceived as costly and time-consuming. Using preliminary macroinvertebrate and associated salin¬ ity data recorded at Lake Toolibin and adjacent wetlands, we discuss the salinization process and how observed faunal assemblages might be used to assess a rehabilitation program. Ninety taxa were collected from two sampling occasions, most of them being saline-tolerant forms represent¬ ing widespread groups with high dispersive powers. Predominantly freshwater forms not found, but expected to occur, can be used as indicator taxa for recovery. In addition to the need for regular monitoring of the fauna, we argue for a greater emphasis on ensuring the recovery of Lake Walbyring, which contained a demonstrably different suite of macroinvertebrates. Introduction Lake Toolibin, approximately 200km south-east of Perth, lies at the head of a series of seasonal or ephem¬ eral lakes which form the headwaters of the Arthur River, a tributary of the Blackwood River in south-west¬ ern Australia, The clearing of native vegetation has resulted in both a rise in saline groundwater levels and an increase in saline surface run-off into the headwaters of the Arthur River (Halse 1988). Most wetlands of the system, and other parts of the Western Australian wheatbelt, have been severely affected by secondary salinization. However, Lake Toolibin has remained com¬ paratively fresh. Lake Toolibin Nature Reserve is a registered "Wet¬ land of International Importance" under the Ramsar Convention. It has important conservation value as a breeding habitat for native waterfowl, including rare species (Halse 1987), and because of the occurrence of restricted vegetation types (Froend ct al. 1987). The Department of Conservation and Land Management of Western Australia and the local Lake Toolibin Catchment Committee are engaged in tree planting and overland flow interception and diversion operations in the Lake Toolibin catchment in an effort to prevent any further decline of this Reserve. Recent efforts to monitor these restorative activities include the work of Bell & Froend (1990), who recorded the decline and mortality of wet¬ land trees at Lake Toolibin over a five year period. The recovery plan for Lake Toolibin (Anon. 1992) © Royal Society of Western Australia 1995 suggested that aquatic invertebrate surveys should be undertaken to provide a measure of water quality, and to indicate whether recovery criteria were being met. Faunal composition should show a rapid response to changing wetland conditions because many macro¬ invertebrate taxa possess a mobile phase in their life cycle, hence they can colonize suitable habitats rapidly. In the case of Lake Toolibin, the hypothesis is that water salinity will largely dictate the type of fauna that will be found. While the "coarse relationship" (sensu Williams et al. 1990) between salinity and faunal composition is unquestionable, the role of salinization in shaping aquatic communities remains poorly understood (De Deckker 1983). The work of Williams et al. (1991) suggested that Australian macroinvertebrates are more tolerant to salinity than previously thought, although an alternative explanation is that only euryhaline forms persist from a once more diverse fauna. At Lake Toolibin, neither the structure of aquatic inver¬ tebrate communities, nor the possible effects of increasing salinity on them, have been described. Baseline data are therefore necessary if we are to understand what is happening in this wetland. If macroinvertebrate assem¬ blages are to be used as indicators for monitoring restorative change, then ideally sampling should be simple to perform (perhaps by relatively untrained people), and yield results upon which rehabilitation end¬ points can be determined. The aim of this paper is to provide a preliminary assessment of macroinvertebrate data for Lake Toolibin and adjacent wetlands as the basis for monitoring a rehabilitation programme. 33 Journal of the Royal Society of Western Australia, 78(2), June 1995 Lake Toolibin Case Studies The Lake Toolibin district had a higher than average rainfall (404 mm cf 358 mm) in 1992. In late September of that year, students and staff from Edith Cowan Uni¬ versity (Joondalup) visited six sites, covering each of three major water bodies and the major habitats in the Toolibin district (Fig 1). Water conductivity (in mS cm'1) was measured at each site using a Wissenschaftlich-Technische Werkstatten conductivity meter. The measurements were converted to parts per thousand (ppt) by multiply¬ ing the value by 0.6, following Williams (1966), in order to provide comparative data to Halse (1987). At each sampling site, recognizably different micro¬ habitats were sampled for macroinvertebrates using a standard Freshwater Biological Association net of 500p mesh size. The sampling technique involved vigorously sweeping the water column from water surface to sedi¬ ments over an area of approximately one square metre. Material collected was then live-picked, with representa¬ tives of morphospecies being preserved in 70% alcohol for later identification in the laboratory. A drift net (DN) with an opening of 300 mm x 330 mm and mesh size of 500 pm was set for 24 hours in the Arthur River (up¬ stream from site 1) to collect macroinvertebrates. Mate¬ rial collected (approximately 2 litres of invertebrates) was mixed and halved to reduce sample size, then pre¬ served in 5% formalin and subsequently sorted in the laboratory. Zooplankton was sampled at each site by towing a 63 pm mesh size zooplankton net for 5 m in approximately 0.3 m of water near the shore. Material collected was preserved in 5% formalin and subse¬ quently sorted in the laboratory. All collections of inver¬ tebrates were identified to species level wherever pos¬ sible, with adults and larvae being treated as separate taxa. I km Figure 1. Location of study sites in the Lake Toolibin district showing major roads and principle drainage lines. To examine seasonal changes and the effects of evapo-concentration, the sites were revisited by RGD in mid April 1993. Only Lake Walbyring (site 6) was resampled as before, as the other sites were dry. Table 1 summarizes the salinity, species richness and the number of species found exclusively at each site ( i.e . site-specific species). During the September sampling. Lake Toolibin and the three upstream sites had a salinity of 2.4-4.4 ppt, which is about one-tenth that of seawater. At this time, the salinity at Lake Walbyring was less than half the concentration found at other sites, and was the only site with "fresh" waters (where fresh might be regarded as <1.6 ppt after converting the salinity bound¬ aries provided by Semeniuk, 1987). Whereas other sites became dry and salt encrusted over summer, Lake Walbyring in April 1993 still had a salinity below that of Lake Toolibin when it was full in September 1992. Table 1. Salinity and macroinvertebrate species richness at Toolibin district lake sites. Site Location Salinity mS cm'1 (ppt) Species Richness Site-Specific Species 1 Arthur River 5.8 (3.5) 33 8 DN Drift net (see above) 29 6 2 Lake Toolibin (Western side) 5.0 (3.0) 25 2 3 Lake Toolibin (Eastern side) 4.0 (2.4) 26 3 4 Lake Dulbinning 5.8 (3.5) 31 6 5 Shallow marsh area 7.3 (4.4) 20 3 6 Lake Walbyring 1.7 (1.0) 31 14 6a Lake Walbyring 2.6 (1.6) 27 0 a indicates April 1993 data. A total of 90 macroinvertebrate taxa (82 of which were found in the September 1992 samples, and 27 in the 1993 sample) were recorded from the study (Appen¬ dix 1). Most of the major inland aquatic orders of inver¬ tebrates were represented. Some notable absences from the aquatic fauna were the mayflies (Ephemeroptera) and stoneflies (Plecoptera). These groups are among the insect orders thought to be sensitive to salinity levels greater than about 1.0 ppt (Hart et al. 1991). Lake Walbyring contained substantially more site-specific species than any other site (Table 1). Discussion Degradation of Lake Taarblin, immediately down¬ stream of Lake Walbyring, became obvious in the 1930s when the death of wetland trees was thought to be due to the rise in water levels (Anon. 1987). According to Sanders (1991), salinization of Lake Taarblin was thought to have begun in the late 1950s, with Lakes Toolibin, Dulbinning and Walbyring following a similar pattern in the 1960s. Along with this historical pattern is a seasonal variability in salinity with lake waters becom¬ ing more saline as they become shallower (through evapoconcentration), and a tendency for salinities to vary interannually, presumably as a result of annual 34 Journal of the Royal Society of Western Australia, 78(2), June 1995 rainfall variations and the occasional flushing of salt when lakes overflow (Halse 1987). We know that the rainfall received in 1992 was above average and that the lakes of the Toolibin district contained water for longer than normal. Similarly, September sampling reflects wetter conditions and lower salinities compared to other times of the year. For both these reasons, our data should represent the “fresher7' end of the val¬ ues collected recently for these wetlands. However, his¬ torical depth and salinity data of Lakes Toolibin and Walbyring collected by the Department of Conservation and Land Management (JAK Lane, unpub. data) show the reverse pattern. Salinities for Lake Toolibin were higher in 1992 than they were in 1983 or 1990, the two other years in which the lake was full of water. Over this period. Lake Walbyring was consistently fresher than Lake Toolibin, and does not seem to have had as strong a trend toward salinization, as we also found. The hydrology and hydrogeology of the Toolibin area is complex; however, Lake Walbyring does appear anomalous, with comparatively fresher waters than the other lakes, when full. Lake Toolibin is thought to be a groundwater recharge area, where waters enter a shal¬ low, saline aquifer beneath the lake. Stokes & Sheridan (1985) believe these waters realize superficial expressions at Lake Walbyring and the severely salt-scalded Lake Taarblin, with Lake Walbyring receiving overflow from Lake Toolibin at over bank-full stage. It is unclear how the waters of Lake Walbyring are being maintained, es¬ pecially in its comparatively freshwater state. Perhaps the lacustrine deposits are creating a “perched" effect, and/or this waterbody receives run-off from part(s) of a catchment that are less salt-affected than those drainage lines which supply other wetlands of the system. The aquatic macroinvertebrate taxa found at Lake Toolibin and adjacent wetlands are predominantly eury- haline forms typical of brackish /slightly saline habitats, and most have wide dispersive powers. The relative dissimilarity of the Lake Walbyring fauna may imply that this wetland has retained fresher elements other¬ wise lost due to the salinization of the Lake Toolibin area. Alternatively, faunal assemblages may be season¬ ally opportunistic with respect to salinities. In wetlands that experience fluctuations in salinity, there may be a succession of species as salinities change. For instance, unpublished data (S Halse, pers. comm.) taken at Lake Walbyring in September 1985 and August 1986 (when salinities were far greater) show a fauna much more tolerant of saline conditions; 28 taxa were found on these two sampling occasions, yet no more than one third of these taxa were found by us at Lake Walbyring. No Ephemeroptera or Plecoptera, were found in any collec¬ tions. As with many biological surveys, our interpretation of the data is based on a “snapshot" sampling regime, and some caution is required. Nevertheless, the magnitude of the differences between the fauna of Lake Walbyring and Lake Toolibin, apparently related to differences in salinity, suggest that a single macroinvertebrate collec¬ tion by relatively untrained, but supervised personnel, can yield valuable baseline data. Any remaining ambiguities, like those discussed above, could be resolved best through the longer term monitoring and detailed analyses of macroinvertebrate assemblages. Monitoring over time would be required to document the full range and sea¬ sonality of species, and determine whether a more regional, less cosmopolitan fauna becomes established as tree-planting and hydrological manipulations have effect. We propose that the faunal characteristics at Lake Walbyring found in September 1992 could be the mini¬ mum required to indicate successful rehabilitation if consistently found at Lake Toolibin; other fauna like mayflies (Ephemeroptera) could also be indicators of wetland recovery from saline effects. Furthermore, the fresher waters of Lake Walbyring are a habitat for po¬ tentially re-colonizing species and should be monitored. The production of these preliminary findings are timely, given that decisions regarding water and salinity management at Lake Toolibin are imminent. Engineer¬ ing solutions, including water diversion measures, may be the only short-term method of saving Lake Toolibin, yet our results indicate that in a wetland chain, one lake cannot be treated in isolation from neighbouring wet¬ lands, and that a greater understanding of catchment conditions at least at Lake Walbyring, is required. Deci¬ sions about rehabilitating Lake Toolibin should not ex¬ clude the value of Lake Walbyring or any other regional wetland which may, at one time or another, harbour an assemblage of freshwater species. Acknowledgements: Permits to conduct the study were provided by the Department of Conservation and Land Management. The following are thanked for their contributions to this paper: the students from Edith Cowan University (Water and Wetlands Management unit, 1992); Darren and Fiona Ryder for their initial sorting of 1992 samples; Mr Kim Richardson for assistance with sampling and with identifications. Drs Mark Harvey, Don Edward and the late Shirley Balia, Mrs Shirley Slack- Smith, Ms Faye Cheal and Ms Sue Harrington identified species. Rainfall information was supplied by the Western Australian Bureau of Meteorol¬ ogy. Mr Jim Lane and Dr Stuart Halse (CALM) allowed us to use and cite unpublished data. Dr Halse also suggested the work be done, and kindly provided a critique of an earlier draft of this manuscript. References Anon. 1987 The status and future of Lake Toolibin as a Wildlife Reserve. A report prepared by the Northern Arthur River Wetlands Rehabilitation Committee. Report Number WS 2, Water Authority of Western Australia. Perth, 26pp. Anon. 1992 Recovery plan for Lake Toolibin and surrounding reserves. A report prepared for the Department of Conserva¬ tion and Land Management, Western Australia, under the Australian National Parks and Wildlife Sendee Endangered Species Program, 1991/92. Department of Conservation and Land Management, Western Australia, 57pp. Bell D T & Froend R H 1990 Mortality and growth of tree species under stress at Lake Toolibin in the Western Australian Wheatbelt. Journal of the Royal Society of Western Australia 72:63-66. De Deckker P 1983 Australian salt lakes: their history, chemistry and biota - a review. Hydrobiologia 105:77-84. Froend R H, Heddle E H, Bell D T & Me Comb A J 1987 Effects of salinity and waterlogging on the vegetation at Lake Toolibin, Western Australia. Australian Journal of Ecology 12:281-298. Halse S A 1987 Probable effect of increasing salinity on the waterbirds of Lake Toolibin. Department of Conservation and Land Management, WA, Technical Report 15, 26 pp. Halse S A 1988 The Last Lake. Landscope 3(4), 17-22. 35 Journal of the Royal Society of Western Australia, 78(2), June 1995 Hart B T, Edwards R, Hortle K, James K, Me Mahon A, Meredith C & Swadling K 1991 A review of the salt sensitivity of the Australian freshwater biota. Hydrobiologia 210:105-144. Sanders A 1991 Oral histories documenting changes to wheatbelt wetlands. Department of Conservation and Land Management, Perth, WA. Occasional Paper 2/91, 47 pp. Semeniuk C A 1987 Wetlands of the Darling System - A geomor- phic approach to habitat classification. Journal of the Royal Society of Western Australia 69:95-1 1 1. Stokes R A & Sheridan R J 1985 Hydrology of Lake Toolibin. Water Authority of Western Australia. Report WH 2, 48 pp. Williams W D 1966 Conductivity and the concentration of total dissolved solids in Australian lakes. Australian Journal of Marine and Freshwater Research 17:169-176. Williams W D, Boulton A J, & Taafe R G 1990 Salinity as a determinant of salt lake fauna: a question of scale. Hydrobiologia 197:257-266. Williams W D, Taafe R G & Boulton A J 1991 Longitudinal distribution of macroinvertebrates in two rivers subject to salinization. Hydrobiologia 210:151-160. 36 Journal of the Royal Society of Western Australia, 78(2), June 1995 Appendix 1. Invertebrate taxa found in the Northern Arthur River wetlands. (For explanation of site codes, see Table 1 and Figure 1). PHYLUM CLASS (ORDER) Family Species 1 DN* 2 Site 3 4 5 6 6* PLATYHELMINTHES Sp X X NEMATODA Sp X X ANNELIDA OLIGOCHAETA Spl X Sp2 X X Sp3 X HIRUDINEA Sp X X X X ARTHROPODA ARACHNIDA (ACARINE) Eylais sp X Limnesia sp X (ARANEAE) Singotypa sp X Tetragnathidae Tetragnatha sp X X CRUSTACEA (NOTOSTRACA) Lqjidurus apus viridis Baird X (CLADOCERA) Chydoridae Pleuroxus sp X X X Daphniidae Ceriodaphnia sp X X X X Daphnia carinata King X X X X X X X Simocephalus sp X X X X Macrothricidae Echinisca sp X X X X X Macrothrix Ibreviseta Smirnov X Moinidae Sp X X X X X (OSTRACODA) Cypridacea Cyprinotus ledioardi McKenzie X X X X X X Diacypris spinosa De Deckker X Mytilocypris lambiguosa De Deckker X X X X X X Mytilocypris sp 2 X X X X Alboa worooa De Deckker X X X X X X Bennelongia sp X X Sarscypridopsis aculeata (Costa) X X X X (CHONCHOSTRACA) Cyzicus sp X (COPEPODA - Calanoida) Centropagidae ICalamoecia sp X X X X X (COPEPODA - Cyclopoida) 7 Microcyclops sp X X X X X (AMPHIPODA) Ceinidae Austrochiltonia sp X X X X X X X (DECAPODA) Palaemonidae " Palaemonetes australis " Dakin X Parastacidae Claw of Cherax albidus Clark X X X INSECTA (ODONATA - Anisoptera Aeshnidae (ODONATA - Zygoptera) Coenagridae Lestidae Hemianax papuensis (Burmeister) Xantlwgrion erythroneurum (Selys) Austrolestes annulosus (Selys) Austrolestes io (Selys) (HEMIPTERA) Corixidae ? Agraptocorixa sp Micronecta spl Micronecta sp2 X X X X X X X X 37 Journal of the Royal Society of Western Australia, 78(2), June 1995 Appendix 1 (continued) PHYLUM CLASS (ORDER) Site Family Species 1 DN* 2 3 4 5 6 6r1 Notonectidae Sigara sp Sp5 Anisops spl Anisops sp2 Paranisops sp Sp4 (DIPTERA) Chironomidae Ephydridae Ceratopogonidae Culicidae Tabanidae Unidentified Dipteran Pupae Chironomus aff. altemans Walker Chironomus tepperi Skuse Cryptochironomus griseidorsum Kieffer Dicrotendipes conjunctus Walker Kiefferulns intertinctus Skuse Procladius paludicola Skuse Procladius villosimanus Kieffer Sp Sp Anopheles ( Cellia ) sp Culex spl Culex sp2 Sp Spl Sp2 Sp3 (TR1CHOPTERA) Leptoceridae Triplectides australis Navas (COLEOPTERA) Dytiscidae (larvae) Antiporus sp Bidessus spl Bidessus sp2 Homeodytes scutellaris (Germar) Hydaticus spl ?Hydaticus sp2 Hydrovatus sp Laccophilus spl Lancetes lanceolatus (Clark) Macroporus spl Macroporus sp2 Necterosoma sp Paroster sp IRhantaticus spl tRlwntaticus sp2 (adults) Allodesus sp Hydrophilidae (larvae) Australphilus montanus Watts Copelatus sp Laccophilus sp2 Berosus spl (adults) Berosus sp2 Laccobius sp Berosus sp3 Curculionidae (adult) Sp Haliplidae (adult) Haliplus sp Hygrobiidae (adult) Hygrobia australasiae (Clark) Noteridae (adult) Sp (LEPIDOPTERA) Pyralidae (larvae) Sp MOLLUSCA GASTROPODA (PULMONATA) Planorbidae Physastra sp XX X XXX XXX X X X X X X X X X X X X X X X X X X X X X X XX X X X X X X X X X X X X X X X X X X X X XXX X X X X X X X X X X X X X X X XXX X X X X X X X X X X X X X X X X X X X X X X X X Species Richness (by site) 33 29 25 26 31 20 31 27 Total Species Richness (all sites): 90 species Drift Net site 6 sampled in autumn. 38 X X Journal of the Royal Society of Western Australia, 78:39-42, 1995 An Upper Cretaceous chert nodule, apparently marine ballast, from Princess Royal Harbour, Western Australia J E Glover1, R J Davey2 & C E Dortch3 ’Department of Geology and Geophysics, University of Western Australia, Nedlands WA 6907 2Simon Petroleum Technology Ltd Exploration Services, Llandudno, Gwynedd LL30 ISA, United Kingdom 3Anthropology Department, Western Australian Museum, Francis Street, Perth WA 6000 Manuscript received December 1994, accepted March 1995 Abstract A stone nodule recovered in the excavation of a silo foundation in the Port Authority area of Princess Royal Harbour, Western Australia, is composed of Upper Cretaceous chert, a lithological type anomalous to the region. The chert in this nodule closely resembles black English flint, and contains algal microfossils (dinoflagellates) that are more likely to be European than Australian. The specimen is, almost certainly, a nineteenth century ballast stone from northwestern Europe, and may have come from Thames River gravels or similar deposits in southern England. The investigation shows the danger of making uncritical assumptions about the origin and provenance of chert objects in southwestern Australia. Introduction In 1968, Mr Brian Ayre of Rockingham, W.A., picked a nodule of flint (generally classified by geologists as chert) from the lower part of a six metre-deep silo foun¬ dation in the Port Authority area of Princess Royal Harbour, at Albany (35° 03' S 117° 54' E) on the southern coast of Western Australia (see Fig 1). He observed that the foundation walls consisted of layers of beach sand, marine shell grit, a 60 cm-thick zone of black mud, and two or more layers of whole marine shells (pers. comm.). The stone came from the disturbed floor of the foundation. Mr Ayre passed the specimen to the Western Australian Museum in 1993 for identification. On the basis of superficial examination, the stone was tentatively identified as a naturally flaked chert or flint nodule from England or elsewhere in northwestern Europe. The occurrence of a putative ballast stone at a depth of about six metres can be explained by the massive dredging and earth¬ filling operations done in the Albany Port Authority area around the turn of the century (Garden 1978), several decades after sailing ships first could have dumped loads of English or European chert ballast in Princess Royal Harbour. Although we have not been able to verify the dumping of flint ballast in Princess Royal Harbour, it is reasonable to assume that the practice took place during the half century or more (ending about 1900) when sailing ships arriving from England were loading cargos (Garden 1978). It is also possible that the nodule came from a nearby wreck. © Royal Society of Western Australia 1995 Figure 1. Map of southwestern Australia showing the distribu¬ tion of the Plantagenet Group, and localities mentioned in the text. 39 Journal of the Royal Society of Western Australia, 78(2), June 1995 Results General description The object is irregularly shaped and measured roughly 11.5 X 7 X 6 cm (Fig 2) and weighed 461.8 gm before the removal of cores in the laboratory. It is a flint (or chert) nodule that is randomly flaked and rolled on several parts. The angles between some of the large flake scars and adjacent striking platforms are much more obtuse than normally associated with, or even possible by, knapping (i.e. artificial flaking), but are like those seen on pebbles flaked by wave action. Consistent with this is the rolled condition of the piece, i.e, it has heavy abrasion and multiple chipping (tiny negative flake scars) extending along the ridges formed between the various flake scars. The flaking appears to have been done by wave action in which the nodule would have been flung and dragged against other nodules, as on a pebble beach ( cf . Shackley 1974). Figure 2. The Princess Royal Harbour chert nodule. Note the smooth, slightly patinated surface (bottom centre and bottom left) and the shiny black conchoidal fracture (right and top centre). Length of specimen = 11.5 cm. Petrology The surface of the nodule is mostly rounded, but where flaked it has well-developed conchoidal fracture. The fresh rock is generally shiny black (Nl) but in places it is irregularly mottled and ranges to medium light grey (N6) and greyish orange (10YR 7/4). Locally, there is a surface patination about 0.75 mm thick ranging from greyish orange (10 YR 7/4) to pale yellowish brown (10 YR 6/2) (colours based on Rock-Color Chart Committee, 1963). There is no effervescence in cold dilute hydro¬ chloric acid, indicating the absence of calcite. For determinative purposes, two solid cylindrical cores, one 2 x 1.5 cm and one 2 x 2.5 cm, were drilled from the specimen. The chert fractured easily and irregularly, and the objective of retaining the slightly patinated tops of the cores to restore the original appearance was only partly realised. Portions of the cores were used for thin sections, and for the extraction of palynomorphs. The rock is composed mainly of cryptocrystalline quartz (Fig 3). In thin section, there are various dark grey-brown patches ranging from partly to highly silicified argillaceous material, which are up to three mm long. Figure 3. Typical thin-section field of Princess Royal Harbour chert nodule under crossed polarisers. The central chalcedonic mass, 0.25 mm long, may represent an in-filled fossil cavity. Re¬ mainder of field is cryptocrystalline silica. The argillaceous areas grade fairly sharply into sur¬ rounding chert. There are also rare minute granules of hematite, possibly from the oxidation of pyrite. Fossils are abundant. Most of the fossils are siliceous palimpsests, indeterminate except at broad taxonomic levels; they consist of foraminiferal tests, curved elongate shards that were probably shell fragments, and other debris. Spicules are abundant and may have been siliceous originally. There are some black carbonaceous flakes. The chambers of fossils are filled with cryp¬ tocrystalline quartz, radiating chalcedony, or argillaceous matter. Test walls are commonly converted to cross-fibre silica, apparently chalcedony. A remarkable feature is the presence of sparse but excellently preserved uncompressed dino flagellate cysts that are clearly visible in thin section. The rock is evidently a secondary chert originally consisting of a foraminiferal and spicular sediment. The excellent preservation of the dinoflagellate cysts is con¬ sistent with early post-depositional solidification caused by silicification at shallow depth. The argillaceous patches may represent intraclasts but, if so, other clasts are conspicuously absent. Alternatively, the patches may represent areas disturbed by animals burrowing in poorly consolidated calcareous ooze. Palynology After digestion of the silica from the core sample to concentrate the organic debris, a sparse but diverse and well-preserved assemblage of dinoflagellate cysts was recovered. The identified taxa are listed in Table 1. Provenance and age Dr N G Marshall and Dr A N Bint, who initially examined the palynoflora, concluded (pers. comm.) that it was of Late Cretaceous age and not typically Australian. They could not rule out a Northern Hemisphere origin. Detailed examination of the material has confirmed and extended these observations of Marshall and Bint. The most abundant taxon is Spiniferites ramosus, a long-rang¬ ing dinoflagellate cyst. The dinoflagellate association is completely consistent with a provenance in western Europe, including the United Kingdom. It should be added that 40 Journal of the Royal Society of Western Australia, 78(2), June 1995 Table 1 Abundance of taxa in the palynoflora of the chert object from Princess Royal Harbour Taxon Proportion of palynoflora Marine Microplankton Acanthaulax wilsonii Yun 1981 6% acanthomorph acritarchs 1% lActinotheca aphroditae Cookson & Eisenack 1980 3% Apteodinium deflandrei (Clarke & Verdier 1967) Lucas-Clark 1987 Present Atopodinium perforatum (Clarke & Verdier 1967) Masure 1991 1% Cleistosphaeridium spp 2% Coronifera oceanica Cookson & Eisenack 1958 Present Coronifera striolata (Deflandre 1937) Stover & Evitt 1978 Present ? Cribroperidinium sp Present Dinopterygium cladoides Deflandre 1935 1% Elytrocysta circulata (Clarke & Verdier 1967) Stover & Helby 1987 4% Endoscrinium campanula (Gocht 1959) Vozzhennikova 1967 1% Exochosphaeridium spp 2% Florentinia buspina (Davey & Verdier 1976) Duxbury 1980 ” 4% Florentinia ferox (Deflandre 1937) Duxbury 1980 1% Florentinia tenera (Davey & Verdier 1976) Duxbury 1980 ” 1% Heterosphaeridium sp 3% Flistiocysta palla Davey 1969 1% Hystrichodinium pulchrum Deflandre 1935 2% Hystrichosphaeridium recurvatum (White 1842) Lejeune-Carpentier 1940 1% Isabelidinium cooksoniae (Alberti 1959) Lentin & Williams 1977 1% Laciniadinium sp 1% Odontochitina costata Alberti 1961 Present Oligosphaeridium complex (White 1842) Davey & Williams 1966 1% Palaeohystrichophora mfusorioides Deflandre 1935 1% Palambages sp Present Pterodinium cingulatum (O Wetzel 1933) Below 1981 1% Sentusidinium sp 6% Spinidinium echinoideum (Cookson & Eisenack 1960) Lentin & Williams 1976 1% Spiniferites ramosus group (Ehrenberg 1838) Mantell 1854 41% Surculosphaeridium longifurcatum (Firtion 1952) Davey el al. 1966 1% Tanyosphaeridium sp 1% Trithyrodinium spp 1% Valensiella reticulata (Davey 1969) Courtinat 1989 1% Valensiella sp 4% Xenascus ceratioides (Deflandre 1937) Lentin & Williams 1973 1% Terrigenous pollen grains angiosperm pollen indeterminate 1% bisaccate pollen undifferentiated 1% palynological species were widely distributed during the Late Cretaceous (Costa & Davey 1992), and most of those listed here have also been recorded from Western Australia (Marshall 1975, Helby et al 198 7), which had comparable latitudes with western Europe during the epoch. However, there are several forms, including Balteocysta perforata and species of Conosphaeridium which occurred in the Australian region at the time but have not been observed in this assemblage. The compo¬ sition of the palynoflora is therefore more European than Australian, but species variation in the Chalk Sea assemblages is too slight to show where the chert is likely to come from in western Europe. The flint is Late Cretaceous in age, probably late Turonian to Coniacian, on the basis of significant occur¬ rences of Acanthaulax wilsonii (Fig 4), Florentinia buspina, F. tenera, Surculosphaeridium longifurcatum and Isabelidinium cooksoniae (single specimen). Figure 4. The resistant resting cyst of the dinoflagellate Acanthaulax wilsonii, which composes 6% of the total palynoflora. Width of field 0.23 mm. Discussion Many chert objects, overwhelmingly of Aboriginal origin, have been found on or near the surface in south¬ western Australia (Glover 1984). Most are flakes, but a few larger artifacts have been found: for example the Broke Inlet biface from about 150 km west of Albany (Glover et al. 1993), weighed 1797 grams. The artifacts are composed of both non-fossiliferous and fossiliferous chert. Non-fossiliferous artifacts were derived from the Proterozoic Coomberdale Chert or other Precambrian units. Fossiliferous chert, on the other hand, came either from silicified Eocene rocks along west coast areas now submerged by the sea, or from silicified Eocene Plantagenet Group rocks cropping out in the south coastal region, or from silicified Plantagenet rocks now submerged off the south coast. The estimated age of the silicified rock ranges from Early to Late Eocene. No potential source of chert containing Late Cretaceous fossils in southwestern Australia is known to us. The significance of the fossils in secondary chert should be considered briefly. Fossils give the age of the original sediment, which may be substantially greater 41 Journal of the Royal Society of Western Australia, 78(2), June 1995 than that of the chert. McGowran (1989) stated that a world-wide Eocene process of silicification reached its peak in early Middle Eocene. Fortunately, the time of silicification in some cherts can be inferred from their texture. Delicate uncrushed palynomorphs in the Dunsborough biface (Glover et ah 1978) indicate early solidification of the rock by silicification (B Balme, pers. comm.). Similarly, the presence of uncompressed Late Cretaceous dinoflagellate cysts in the Upper Cretaceous chert of western Europe is consistent with a silicification depth of 5-10 metres or less, according to Clayton (1986). The excellent preservation of dinoflagellate cysts in the Princess Royal Harbour object accords with shallow silicification. In each of these examples, therefore, the fossils represent the age of the chert as well as the age of the original sediment. Not all chert objects found in southwestern Australia are necessarily local. McCarthy (1958) implies that flint implements have more than once been dumped with ballast in Australia, and Tindale has stated that flint tools have been recovered from "Thames gravel" ballast at Port Lincoln, South Australia (see Dortch & Glover, 1983, p. 330). English flint pebbles may have been used in the processing of gold ore in the Kalgoorlie area during the 1890s according to Hutchinson (pers. comm., see Dortch & Glover 1983). Charleton (1903) refers to the use of "quartz pebbles from Norway" in the Hannans Star Mill, but the term may have been a misnomer for chert. Finally, an implement of chert containing fossils of indeterminate age found near Scaddan, Western Aus¬ tralia (400 km ENE of Albany) is considered to be an Acheulian ( i.e . Lower Palaeolithic) biface brought by ship from England, either as a collector's item or ballast (Dortch & Glover 1983). There were thus several ways in which exotic siliceous stone could have entered Western Australia. What is a likely Northern Hemisphere source for the nodule from Princess Royal Harbour? According to Scott (1993), many Quaternary and Recent sand and gravel deposits in southern England are characterised by more than 80% flint of Late Cretaceous age. The chances of flint ballast in Australia having been derived from those areas are obviously significant. The Princess Royal Harbour stone resembles black flint from western Europe. The macroscopic appearance. Late Cretaceous age and well-preserved palynomorphs, combined with the heavily rolled condition, suggest that the Princess Royal Harbour object is a naturally flaked nodule deriving from a pebble or shingle beach (or other gravel deposit) in western Europe, and more specifically from southern England. It is probably a 19th century bal¬ last stone that was inadvertently deeply buried within the highly disturbed, made ground in the Port Authority area. It adds to evidence that untested assumptions of Eocene age for fossiliferous chert objects in southwest¬ ern Australia should not be accepted uncritically. Acknowledgements. Drs N G Marshall and A N Bint, of Woodside Off¬ shore Petroleum Pty Ltd made a preliminary examination and interpreta¬ tion of the palynological material, and suggested that it be sent to one of us (RJD). Helpful discussions were had with Dr B E Balme of the Depart¬ ment of Geology and Geophysics, The University of Western Australia. R J Davey publishes with the approval of the Directors of Simon Petroleum Technology Ltd. References Charleton A G 1903 Gold Mining and Milling in Western Aus¬ tralia. E & F N Spon, London. Clayton C J 1986 The chemical environment of flint formation in Upper Cretaceous chalks. In: The Scientific Study of Flint and Chert (eds G de G Sieveking & M B Hart). Cambridge University Press, Cambridge, 43-54. Costa L I & Davey R J 1992 Dinoflagellate cysts of the Cretaceous System. In: A Stratigraphic Index of Dinoflagellate Cysts (ed A J Powell). Chapman & Hall, London, 99-154. Dortch C E & Glover J E 1983 The Scaddan implement, a re-analysis of a possible Acheulian handaxe found in Western Australia. Records of the Western Australian Museum 10:318- 334. Garden D S 1978 Southern Haven. Port of Albany, Albany. Glover J E 1984 The geological sources of stone for artifacts in the Perth Basin and nearby areas. Australian Aboriginal Studies 1984:17-25. Glover J E, Bint A N & Dortch C E 1993 Typology, petrology, and palynology of the Broke Inlet Biface, a large flaked chert arti¬ fact from south-western Australia. Journal of the Royal Soci¬ ety of Western Australia 76:41-47. Glover J E, Dortch C E & Balme B E 1978 The Dunsborough implement: an Aboriginal biface from southwestern Austra¬ lia. Journal of the Royal Society of Western Australia 60:41-47. Helby R, Morgan R A & Partridge A D 1987 A palynological zonation of the Australian Mesozoic. In: Studies in Australian Mesozoic Palynology (ed P A Jell). Association of Australian Palaeontologists Memoir 4:1-79. Marshall N G 1975 Late Cretaceous Dinoflagellates from the Perth Basin, Western Australia. Ph D Thesis, University of Western Australia. McCarthy F D 1958 Culture succession in south eastern Austra¬ lia. Mankind 5:177-190. McGowran B 1989 Silica burp in the Eocene ocean. Geology 17:857-860. Rock-Color Chart Committee 1963 Rock-Color Chart. Geological Society of America, New York. Scott P W 1993 Distribution of flint and chert in Quaternary and Recent gravel aggregates in the United Kingdom. Transac¬ tions of the Institution of Mining and Metallurgy, Section B, Applied Earth Science 192:B1-B4. Shackley M L 1974 Stream abrasion of flint implements. Nature 248:501-502. 42 Journal of the Royal Society of Western Australia, 78:43-54, 1995 Freshwater biogenic tufa dams in Madang Province, Papua New Guinea WF Humphreys1, SM Awramik2 & MHP Jebb3 terrestrial Invertebrate Zoology, Western Australian Museum, Francis Street, Perth, W A 6000 department of Geological Sciences, Preston Cloud Research Laboratory, University of California, Santa Barbara, CA 93106 USA 3The Christensen Research Institute, PO Box 105, Madang, Papua New Guinea; present address: Department of Botany, Trinity College, Dublin 2 EIRE Manuscript received January 1995; accepted March 1995 Abstract A large, fresh water calcareous tufa deposit occurs on a minor tributary of the lower Gogol River in Madang Province, Papua New Guinea. The tufa, apparently unique in the area, occurs just downstream of a natural river-tunnel in limestone. The river has a constant flow and is fringed by lowland rainforest. The river runs over a series of terraces formed by tufa dams and is unvegetated by macrophytes. The tufa is complex, consisting of porous, unlaminated calcite, laminated and banded stromatolites, and calcite encrusted insect tubes. Surfaces of the tufa have an epilithic community of insects, green algae, and microbes (rich in cyanobacteria and diatoms) that contributed to tufa formation. The lips of the tufa dams have a greater concentration of insect tubes, stromatolites, and filamentous green algae than elsewhere in the tufa. The associated fauna consists of caddis fly larvae (Trichoptera), midges (Chironomidae) and aquatic lepidopteran larvae (Pyralidae). The deposit is a hybrid of real stromatolites and classical tufa. This unusual equatorial tufa deposit indicates that tube-building insects can play a major role in tufa deposit construction. Introduction A series of calcareous tufa dams occur downstream of a river cave in lowland rainforest in equatorial north¬ ern Papua New Guinea (Fig 1). The formation is notable because it is apparently unique within a large area and because close inspection in the field shows that it is intimately associated, inter alia, with a rich arthropod fauna and stromatolites. This paper describes the context, structure and components of the formation. Tufa, a spongy, porous, terrestrial, semifriable variety of travertine (Bates & Jackson 1987), is found forming dams and terraces in a freshwater stream in Madang Province, Papua New Guinea. The formation of tufa involves the localized precipitation of calcium carbonate due to processes that range from abiogenic precipitation caused by C02 degassing to biotic factors such as photo¬ synthesis that also remove C02 (see Pentecost 1981). Tufa often occurs at the mouths of springs, along streams, and on the shores of lakes (Scoffin 1987, Bates & Jackson 1987). The term 'tufa' apparently has its origin from Pliny (23 to 79 AD) who used the term 'tophus' ( thophus ) for encrustations on plant remains and porous volcanic rocks (Julia 1983, Ford 1989). The term traver¬ tine is often reserved for a hard, dense variety of terres¬ trial calcium carbonate deposit (Bates & Jackson 1987). According to Ford (1989), tufa is more or less synony¬ mous with travertine and he prefers tufa. Julia (1983), on the other hand, uses travertine for all freshwater calcium © Royal Society of Western Australia 1995 carbonate encrustations. Chafetz & Folk (1984) use trav¬ ertine as a general term for all freshwater carbonates around springs while Riding (1991) prefers that traver¬ tine be restricted to warm spring carbonate deposits. We follow the usage of Scoffin (1987) and Bates & Jackson (1987). Most tufas that have been studied are from highly seasonal environments, mainly in the northern temperate and continental regions (Ford 1989). Tufas from low latitude regions, however, are locally abundant (e.g. Dunkerley 1981) and may form in seasonally arid settings like those from the Napier Range, Western Australia (Viles & Goudie 1990). The tufa in Madang Province, Papua New Guinea, is a low latitude deposit that forms dams and terraces along a river draining a karst region (Fig 2). A complex community of aquatic insects, algae, and cyanobacteria is associated with the tufa. Unlike in many other riverine tufa deposits (e.g. Pentecost 1987), mosses and macro¬ phytes are not evident. The internal structure of the tufa is variable and ranges from porous, unlaminated carbon¬ ate, to complex associations of calcified insect tubes and small stromatolites (with moulds of filamentous microbes). The calcified insect tubes are common in much of the tufa. The association of the complex biota with the tufa, the calcified insect tubes, the stromatolites, the preserved moulds of filamentous microbes, and the organic-rich nature of the tufa suggest that the biota play a major role in the formation of the tufa. The Madang tufa is important because tufas from tropical environments have not been studied in detail; it also indicates that 43 Journal of the Royal Society of Western Australia, 78(2), June 1995 Figure 1. Location map of the Og tufa in Papua New Guinea. tube-building insects can play a major role in tufa formation and that well-developed stromatolites occur in tufa and contributed to the carbonate buildup. Description of the area Location The tufa deposit occurs in a gorge of a minor lower tributary of the Gogol River (Fig 1), downstream of the exit of a ca. 300 m long natural river-cave (the Og Cave) in the Wandokai Limestone; the river runs briefly through this natural tunnel, the remainder of its course being over ground. The locality is ca. 0.8 km northwest from Sein village, Madang Province, Papua New Guinea (5°18'S 145°43'E). Climate The climate is seasonal. The area comes under the influence of the north-west monsoon from December through March and the south-east trade winds for the rest of the year. The mean annual rainfall of ca. 3300 mm masks the high variability (2000-4500 mm over the last 40 years) that results from the strong impact of the El Nino Southern Oscillation (ENSO) on this coast. The high rainfall, together with a mean monthly temperature of between 23 and 33°C support a lowland humid forest on infrequently drying soils overlain by volcanic ash (Bleeker 1983). Geology The Wandokai Limestone is of Pleistocene-Holocene age and forms the bedrock of the region. This limestone is a massive and crudely bedded biocalcarenite, calcarenite, calcilutite, and calcareous mudstone with subordinate lithic arenite, conglomerate, and clay (Robinson et al 1976). Although there is noticeable surface drainage, the area is a cavernous karst with predomi¬ nantly underground drainage. Methods The tufa deposit was mapped (Fig 3) to Grade 5-3 using standard speleological methods (Ellis 1976) em¬ ploying a tape measure, compass, and inclinometer (Suunto). Samples of the deposit and water were collected for laboratory analyses. Some tufa pieces were fixed in neutral formalin in a local laboratory for later microbio¬ logical studies. Samples of the fauna from a number of distinct habitats within the tufa deposit (Fig 4) were taken for identification. Organisms were also extracted from the collected tufa samples. Sweep net samples were taken during the day from the vegetation fringing the pools and collections made at night on the tufa using a black light. In situ oxygen concentration was measured using a Hanna Instruments HI 8543 portable dissolved oxygen meter. In situ hydrogen ion concentration was measured using a Beckman 031 pH meter, calibrated before and after use with Labchem calibration solutions. Water temperature was measured in situ using the pH meter calibrated against a certified thermometer. All other measures of water chemistry were made on samples collected (both filtered and unfiltered) in triple acid- washed bottles and kept frozen for later analysis. Ca2+, Mg2+, K+, and Na+ were analyzed by atomic absorption spectrophotometry (Varian/Spectra 30/40) and other analyses followed Strickland & Parsons (1972). X-ray powder diffraction analysis of one tufa sample (SBO305) was performed using a Philips Powder Diffractometer (12045/P3) and Random Technology controller (DFC-331 B). Microorganisms were examined as wet mounts prepared from tufa preserved in neutral formalin. The outer surface of the tufa was scraped off and the calcite dissolved in dilute HC1. The insoluble materials were then washed with deionized water and a wet mount prepared. Samples were examined using a Zeiss Photoscope II under white light, phase contrast, and Nomarski Interference Contrast. Freshly fractured surfaces were examined under a Nikon SMZ-10 stereomi¬ croscope to determine the extent to which larger microbes may occur below the surface. Tufa samples were cut with a rock saw and freshly cut surfaces were examined by eye, hand lens, and a Nikon stereomicroscope. Thick (ca. 45 pm) petrological 44 Journal of the Royal Society of Western Australia, 78(2), June 1995 Figure 2. Parts of the Og tufa. a. general view of lower left section of Fig 3, b. 'riffle' flow over gently sloping tufa denoted by the broad tufa areas in Fig 3, c. detail of "a" showing overgrowth of lip, d. scalloped area of Fig 5 in situ with water partly diverted by arm (upper left). Figure 3. Plan and extended section of the tufa deposit below Og Cave with the apparently active tufa faces shown in black. The stream emerges from Og Cave (1) and spreads through a series of pools formed by tufa dams (3-8), before the confluence with another stream below the final dam (8). Some tufa dams have a narrow sill and a nearly vertical downstream face from which the water falls (3, 5-8; see Fig 2c), while others have a broad sloping sill over which the water has a riffle flow (4, D see Fig 2b). Main dam (5); B marks the site of the sill referred to in the text; dam of the upper pool (3); D denotes sites from which tufa samples were collected; E marks the approximate site of the fracture in the lower tufa dam caused by an earthquake in 1972. 45 Journal of the Royal Society of Western Australia, 78(2), June 1995 Figure 4. The location of the tufa samples: diagrammatic down¬ stream profiles of three tufa dams indicating the relative positions from which the tufa samples were cut (numbered blocks). Left, gentle slope as in Fig 2b; middle, lip overgrowth as in Fig 2c. sections, 151 x 176 mm in size, were prepared from six different tufa samples. These were examined under a Zeiss Photomicroscope II and Nikon SMZ stereomicro¬ scope, with white and polarized light. Morphological parameters of tine tufa and its components {e.g. lamina thickness, size of microbes, and insect tube diameter) were analyzed using Optimas image analysis software (Bioscan). Statistics were performed using Microsoft Excel and SPSS for Windows. Description of the tufa deposit Overview Owing to its proximity to Og Cave, we have named the deposit the Og tufa. The tufa is apparently unique within the extensive area known to the people of Sein (>10,000 km2) although superficial deposits of tufa, lacking dams or insect tubes, are found coating rocks associated with seepages entering the Sein River from its west bank, as well as downstream of the Og confluence. The Og tufa occurs in a gorge about 20-30 m deep and 7-40 m wide, that is fringed by lowland tropical vegetation. Hence, the solar radiation incident on the tufa deposit is quite restricted, especially close to the cave. The tufa commences within 20 m of the outflow from Og Cave and extends for approximately 60 m downstream to a confluence with another tributary. There is no visible tufa deposit at the tunnel outflow but it may be obscured by a recent rockfall. The river descends 12.5 m through a series of tufa dams over a horizontal distance of approxi¬ mately 60 m (Fig 3). The volume of tufa represented in Fig 3 is estimated to be in the order of 2000 m3, assuming a constant gradient in the underlying rock. We have not measured the rate of flow for the waters in the study area. Og Cave contains large bat colonies (thousands of bats belonging to at least several species; T Reardon, pers. comm. 1994), leeches (V M van der Lande, pers. comm. 1994), amblypygids, and millipedes. The down¬ stream water sometimes smells of bat guano. A fracture associated with an earthquake in 1972 split the tufa dam of the main pool near the left bank (Fig 3, site E). This caused the pool to drain and it is the only time known to the people of Sein, who use the pools daily, that water has not flowed over the face of the tufa dams. While the fissure is still visible to within 15 cm of the next drop-off, it has since been filled naturally with gravel; there has been no regrowth of tufa in the fissure. This lack of overgrowth by tufa suggests a slow rate of tufa formation in this environment. The general mor¬ phology of the deposit and the fissure suggest that the system develops by downstream progression with only minimal vertical accretion at the rims. The river level is reported to be have been constant, even during periods of the ENSO drought. A massive log wedged in the tunnel indicates occasional major flooding. At the time of sampling in May 1990, the mean depth of water flowing over the lip of the main dam was 19±3 mm (n=20) and this depth was reported to be normal by the people of Sein. The Og tufa is highly variable in morphology, especially at a fine scale, and we present our description of the tufa at four observational levels: megastructure, macrostructure, mesostructure, and microstructure (K Grey, S M Awramik, J Bertrand-Sarfati, H J. Hofmann, B R Pratt, MR Walter & Zhu Shixing, unpublished observations) and we focus on those features that stand out at the observa¬ tional level under discussion. Megastructure The large-scale configuration of the tufa deposit, as seen in the field, consists of a series of dams, a few to several metres wide and up to 17 m long, that stand from less than a metre to four metres high above the next, lower tufa /pool complex (Fig 3). The upper pool contains a series of crescent shaped lips and their con¬ stant level and shape suggest that they represent fossil tufa dams, as found in some Australian tufa deposits (R Drysdale, pers. comm. 1994). The lips occur between 8 and 22 m upstream of the retaining dam and are progressively deeper in the pooled water upstream, the successive mean depths (± standard deviation, n) being 22±8(8), 32±3(8), and 50±1(3) cm. The rims of the extant dams are up to ca. 17 m long, are level (±3 mm), and appear to be self-regulating. A riffle-flow, several metres long, forms as the pool shallows toward the lip of the dam (the water was 19±3(8) mm Figure 5. Diagrammatic section through an active face of the tufa formation showing water flow, and detail showing the scalloped edge (cf Fig 2d) and the location of caddis fly tubes and cyanobacteria /algal bands. 46 Journal of the Royal Society of Western Australia, 78(2), June 1995 Figures 6. Photographs of the surface and gross sections of the formation at Sein. a. Surface of tufa upstream of the sill showing small protuberances (similar to the chou-fleur type of Freytet & Piet 1990); sample 2a in Fig 4. Scale bar = 10 mm. b. Surface growth forms as broad papillae on the sill; sample 2b in Fig 4. Scale bar = 10 mm. c. Surface view of the scalloping below the lip of the sill resulting from the alternating layers of cyanobacterial growth and insect tubes (arrows); sample 5 in Fig 4. Scale bar = 20 mm. d. Section of Fig 6b above showing the columnar growth form; sample 2a in Fig 4. Scale bar = 20 mm. e. Section of Fig 6a above showing the columnar growth form and the banding. Scale bar = 10 mm. f. Section of Fig 6c above showing alternating bands of calcareous insect tubes and denser 'algal' layers. Note macroalgae on the surface. Scale bar = 20 mm. 47 Journal of the Royal Society of Western Australia, 78(2), June 1995 deep on the main dam at the time of sampling). Tufa is deposited on the steep downstream side of the dam and, as well as the expected 'microgours', has in some places a scalloped surface (Figs 2d, 5, 6c). Macrostructure The overall structure of the individual components of the tufa is termed the macrostructure. The tufa is primarily massive; nevertheless, there are certain obvious indi¬ vidual components of the tufa such as the banded ap¬ pearance (Fig 6f) that is present even in the absence of insect tubes, fence-like linear arrays of calcified insect tubes (Fig 6c, 6f), stromatolites (small columns, pseudocolumns, and columnar-layered structures; Fig 6d), and nodules or bumps on the surface (Figs 6a, 6b). The surface of the tufa is dominated by bumps or nodules. These often appear as small protuberances, commonly a few millimetres in size but up to 2 cm across and up to 0.8 cm high (Figs 6a, 6b). Many of the nodules, when studied in cut specimens and in thin sections, appear to be the product of the complex growth of columnar-layered and pseud ocolumnar stromatolites, some with a pillared microstructure (Figs 6a, 6e). Up¬ stream of the sill (site B in Fig 3) the surface of the tufa is not regularly sculptured, but below the lip it is often deeply scalloped (Figs 2d, 5, 6c) with indentations on the order of 2.5 cm deep. The scalloping appears to be produced by the alignment of aquatic insect tubes (Figs 5, 6c, 6f, 7a). The general scalloping associated with the leading edge of the sill seems to be influenced by algae/ cyanobacteria, an hypothesis that is supported by the highest density of stromatolites in these samples (Figs 6f, 8d). Mesostructure The mesostructure features (at a scale between mac¬ rostructure and microstructure; see below) include insect tubes, stromatolites and their lamination (and banding), and other small-scale features observable with the eye or hand lens. The tufa is variable from site to site and our limited sampling constrains our ability to discern any but the most superficial patterns. Edges of dams (sills) and other areas of greater water turbulence appear to have the densest occurrence of insect tubes. Sites with denser (less porous) tufa have few large insect tubes but many smaller tubes, and a fabric that is characterized by pillared to arborescent micritic bushes often originating on insect tubes with stromatolites oriented normal to the upper surface (Fig 8c). The best-developed stromatolites (Fig 8d) occur in tufa with dense accumulations of insect tubes (Figs 6f, 8a, 8b); however, thin, somewhat columnar, columnar-layered, and pseudocolumnar stromatolites are common in all samples (terminology from K Grey, S M Awramik, J Bertrand -Sarfati, H J. Hofmann, B R Pratt, MR Walter & Zhu Shixing, unpublished observations). In samples with few insect tubes, nodules appear to be formed by pseudocolumnar tufa and stromatolite for¬ mation (Fig 6d). The insect tubes, produced primarily by caddis fly larvae (Trichoptera), occur in all samples studied and range in size from 0.18 to 3.06 mm in diameter (mean=0.93 mm; median^ 0.64 mm; n=384). They have arbitrarily been grouped into small (=^0.64 mm) and large (>0.64 mm) tubes. Tubes are not uniformly distrib¬ uted with respect to size and density. For example, the sample illustrated in Fig 6f has dense arrays of large tubes with a mean diameter of 1.7 mm (median=1.83 mm; n=109). In another sample (SBO304) which has abundant tubes (ranging in size from 0.19 to 2.41 mm dia; mean=0.56 mm; median=0.44 mm; n=107), large tubes (>0.64 mm in diameter) are rare. Walls of the calci¬ fied tubes, which are up to 1 mm thick, are composed of Figures 7. Gross sections of the formation at Sein. a. Section of Fig 6c showing alternating bands of calcareous insect tubes and denser layers which are stromatolites or microbial carbonates. Scale bar = 10 mm. b. End view of Fig 6f. Scale bar - 20 mm. 48 Journal of the Royal Society of Western Australia, 78(2), June 1995 Figures 8. Photomicrographs of microbes in formalin preserved samples and thin sections of dried tufa. a. Oscillatoriacian cyanobacteria on surface of tufa (from formalin preserved sample). Scale bar ca. 10 pm. b. Filamentous cyanobacteria in laminae of the stromatolites. Scale bar ca 25 pm. c. Portion of stromatolite on calcified insect tube with pillared microstructure or tufts; pillars are separated by voids filled with sparite. Scale bar = 1 mm. d. Microbial carbonate with radiating filamentous microbial fossils (clear tubes) in a micritic matrix, forming small, dense bushes. Scale bar ca. 25 pm. e. Cross section of filamentous microbial fossils (arrows) similar to Fig 8d. Scale bar ca. 25 pm. f. Stromatolite encrusting and bridging calcified insect tubes. Scale bar = 1 mm. g. Coarsely laminated stromatolites. Scale bar = 1 mm. laterally discontinuous micritic layers often organized in a lacework to cellular pattern (Figs 8c, 8f). The inner surface of tubes often has a yellowish-brown stain. Tufa specimens with moderate to abundant calcified insect tubes also appear to have the most stromatolites. In thin section, the laminae of the stromatolites are seen be¬ tween and encrusting insect tubes (Fig 8f). Frequently, a variety of small (<1 cm dia) calcareous tubes is found in tufa ( e.g . Wallner 1935; Fleimann & Sass 1989). Some of these are carbonate precipitated 49 Journal of the Royal Society of Western Australia, 78(2), June 1995 around plants (Pedley 1992) while others are calcified insect tubes (Scholl & Taft 1964). In contemporary deposits, direct observation can be made of the organisms making the tube but in ancient tufa deposits it may be difficult to differentiate between plant and insect tubes. However, in fossilized examples a limited variability in tube diameter would suggest that small plants ( e.g . bryo- phytes) and insects were involved, whereas tubes many centimetres in length would indicate plant stems or roots (root hairs may be preserved). We suggest that the nature of the calcified wall, like the lacework found in the Og tufa insect tubes, may be a feature characteristic of tube-building insects that could be used to identify them in ancient tufas. The tufa contains numerous small (<1 cm dia) pseudocolumnar and columnar-layered stromatolites (columnar forms are rare), less than a centimetre high, that frequently encrust masses of calcified tubes (Fig 8f). The stromatolites appear to be most abundant in tufa with the densest population of tubes and these occur at the sill; however, they are not abundant in tufa with the fence-like arrays of tubes (Fig 6c). The thickness of the stromatolitic laminae varies con¬ siderably. However, there is a clear pattern of predomi¬ nantly light laminae (average 114.8 pm thick) and dark laminae (average 56.4 pm thick). Some areas have a coarsely laminated, banded appearance (Fig 8g) which is a variety of stromatolitic laminae (Fig 8f). The bands are alternating layers of thinner dark micrite and thicker light-colored microspar and differ from stromatolite laminae in their more diffuse boundaries and overall coarseness of the calcite. Unlike banding in many lacustrine stromatolites, these bands are not composed of finer laminations. Microstructure Much of the tufa is structureless and composed of micrite with numerous voids. Thin sections of this tufa appear very dull, blue-green in color under cathode luminescence indicating primary, fresh calcite that has not be influenced by meteoric/diagenetic processes. There are two obvious microstructures, pillared and stromatolitic. Stromatolitic microstructure consists of alternating laminae of thinner dark micrite and thicker light microsparite (see above). The preservation of the material is sufficient to preserve filamentous microbial fossils (presumably cyanobacteria) in the laminae (Fig 8b). An interesting microstructure is a pillared form which is dominant in sample SBO 305. The pillars con¬ sist of small, possibly cylindrical structures (25 to 82 pm dia), some occasionally bifurcate, arranged in a radial manner on large calcified insect tubes, and composed of micrite (Fig 8c). Individual pillars have a fibrous inter¬ nal structure and are composed of remnants of micro¬ bial filaments arranged parallel to the long axis of the pillar. This structure resembles the stromatolitic tufts de¬ scribed by Freytet & Plaziat (1982, Plate 9a on page 113). Most filaments are not well preserved and it is uncertain if the filaments branch. The diameter of filament moulds of clear calcite is approximately 3 pm. Filaments of this size are apparently common in tufa (e.g. Pedley 1987). Biogenic elements in the tufa Cyanobacteria and algae. In formalin preserved speci¬ mens, the surface of the tufa contains a complex micro¬ bial community dominated by filamentous cyanobacteria (Fig 8a) and diatoms. Green algae live on the surface of the tufa, but we have not found compelling evidence of green algae preserved in the tufa. The pillared micro¬ structure shown in Fig 8c contains in longitudinal section what might appear to be the remains of large filaments. However, we have not found circular cross sections of the same diameter that would support this interpretation (see Figs 8d, 8e). While the cyanobacteria appear to dominate numerically, the larger green algae would dominate the photosynthetic biomass. Because of limited sampling of fixed material and the degradation that occurred between the time samples were collected and fixed in the laboratory, we have not attempted to identify the cyanobacteria or algae in more detail. Fauna. An abundant and diverse fauna is sometimes found in and on tufa deposits (e.g. Diirrenfeldt 1978) and the Og tufa is no exception. Faunal samples were taken for identification from a number of distinct habitats within the system. The dominant biomass of these samples is for several species of sedentary caddis fly (Trichoptera), Diptera ('midges') and aquatic lepidopteran larvae (Pyralidae; two species). Some of these species inhabit tubes in the tufa, and spin nets. Adult Trichoptera collected represent eight genera (ca. 14 species; Table 1). Larvae were present in various instars at the one time observed. Adults of the dipteran family Chironomidae represent probably eight species of eight genera (P Cranston, pers. comm.). The lepidopteran family Pyralidae was represented by larvae of probably one genus (Table 1; the subfamily Nymphulinae within Australasia requires revision; J Hawking, pers. comm.). In addition, an errant fauna including flatworms (Turbellaria) and water mites (Hydracharina) was observed. Many of the caddis-fly larvae (Trichoptera; Hydropsychidae: Philopotamidae) represented in the samples mostly strain food (algae, fine organic particles and small invertebrates) from the flowing water by con¬ structing capture nets of silk. Leptoceridae may be large or small particle detritivores with some tendency to herbivory (especially in Triaenodes PT-773), or predatory (St Clair 1994). Larvae of the predatory Leptoceridae construct tubular cases of mineral and plant materials, and the Hydroptilidae, or micro-caddises, are free-living for the first four instars and only construct a purse shaped case in the fifth instar (Neboiss 1991). The Nymphulinae (Lepidoptera) have aquatic larvae which may be case-making, web-spinning or free-living, and feed on algae or aquatic plants, and may live in flat cases formed from pieces of the food plant. The chironomid taxa represented are all rheophilic, with the possible exception of ?Brijop}wenocladius, which belongs to a terrestrial /semi-terrestrial clade. Most are relatively cosmopolitan genera, although Riethia has South American/Australian affinities, and Skusella has Afrotropical/ Australian affinities. Microtendipes is ap¬ parently monotypic and quite uncommon in Australia, and the only adult species that has been named is 50 Journal of the Royal Society of Western Australia, 78(2), June 1995 Table 1 Insect larvae (L) identified from the tufa, and adults (A) collected at a black light situated above the Og tufa. Species Family TRICHOPTERA Cheumatopsyche sp PT-1862 Hydropsychidae (A) Cheumatopsyche sp PT-1863 Hydropsychidae (A) Cheumatopsyche sp PT-1868 Hydropsychidae (A) Cheumatopsyche sp Hydropsychidae (L) Chimarra sp PT-1864 Philopotamidae (A) Chimarra sp nr C. goroka Sykora PT-1866 Phiiopotamidae (A) Chimarra sp nr C. papuana Kimm. PT-1865 Philopotamidae (A) Chimarra sp Philopotamidae (L) Ecnomus sp Ecnomidae (A) Hellyethira sp nov A Wells ( in press ) Hydroptilidae (A) Oecetis sp; 70. buitenzorgensis Ulmer Leptoceridae (A) Orthotrichia sp nov1 Hydroptilidae (A) Tinodes sp; IT. aberrans Kimm Psychomyiidae (A) Triaenodes sp PT-1867 Leptoceridae (A) Triaenodes sp B Leptoceridae (A) ?genus Leptoceridae (L) Triaetiodes sp C (different from PT-1867) Leptoceridae (A) ?genus sp 2 Hydrophilidae (L) DIPTERA: CHIRONOMIDAE genus nr Thienemannimyia Tanypodinae: Pentaneurini Cricotopus cf albitibia Kieffer Orthocladiinae (A) Riethia Istictoptera Chironominae (A) Rheotanytarsus Chironominae (A) Rheocricotopus Orthocladiinae (A) nr Bryophaenocladius Orthocladiinae (A) genus nr Skusella Chironominae (A) Microtendipes sp Chironominae (A) LEPIDOPTERA: PYRALIDAE Potamomusa sp Nymphulinae (L) EPHEMEROPTERA ? genus Caenidae (L) Baetis sp Baetinae (L) 1 to be described by A Neboiss conspecific with the Afrotropical species, although rear¬ ing would be required to confirm this (P Cranston, pers. comm.). Water chemistry The general water chemistry from samples collected above and below the Og tufa are presented in Table 2. The pH of the water increased significantly while flow¬ ing over the formation rising from pH 7.6 to ca. pH 8.0 (E i,4 = 36.7; p= 0.004) but the temperature (26.7±0.33°C; n=10) and oxygen content (mean 7.8 ±0.13 ppm Oz, n= 14) showed no trend, with the latter being ca. 95-99% saturated throughout its passage across the formation. The water is nutrient rich (as reflected by the nitrate levels; Table 2) having received the waste products of a large bat colony inhabiting the tunnel immediately up¬ stream. Table 2 The mean constituents of water flowing over the Og tufa on 24 May 1990 (n=2). mg L'1 Calcium 27.65 Magnesium 2.15 Potassium 0.35 Sodium 2.3 Total iron 0.34 Nitrite-N 0.006 Nitrate-N 0.47 Ammonium-N 0.39 Phosphate-P 0.025 Total-P 0.031 The water chemistry is unremarkable except that the calcium concentration of <30 mg L'1 is very low for a tufa-depositing stream and perhaps provides support for the biotic involvement in the tufa formation. By com¬ parison, fresh groundwater in a tropical karst in Cape Range, Western Australia, has a mean of 82 mg L"1 Ca2+ (sd= 4.9, n= 7). Detailed examination of the water and tufa chemistries is required to determine the relative contribution of physical and biological processes to tufa deposition and nature of the carbonates involved. Microenvironment The water velocity increases as it approaches the lip of the sill (dam) and passes down the face. The distribution of the taxa with respect to the sills in the tufa is shown in Fig 9. The front face is saw-toothed with the lee positions of the formation being occupied by tubes of net-making caddis flies; the growth appears to be by forward progression of the cyanobacteria-rich laminae (Fig 5). The internal structure of the formation varies in a manner apparently related to the nature of the water flow in the area. Sections of the rock show clearly that tufa growth has progressed outwards and that it has changed position (elevation; see fossil sills in Fig 3) so that the deposited band of tufa/calcite is sinuous, similar to turbulent water flow (Figs 6f, 7a, 7b). The rows of net-trapping Trichoptera were mostly collected under the overhangs on the face of the dams, while the bigger, long-jawed caddis flies were more fre¬ quent in the riffles, with tube-living chironomids throughout. These are different microhabitats even within the front face of the falls. The microturbulence of the water may be of considerable importance in deter¬ mining suitable microhabitats for the cyanobacteria, al¬ gae, and the arthropods. Although many of the insects inhabiting the formation are net-builders, they would trap little non-calcareous material from the stream; the acid insoluble fraction of the tufa comprised only 0.8% by weight (range 0.30 to 1.0; n=3); this compares with from 4-9.1% acid insoluble fraction in tufas from Europe and Australia (Viles & Goudie 1990) and a mean of 3.44% (rangel.0-10.55%) in ten accreting deposits from active sites in Great Britain (Pentecost 1993). 51 Journal of the Royal Society of Western Australia, 78(2), June 1995 2 1 l . » 0 - 1 r— - t m - ► '/ / * * * / i i 1 1 g Tufa \ 1 \ Adults at I ■ light trap TRICHOPTERA ?Leptoceridae x . - . x . . X C he umato psyche x . x - . X X Chimarra x . x - . X X Ecnomus sp. X Tinodes X Triae nodes X Ortholrichia X Oeceiis X Hellyethira X Hydrophilidae X CHIRONOMIDAE Cricoiopus x - . - . x X ?Skusella x IThienemannimyia x . . X Microiendipcs Bryophaenocladius x Rheocricotopus x - . - . x . X Rheolany tarsus x . x Riethia X LEPIDOPTERA Potamomusa x --- . X Figure 9. The distribution of the taxa with respect to the sills in the tufa formation. The water is pooled upstream of the sill and then forms riffles over the lip and down the face. The light trap data indicate adults present at the site. Discussion The series of drowned tufa dams in the upper pool (Fig 3; another example is also discussed by Ford 1989) provides evidence consistent with the face of the tufa dams both growing in height and progressing down¬ stream. The sweeping convex downstream faces and the remarkable evenness of the lips of the tufa dams suggest they are self-regulating and that they progress most rapidly where the current is strongest in the centre of the stream. The water is rich in nitrates (Table 2) possibly result¬ ing from bat guano. Marine stromatolites are often pre¬ vented from forming by an increase in the nutrient lev¬ els in the water which promotes growth of algae over the cyanobacteria (Pentecost 1978, 1992) and the same could occur in non-marine environments. However, with the Og Tufa, algae (other than diatoms) and mosses are not sufficiently abundant, despite the high nutrient levels, to dominate tufa surfaces and prevent cyanobacterial and diatom participation in the construc¬ tion of the tufa. It is uncertain what factor(s) prevents luxurious growth of higher photosynthesizers. Tufa dams and terraces are normally discussed in the context of examples from highly seasonal climates in which both diurnal and annual growth layers can be observed in the formations (Ford 1989). Ford's recent review mentions no large tufa deposits from equatorial regions lacking seasonality - the constancy of water flow over the Og tufa was noted above - yet widespread and varied tufa deposits were reported by Hossfeld (1951) to occur in northern Papua New Guinea, including the Madang area. 52 Journal of the Royal Society of Western Australia, 78(2), June 1995 Five methods of tufa deposition have been identified (Ford 1989): 1) chemical reactions in saline or alkaline lakes; 2) the cooling of thermal waters 3) inorganic degassing of C02 (often promoted by microturbulence); 4) indirect biochemical precipitation caused by photo¬ synthetic uptake of C02; and 5) direct metabolic precipi¬ tation of calcium carbonate by organisms such as cyanobacteria, algae, and mosses. The first two factors are unlikely candidates for the genesis of the Og tufa because the conditions are not appropriate. The third factor is unlikely to be the prime agent in the formation of the Og tufa as there is only minor evidence for tufa formation where it might be expected elsewhere in the system, in the absence of biotic associates. The fourth factor is unlikely as living stems and leaves immersed in the water were not covered by tufa either above or below the Og tufa. The presence of cyanobacteria and algae within the tufa and the development of stromato¬ lites within suggest a role of the fifth factor in the formation of the Og tufa, although the effect might be indirect (Pentecost 1978, 1981, 1984, 1985). Our study of the Og tufa suggests that a sixth factor needs to be added to the list of tufa depositing agents, namely the presence of tube-building insects living amongst the stromatolitic layers (Figs 6f, 7a, 7b, 8a, 8b). The tubes provide a stable substrate for benthic microbes (like the cyanobacteria and diatoms) that are adapted to an environment characterized by turbulence and the pre¬ cipitation of calcium carbonate. The Og tufa is an example of a tufa deposit formed by a combination of factors (3, 5, and 6) where local water turbulence, a rich benthic photosynthetic micro¬ bial community, and calcite tube-building invertebrates combine to form tufa. There is a commonly held notion that microbial mats and stromatolites form in environments that are extreme (e.g. Jorgensen & Revsbech 1982). Microbial mats of cyanobacteria and diatoms are very common in a vari¬ ety of environments and form in areas actively grazed and burrowed by macro and micro invertebrates. The accretion of cyanobacteria into stromatolites in an envi¬ ronment seemingly benign - constant water, equatorial climate, buffered, moderate water chemistry - and biotically rich may be puzzling. However, the key ingre¬ dient here is the rapid precipitation of calcite that miner¬ alizes the structure; if mineralization is contemporane¬ ous with microbial growth, stromatolites can form. Acknowledgements We are grateful to the landowners of Sein village for permission to work in the area, for their assistance with the work and their hospitality. For determining the fauna we thank Dr A Neboiss, Museum of Victoria (Trichoptera); Mr J Hawking, Murray Darling Fresh Water Research Centre, Albury (Pyralidae) and Dr P Cranston, Australian National Insect Collection (Chironomidae). Laboratory water analyses were conducted by P L Osborne, Department of Biology, University of Papua New Guinea. Caroline Lawrence assisted in the preparation of the map and Douglas Biford printed photographs of the tufa samples. David Pierce performed the XRD analysis and printed photomicrographs. SMA acknowledges the financial assistance of NASA Grant NAGW-1940 on lacustrine stromatolites. The field work was conducted while WFH was in receipt of a Christensen Research Institute Fellowship and with the approval of the Madang Provincial Council. This is contribution 143 from the Christensen Research Institute, Madang, Papua New Guinea. We thank Russell Drysdale for his pertinent comments on a draft of this paper. References Bates R L & Jackson J A 1987 Glossary of Geology. American Geological Institute, Alexandria, Virginia. Bleeker P 1983 Soils of Papua New Guinea. Commonwealth Sci¬ entific and Industrial Research Organisation and Australian National University Press, Canberra. Chafetz H S & Folk R L 1984 Travertines: Depositional morphol¬ ogy and the bacterially constructed constituents. Journal of Sedimentary Petrology 54:289-316. Dunkerley D L 1981 Australian landform example No. 39: tufa dam. Australian Geographer 15:58-61. Durrenfeldt A 1978 Untersuchungen zur Besiedlungsbiologie von Kalktuff — faunistische, okologische und elektronenmikroskopische Befunde. Archiv fur Hydrobiologie, Supplement 54:1-79. Ellis B M 1976 Cave surveys. In: The science of speleology (eds T D Ford & C H D Cullingford). Academic Press, London, 1-10. Ford T D 1989 Tufa - the whole dam story. Cave Science 16:39- 49. Freytet P & Plaziat J-C 1982 Continental carbonate sedimenta¬ tion and pedogenesis-Late Cretaceous and Early Tertiary of Southern France. Contributions to Sedimentology 12:1-213. Freytet P & Piet A 1990 Les formations stromatolitiques (tufs calcaires) recenles de la region de Toumus (Saone et Loire). Geobios 24:123-139. Heimann A & Sass E 1989 Travertines in the northern Flula Valley, Israel. Sedimentology 36:95-108. Hossfeld P S 1951 Calcareous tufa deposits in northern New Guinea. Transactions of the Royal Society of South Australia 74:108-114. Jorgensen B B & Revsbech N P 1982 Photosynthesis and struc¬ ture of benthic microbial mats: microelectrode and SEM stud¬ ies of four cyanobacterial communities. Limnology and Oceanography 28:1075-1093. Julia R 1983 Travertines. In: Carbonate Depositional Environ¬ ments. American Association of Petroleum Geologists, Memoir 33:64-72. Neboiss A 1991 Trichoptera. In: Insects of Australia: A textbook for students and research workers. 2nd edition. Melbourne University Press, Melbourne, 787-816. Pedley H M 1987 The Flandrian (Quaternary) Caerwys Tufa, North Wales: an ancient barrage tufa deposit. Proceedings of the Yorkshire Geological Society 46:141-152. Pedley M 1992 Freshwater (phytoherm) reefs: the role of biofilms and their bearing on marine reef cementation. Sedimentary Geology 79: 255-274. Pentecost A 1978 Blue-green algae and freshwater carbonate de¬ posits. Proceedings of the Royal Society, Series B 200: 43- 61 . Pentecost A 1981 The tufa deposits of the Malham district. Field Studies 5:365-387. Pentecost A 1984 The growth of Chara globularis and its relation¬ ship to calcium carbonate deposition at Malham Tam. Field Studies 6:53-58. Pentecost A 1985 Association of cyanobacteria with tufa depos¬ its: Identity, enumeration and nature of the sheath material revealed by histochemistry. Geomicrobiological Journal 4: 285-298. Pentecost A 1987 Some observations on the growth rates of mosses associated with tufa and the interpretation of some postglacial bryoliths. Journal of Bryology 14:543-550. Pentecost A 1992 Travertine: Life inside the rock. Biologist 39:161-164. Pentecost A 1993 British travertines: A review. Proceedings of the Geologists Association 104: 23-39. Riding R 1991 Classification of microbial carbonates. In: Calcare¬ ous Algae and Stromatolites (ed R Riding) Springer Verlag, Berlin, 20-51. 53 journal of the Royal Society of Western Australia, 78(2), June 1995 Robinson G P Jacques AC & Brown C M 1976 1:250 000 Geological Series, Explanatory Notes: Madang, Papua New Guinea. Sheet 5B/55-6 International Index. Australian Government Publish¬ ing Service, Canberra. Scholl D W & Taft W H 1964 Algae, contributors to the forma¬ tion of calcareous tufa. Mono Lake, California. Journal of Sedi¬ mentary Petrology 34:309-319. Scoffin T P 1987 An Introduction to Carbonate Sediments and Rocks. Blackie, Glasgow. St Clair R M 1994 Diets of some larval Leptoceridae (Trichoptera) in south-eastern Australia. Australian Journal of Marine and Freshwater Research 45:1023-1032. Strickland J D H & Parsons T R 1972 A practical handbook of sea water analysis. Bulletin 167. Fisheries Research Board of Canada, Ottawa. Viles H A & Goudie A S 1990 Reconnaissance studies of the tufa deposits of the Napier Range, N.W. Australia. Earth Surface Processes and Land forms 15:425-443. Wallner J 1935 Zur weiteren Kenntnis der sog. Chironomidentuffe. Botanische Archiv 37:128-134. 54 Journal of the Royal Society of Western Australia, 78:55-56, 1995 Recent Advances in Science in Western Australia Earth Sciences Paiaeomagnetic dating of dolerite dykes and sills in the Mesoproterozoic Albany Mobile Belt and Neoproterozoic Stirling Range Formation by researchers from the University of Western Australia provides evidence for extensional events between Australia, east Antarctica and Greater India. Three periods of dyke and sill intru¬ sion are distinguished. The first, earliest Cambrian, phase took place during crustal extension between the Yilgarn and east Antarctic era tons and is related to Gondwana assembly. A second phase of mid-Carboniferous dykes and a third, possibly Triassic, phase are related to rifting and extension between Greater India and the Western Australian Shield during early stages in the formation of the Perth Basin. Harris LB & Li Z X 1995 Falaeomagnetic dating and tectonic significance of dolerite intrusions in the Albany Mobile Belt, Western Australia. Earth and Planetary Science Letters 131:143-164. Lithographic strata of the traditional "laterite profile" (i.e. pallid zone, mottled zone, laterite, 'residual' sand) and late Cenozoic valley fills of the "palaeodrainage system" (i.e. hardpan, calcrete, alluvium, colluvium) throughout the Yilgarn Sand land Region have been re¬ interpreted to be mainly desert-aeolian sediment. The traditional approaches used to resolve the origin of these units (e.g. photogeology, geomorphology, pedology, geochemistry') are concluded by the authors to be largely inappropriate and to have resulted in models that incor¬ rectly interpret "laterite profile" units soley as deeply weathered basement rocks. If the regolith is to be correctly interpreted, then the authors suggest that the traditional residual and "locally reworked" views, and the alterna¬ tive aeolian views, need to be investigated and critically appraised from the perspective of sedimentary geology. Glassford D K & Semeniuk V 1995 Desert-aeolian origin of late Cenozoic regolith in arid and semi-arid Southwestern Aus¬ tralia. Palaeogeography, Palaeoclimatology, Palaeoecology 114:131-166. Intracratonic sedimentation started in the onshore northern Perth Basin in the Early Permian and continued until the breakup of Gondwana in the early Cretaceous. While mature source rocks are widespread, reservoirs are abundant, and structures are well timed for hydro¬ carbon entrapment, a critical factor is seal since structures covered by shales that are thinner than the throw of faults may lose their trapping potential. The Allanooka High, Dongara Terrace, Beharra Springs Terrace and Cadda Terrace — with good source rock potential, Early Triassic seals and Late Permian objectives at drillable depths — are thought to offer the best chances for further hydrocarbon discoveries. Crostella A 1995 An evaluation of the hydrocarbon potential of the onshore Perth Basin, Western Australia. Geological Sur¬ vey of Western Australia. Report 43. Potentially economic coal seams in the Permian Irwin River Coal Measures (IRCM) on the Irwin Terrace are described by researchers from the Geological Survey of WA to cover an area of about 170 km2. The IRCM and underlying sedimentary rocks constitute a broadly upward-coarsening glaciomarine alluvial deltaic succes¬ sion. Most of the coal resources are deep, with possibly 5% lying in the current opencut window'. The estimated coal resources are 1000 Mt inferred (Class 1) of the Australian Code. In general, the coal is high in moisture (20-30%), high in volatiles (24-37%), moderate to low in sulphur (0.5%) and moderate to high in ash (7-30%); specific energy is between 16.38 and 21.66 MJ kg'1 on an as-received basis. Le Blanc Smith G & Mory A J 1995 Geology and Permian coal resources of the Irwin Terrace, Perth Basin, Western Australia. Geological Survey of Western Australia. Report 44. Groundwrater in the Mesozoic and Cenozoic succes¬ sion is currently exploited to provide about 40% of scheme w'ater to the Perth region; the remainder comes from surface water catchments. Perth will become more dependent on groundwater as urban development con¬ tinues to expand and the surface catchments become fully utilized. Under present landuse conditions, the maximum projected total and sustainable groundwater from the aquifers in the region would be about 500 106 m3 year1, and may be raised to 600 106 m3 year1 with further urban development and increased groundwater recharge from stormwater catchments. The current abstraction rate of about 300 10* m3 year1 is well within sustainable limits and potential exists for significant additional abstraction. Davidson W A 1995 Hydrogeology and groundwater resources of the Perth region, Western Australia. Geological Survey of Western Australia. Bulletin 142. The stromatolite forms Acaciella australica and Basisphaera irregularis are recorded from the Skates Hill Formation by K Grey of the Geological Survey of WA. These forms support correlation of the unit with the Bitter Springs Formation and other units of Supersequence 1 in the Centralian Superbasin, and indicate a Neoproterozoic age for the Savory Basin succession. Grey K 1995 Neoproterozoic stromatolites from the Skates Hills Formation, Savory basin. Western Australia, and a re¬ view of the distribution of Acaciella australica. Australian Jour¬ nal of Earth Sciences 42:123-132. Life Sciences The survival of adult western long-billed corellas and Major Mitchell cockatoos, studied in the wheatbelt of Western Australia from 1977 to 1983 by researchers from the CSIRO Division of Wildlife and Ecology (Midland), ranged from 81.3% for female Major Mitchell cockatoos (92.9% for males) to 94.2% for male western long-billed corellas (93.2% for females). The survival of immature birds was lower. This is attributed to predation of imma¬ ture birds in locally nomadic flocks; immature survival was only slightly negatively affected by dispersal. The populations of both western long-billed corellas and Major Mitchell cockatoos are predicted to be stable, or slowly increasing, as appears to be the case. Smith G T & Rowley I C R 1995 Survival of adult and nestling western long-billed corellas, Cacatua pastinator , and Major Mitchell cockatoos, C. leadbeateri, in the wheatbelt of Western Australia. Wildlife Research 22:155-162. 55 Journal of the Royal Society of Western Australia, 78(3), September 1995 A collaborative study by researchers from the Swedish University of Agricultural Sciences and Curtin Univer¬ sity of Technology documented the invasion of native sclerophyll woodland vegetation by exotic weed species, after fire, in linear remnants along a highway in south¬ western Australia. The most common weeds were the perennial grasses Eragrostis curuula and Ehrharta calycina. The weeds spread mainly from the road side of the linear sclerophyll remnants, rather than the fence side. Grasses are normally an insignificant component of sclerophyll forest, and their presence after fire increases the fire proneness. The impact of fire was still evident after 7 years. Restrictions on the frequency of burning, and on further narrowing of the linear sclerophyll woodland corridors are advocated. Milberg P & Lamont B B 1995 Fire enhances weed invasion of roadside vegetation in southwestern Australia. Biological Con¬ servation 73:45-49. The relationship between species and genus richness of local Australian ant faunas is examined by A Andersen, of the CS1RO Tropical Ecosystems Research Centre (Winnellie), to test the validity of the hypothesis that higher- taxon categories be used, rather than species, in rapid biodiversity surveys. Although the relationship between species and genus richness was strong within regions for 24 sites distributed throughout Australia, overall it varied substantially and there was only a weak relationship. The relationship was confounded by bio¬ geographic factors and influenced by sampling area and intensity. It is concluded that genus richness of ants is an unreliable substitute for species richness except in limited circumstances, and it is suggested that this con¬ clusion may also apply to other taxa for which a small number of genera can contribute a large number of spe¬ cies. Andersen A N 1995 Measuring more of biodiversity: genus richness as a surrogate for species richness in Australian ant faunas. Biological Conservation 73:39-43. Estimates by researchers from the Department of Conservation and Land Management, Perth, of the sus¬ ceptibility to Phytophthora cinnatnomi of plant species in 63 active disease centres and 17 old centres of Eucalyptus marginata forest, north of the Preston River, indicated that the impact was intermediate (scattered deaths) in 46% of the active disease centres and high (most suscep¬ tible plants dead) in 29% of active centres, compared to a high impact in 65% in old disease centres. Cross tabu¬ lation of species by frequency of death and isolation of P. cinnamomi from plant and soil allowed the classifica¬ tion of the response of plant species to infection. Shearer B L & Dillon M 1995 Susceptibility of plant species in Eucalyptus marginata forest to infection by Phytophthora cinnamomi. Australian Journal of Botany 43:113-134. Researchers from the Department of Conservation and Land Management, Perth, have extended the geo¬ graphic and host range of the canker fungus Cryptodiaporthe melanocraspeda. Large numbers of Bank- sia coccinea were observed dying downward from apical branches, in the south coast region of Western Australia, in 1989; rapid complete death of stands was typical for many diseased stands. Lesions of C. melanocraspeda girdled stems and were concluded to be associated with the death of B. coccinea. In contrast, the fungi Zythiostroma spp, which form lesions that girdle the stems of B. coccinea, and Botryosphaeria ribis, which forms non-girdling lesions, were considered to be only weak pathogens. Shearer B L, Fariman R G & Bathgate J A 1995 Cryptodiaporthe melanocraspeda canker as a threat to Banksia coccinea on the South Coast of Western Australia. Plant Diseas 79:637-641. Note from the Hon Editor: This column helps to link the various disciplines and inform pthers of the broad spectrum of achievements of WA scientists (or others writing about WA). Contributions to "Recent Advances in Science in Western Australia" are welcome, and may include papers that have caught your attention or that you believe may interest other scientists in Western Aus¬ tralia and abroad. They are usually papers in refereed journals, or books, chapters and reviews. Abstracts from conference proceedings will not be accepted. Please sub¬ mit either a reprint of the paper, or a short (2-3 sen¬ tences) summary of a recent paper together with a copy of the authors' names and addresses, to the Hon Editor or a member of the Publications Committee: Dr S D Hopper (Life Sciences), Dr A E Cockbain (Earth Sci¬ ences), and Assoc Prof G Hefter (Physical Sciences). Fi¬ nal choice of articles is at the discretion of the Hon Edi¬ tor. "Letters to the Editor" concerning scientific issues of relevance to this journal are also published, at the dis¬ cretion of the Hon Editor. Please submit a word process¬ ing disk with letters, and suggest potential reviewers or respondents to your letter. P C Withers, Honorary Editor, Journal of the Royal Society of Western Australia. 56 Journal of the Royal Society of Western Australia, 78:57-66, 1995 An early Triassic fossil flora from Culvida Soak, Canning Basin, Western Australia G J Retallack Department of Geological Sciences, University of Oregon, Eugene, Oregon USA 97403 Manuscript received May 1995; accepted October 1995 Abstract New collections of fossil plants and reinterpretation of previous paleobotanical work indicate that the Culvida Sandstone is late Early to early Middle Triassic (late Scythian-Anisian), or some 244-248 Ma in age. Most of the fossils have been identified with well known species, with the exception of Chiropteris whitei sp nov and Pleuromeia dubia (Seward) comb nov. This fossil flora is most similar to that of the Newport and Camden Haven Formations of New South Wales. There are also strong similarities with the fossil flora of the Parsora Beds of the South Rewa Basin, India, and the Burgersdorp Formation of South Africa. The Culvida flora is dominated by Dicroidium zuberi like southeastern Australian floras of humid cool temperate paleolatitudes, but it also contains D. hughesii wrhich dominated monsoonal subtropical floras of India and South Africa. The Culvida flora was thus transitional between these two floristic regions. Despite these regional differences in dominance, early Triassic floras were surprisingly cosmopolitan and low in diversity following the Permian-Triassic life crisis. Introduction The early Triassic was peculiar for world vegetation because of low diversity cosmopolitan floras (Retallack 1995). The homogeneity of early Triassic floras across the great Pangean landmass presents a challenge to modern concepts of uniformitarianism, and is in stark contrast to diverse and provincialized floras of the late Permian and Middle Triassic. Low diversity early Triassic fossil floras have been found in Argentina, New South Wales, central Queensland and Tasmania, all high latitude humid parts of the Gondwana supercontinent (Retallack 1977). The Canning Basin of Western Australia is one of the few regions with Early Triassic fossil floras from the low latitude northern edge of the supercontinent. This paper describes one of these fossil floras that have remained until now poorly known. A newly described collection A collection of fossil plants was made in 1973 by G R Evans and L N Brown of Mines Administration Pty Ltd of Brisbane, and forwarded to me for study by R J Paten. The specimens are now housed under numbers F9143 to 9160 in collections of the Geological Survey of Western Australia. The fossils are impressions in white shale, with irregular areas of red and purple stain. This colour is typical for the mottled zone of a deep lateritic paleosol, and is presumed to be a Cenozoic alteration of a pre¬ existing Triassic plant-bearing shale that may have been grey to brown in color. The fossils were collected from the Culvida Sandstone at Culvida Soak (Fig 1: grid reference 506468 on Cornish 1:250,000 sheet). The matrix of the fossils is most similar to the interval 140-170 m © Royal Society of Western Australia 1995 above the base of the Culvida Sandstone in BMR Cor¬ nish number 2 borehole (Yeates et al. 1975). The following fossil species were identified from this collection and are discussed in more detail at the end of this paper. Pleuromeia dubia (Seward) comb nov Cladophlebis carnei Holmes & Ash 1979 Chiropteris whitei sp nov Dicroidium hughesi (Feistmantel) Lele 1962a Dicroidium narrabeenense Jacob & Jacob 1950 Dicroidium zuberi (Szajnocha) Archanglesky 1968 Umkomasia sp indet Lepidopteris madagascariensis Carpentier 1935 Reinterpretation of previous collections Earlier paleobotanical work by White (1961) and White & Yeates (1976) can now be revised. Updated lists for localities (Fig 1) currently recognized within the Culvida Sandstone (by Yeates et al. 1975) are given below, with commentary following. C08 Equisetales gen et sp indet (White 1961 PI 2, Figs 5,6, PI 3, Fig 1) Dicroidium zuberi (Szajnocha) Archangelsky 1968 (White 1961) Taeniopteris sp indet (White 1961) C062 Dicroidium hughesii (Feistmantel) Lele 1962a (White 1961, PI 3, Fig 6) Dicroidium narrabeenense (Walkom) Jacob & Jacob 1950 (White 1961, PI 3, Figs 3,4B) Pteruchus barrealensis (Frenguelli) Holmes & Ash 1979 (White 1961, PI 3, Fig 4A) Chiropteris whitei sp nov (White 1961, PI 3, Fig 5) 57 Journal of the Royal Society of Western Australia, 78(3), September 1995 CO2076 Dicroidium hughesii (Feistmantel) Lele 1962a (White & Yeates 1976) CO2107 Tomiostrobus sp indet (White & Yeates 1976) Pleuromeia dubia (Seward) comb nov (White & Yeates 1976, PI 13, Figs 47,48) Sphenopteris sp indet (White & Yeates 1976) Umkomasia sp indet (White & Yeates 1976, PI 13, Figs 48,49) CO2108 Equisetales gen et sp indet (White & Yeates 1976) Culvida Dicroidium narrabeenen.se (Walkom) Jacob & Soak Jacob 1950 (White & Yeates PI 13, Fig 46) Taeniopteris sp indet (White & Yeates 1976) White's identification of Dicroidium odontopteroides from these localities is unproven, because it was based on fragments without a fork. These are more likely to be¬ long to Dicroidium zuberi, a leaf taxon thought to have had ovuliferous organs of Umkomasia and pollen organs of Pteruchus barrealensis [the latter listed by White (1961) as " Lycopodites sp"). Several relatively complete specimens of Dicroidium hughesii and Chiropteris whitei in the newly described collection are identical to fragments identified by White (1961), as "Lingui folium" and "Ginkgoites antarctica" respectively. Small wedge shaped sporophylls widely referred to Araucarites have been referred to Tomiostrobus by Sadovnikov (1982). The rounded tips of sporophylls in the cone figured by White & Yeates (1976, PI 13, Fig 47) are similar to those of Pleuromeia dubia (Seward) comb nov. Geological age of the fossil flora The flora of the Culvida Sandstone can be identified confidently as part of the Dicroidium zuberi oppelzone, of late Early to early Middle Triassic age (Scythian- Anisian: Retallack 1977). Seven of its 13 species are known also from the Newport Formation of the Sydney Basin, NSW (Retallack 1980a), five from the Parsora Beds of the South Rewa Basin, India (Lele 1962a,b; Rao & Lele 1963; Bose 1974), three from the Camden Haven Forma¬ tion near Laurieton, NSW (Holmes & Ash 1979) and three from the Burgersdorp Formation of South Africa (Anderson & Anderson 1985). Dicroidium zuberi is espe¬ cially common in low diversity assemblages in the low¬ est Triassic fossil plant horizons of the Barreal and Cacheuta Basins of Argentina (Frenguelli 1944a,b, 1948; Anderson & Anderson, 1983), and is also known from early Triassic (Malakhovian or Scythian) marine rocks in New Zealand (Retallack 1985). There is no evidence of Dicroidium odontopteroides or any other Middle Triassic forms. Considering recent radiometric dating of Middle Triassic rocks with D. odontopteroides in New South Wales and New Zealand (Retallack et al. 1993), the Culvida Soak flora is some 248-244 million years old. 58 Journal of the Royal Society of Western Australia, 78(3), September 1995 Thus, the Culvida Sandstone is about the same age as the Erskine Sandstone and Blina Shale, which succes¬ sively underlie the Culvida Sandstone in the Canning Basin. The Blina Shale has also yielded the lycopsid Pleuromeia indica (Lele) Dobruskina (1985: see White & Yeates 1976, PI 6, Figs 18-21, PI 12, Figs 42-44) as well as a sparse marine fauna (Gorter 1978). Plant fossils from the Erskine Sandstone include the characteristic rounded sporophylls, as well as other remains of Pleuromeia sternbergi (Munster) Corda in Germar (1852: see Foord 1890; Brunnschweiler 1954; White & Yeates, 1976, PI 6, Fig 17b, PI 8, Fig 29) as well as abundant Dicroidium zuberi (Szajnocha) Archangelsky (1968: see Antevs 1913; Townrow 1957). These rock units probably represent beach ridges and coastal lagoons respectively, outboard of the fluvial Culvida Sandstone. Marine regression that created the general sequence Blina-Erskine-Culvida For¬ mations in the Canning Basin (Gorter, 1978) also created the correlative sequence of Garie-Newport Formations in the Sydney Basin (Retallack 1975, 1980a). Paleoecology and paleoclimatology of Culvida Soak The Culvida Soak fossil flora has most plants of the Dicroidietum zuberi fossil plant association of eastern Australia (Retallack 1977). This was a broadleaf flora dominated by the extinct seed fern Dicroidium zuberi, whose presence is confirmed for the Culvida flora by its characteristic pollen-bearing and ovuliferous organs. This widespread early Triassic assemblage was probably a heath or forest assemblage of nutrient poor soils. These indications of oligotrophy include low species diversity and thick leathery leaves. The Culvida Soak flora shows rather smaller leaf size in Dicroidium zuberi and D. narrabeenensis, and fewer pteridophytes than comparable floras of eastern Australia, and this may be an indication of a drier climate. The Culvida flora also contains a few specimens of Dicroidium hughesii, which is the dominant taxon of a fossil plant association that can be termed the Dicroidietum hughesii. Based on the fossil flora of the Burgersdorp Formation near Aliwal North, South Africa, this association contains a variety of pinnate cycadophyte and ginkgolike leaves not found in early Triassic floras of southeastern Australia (Anderson & Anderson 1985). Cycadophytes and ginkgolike remains also are found in the flora of the Parsora Beds of India (Lele 1962a,b; Rao & Lele 1963; Bose 1974), which is another example of the Dicroidietum hughesii. This fossil plant assemblage of India and South Africa lies within the dry to monsoonal subtropical paleoclimatic belt of the Gondwana supercontinent (Parrish 1990), whereas the Dicroidietum zuberi of southeastern Australia is within the humid cool temperate paleoclimatic region within the Antarctic circle (Anderson & Anderson 1985). The Culvida flora was probably near the ecotone between these two floristic regions. Its mix of species from north and south may be an indication of continuity of temperate woodlands through central Australia during the early Triassic. Despite this regional variation in floras, the uniformity and low diversity of early Triassic floras is impressive. It may be a lingering artifact of the Permian- Triassic life crisis. Extinctions of land plants have re¬ cently been shown to have been coeval with, and as severe as, the great dying of marine organisms (Retallack 1995). In addition, low diversity oligotrophic floras dominated by conifers and lycopods persisted for many millions of years after the Permian-Triassic boundary, with diversity of pteridosperm-dominated floras only at¬ taining levels found in the Late Permian by Middle Tria¬ ssic time with the Dicroidium odontopteroides flora of Gondwana and Scytophyllum flora of Laurasia. This low diversity oligotrophic interregnum was a time when there was no peat deposition anywhere in the world (Retallack et al. 1995). The record of land animals similarly shows a dramatic decline in diversity at the Permian-Triassic boundary followed by a cosmopolitan Lystrosaurus fauna of the early Triassic and then a diverse and provincial tetrapod fauna of the middle Triassic (Benton 1987). Similarly, in the sea, Early Triassic faunas were depauperate (Erwin 1994) and there are no known early Triassic reefs (Flugel 1994). These peculiarities of the early Triassic mark it as an unusual time. The big life crises in the history of life may indeed change the rules, making difficult the application of uni- formitarianism in interpreting the past. Systematic palaeobotany CLASS LYCOPSIDA ORDER ISOETALES GENUS PLEUROMEIA Corda for Germar 1852 Pleuromeia dubia (Seward) comb nov Holotype. Stem cast (South African Museum 13727) Type locality. Alcocks Quarry, near Aliwal North, South Africa; Burgersdorp Formation: Early Triassic (Anderson & Anderson 1985). Description. The single leaf found at Culvida Soak (F9155, Fig 2E) is very similar to "Cylomeia undulata" (White 1981), but that name is unsuitable for a variety of reasons. The holotype of "C. undulata" is a fossil leaf identical to the Culvida fossil, called "Taenioptcris undulata" by Burges (1935). Its lateral wrinkles are probably due to compression or drying of a thick and fleshy lycopsid leaf (White 1981). I could not make out a supposed "very delicate" lateral venation spaced at "30 per cm" noted by Burges (1935) as evidence of cycadophytic affinities on either the type or Culvida material. Kimura (1959) was evidently not aware of this species when he erected the name Taeniopteris undulata for some quite different cycadophyte leaves of Upper Jurassic age in Japan. Confusion is now removed by the discovery of "Cylomeia undulata" leaves attached to lycopsid stems of " Gregicaulis " dubius in both the Newport Formation near Sydney Australia (Retallack 1973) and the Burgersdorp Formation near Aliwal North, South Africa (Anderson & Anderson 1985). As¬ sociated oval sporophylls with rounded tips have been found both in South Africa (Anderson & Anderson 1985) and in the Culvida Sandstone (White & Yeates 1976, PI 12, Figs 44,45). Pleuromeiacean lycopsids are best identified according to reproductive structures 59 Journal of the Royal Society of Western Australia, 78(3), September 1995 ™Xpsnp?HvptrvnyCr^PS1wandreL fe?*(r°T5:UlV!da Soak; A' B- Clad°Phlebis carnei (F9160); C, G, Dicroidium narrabeenense (F9154, h9149 respectively), D, Dicroidium hughesn (F9155); E, Cylomeia undulata (F9155); F, Dicroidium zuberi (F9144). 60 Journal of the Royal Society of Western Australia, 78(3), September 1995 (White 1981; Sadovnikov 1982; Dobruskina 1985), and these sporophylls leave little doubt that "Gregicaulis" is a junior synonym of Pleuromeia, here taken in a stricter than usual sense. Pleuromeia dubia was originally described as "Stigmatodendron dubium" by Seward (1908), but that genus is based on remains of a Carboniferous arborescent lycopsid from Russia. Distribution. Pleuromeia dubia is known from the Burgersdorp Formation of South Africa (Anderson & Anderson 1985) and the Newport Formation (Retallack 1973; White 1981) and Camden Haven Claystone (Holmes & Ash 1979) of southeastern Australia, all of late Early to early Middle Triassic age. CLASS PTEROPSIDA ORDER & FAMILY INCERTAE SEDIS GENUS CLADOPHLEBIS Brongniart emend Frenguelli 1947:12 Cladophlebis carnei Holmes & Ash 1979 (Figs 3A,B) Holotype. Australian Museum partial frond F59425 Type locality. Camden Head near Laurieton, NSW; Camden Haven Claystone; late early to early Middle Triassic. Description. One of the fossils from Culvida Soak (F9160, Fig 2A,B) has the characteristic falcate pinnules and singly forked lateral venation of this species. It is most similar to Cladophlebis oblonga Halle, which has somewhat blunter, wider pinnules and thicker rachis (Frenguelli 1947). Affinities. Similar leaves are found with sporangia similar to those of Asterotheca in the Newport Formation of NSW (Retallack 1973, 1980a), Estratos del Alcazar of Argentina (Menendez 1957), and Burgersdorp Formation of South Africa (Anderson & Anderson 1985), so these leaves were probably marattialean ferns. Distribution. In addition to the type locality, this spe¬ cies is known from the early Triassic Newport Forma¬ tion near Sydney, NSW (Retallack 1973, 1980a), Estratos del Alcazar near Hilario, Argentina (Menendez 1957), Burgersdorp Formation near Aliwal North, South Africa (Anderson & Anderson 1985), and Middle Triassic Gunnee Beds near Delungra, N.S.W. (Bourke et al. 1977). GENUS CH1ROPTER1S Kurr emend Riihle von Lilienstern 1931:273 Chiropteris zvhitei sp nov (Figs 3F-J,4C) 1961 Ginkgoites antarctica White p 302, PI 3, Fig 5 Holotype. The most complete leaf (F9158, Fig 4C), whose outline is reconstructed in Fig 3G. Type locality. Culvida Soak, Canning Desert, Western Australia; Culvida Sandstone, late Early to early Middle Triassic. Diagnosis. Reniform leaves, width about 40 mm (25-60 mm), with a petiole at least 12 mm long attached at a wide angle to the blade; leaf marine crenate and slightly recurved; venation fine, evenly radiating from petiole. anastomosing and dichotomising throughout the leaf, obscured by interveinal woody striae. Derivation. The specific epithet is from Mary E White, who pioneered paleobotanical studies of the Canning Basin and other outback Australian localities. Comparison. Although very similar to the leaves of Ginkgo, these fossils (F9150, F9156-9) differ in their anastomosing venation (Fig 4C), reniform shape (Fig 3G) and angle of insertion of the petiole (reconstructed in Fig 31). The venation is obscured by abundant woody interveinal striae, and the leaf texture was stiff enough for the conical shape of the leaf blade and a slightly recurved margin to have resisted compaction (Fig 4C). The leaf was also wrinkled in broad zones corresponding to growth ridges. These also are differences from ginkgoalean leaves. There are two distinctive kinds of species within the genus Chiropteris now that the apetiolate woody species have been removed to Ginkgophytopsis (by Retallack 1980b). One group includes Chiropteris zeilleri (Seward 1903), and C. barrealensis (Frenguelli 1942) and has pal¬ mate, flabellate leaves with widely spaced clear vena¬ tion, characteristically without anastomosis between the central two veins, unlike C. zvhitei. Another group in¬ cludes the type species Chiropteris lacerala (Riihle von Lilienstern 1931), C. reniformis (Kawasaki 1925), C. kazvasakii (Kon'no 1939) and C. harrisii (Archangelsky 1960). These are conical, circular or reniform leaves with dense venation and common interveinal striae, like C. zvhitei. Of these species, C. zvhitei is most like C. harrisii, which differs in being twice the size, with wider lateral lobes. These species have much less developed lateral lobes than in C. reniformis and C. kazvasakii. The type species of the genus, C. lacerata is more strongly conical, cutinized and lignified than C. zvhitei, and has in addi¬ tion a lamina more dissected and lobes not extending backward past a right angle to the petiole. Affinities. Affinity of these plants with ginkgoaleans has long been thought unlikely, and Riihle von Lilienstern (1931) proposed that the type species was a dipteridacean fern on the basis of its reflexed conical leaf shape, anastomosing venation and supposed sori. These latter look more like insect domatia (of Stace, 1965) than sori and the details of the venation are quite unlike the nearly square meshes of typical dipteridacean ferns. Another possibility is that these are progymnosperms allied to Archaeopteris as suggested for the similar Ginkgophytopsis by Retallack (1980b). A pos¬ sible progymnosperm sporangial axis has been found in a New Zealand locality yielding Ginkgophytopsis (Pole & Raine, 1994), but both Gingkophytopsis and Chiropteris re¬ main problematic. CLASS PTERIDOSPERMOPSIDA FAMILY CORYSTOSPERMACEAE Thomas 1933 GENUS DICROIDIUM Gothan emend. Townrow 1957:26. Dicroidium hughesii (Feistmantel) Lele 1962a (Figs 2D, 4A) Remarks. This large unipinnate leaf (F9155) was evi¬ dently forked, as two rachides converge on the slab (Figs 61 journal of the Royal Society of Western Australia, 78(3), September 1995 F91436 F9LS2 rP^rHvph!? and Pr°blematlca from Culvida Soak; A, Lepidopteris madagascariensis (F9148); B-D, Dicroidium zuberi (F9151, lively). resPectlvelY)' E' Umkomasm sp indet (F9147); F-J, Chiroptens whitei (F9159, F9158, F9150, reconstructed leaf, F9156 respec- 62 Journal of the Royal Society of Western Australia, 78(3), September 1995 Figure 4. Fossil seed ferns and problematica from Culvida Soak: A, G, Dicroidium narrabeenense (F9154, F9149 respectively); B, Dicroidium hughesii and Pleurotneia dubia (F9155); C, Chiropteris whitei (F9158); D, Umkomasia sp (F9147); E, F, Dicroidium zuberi (F9152, F9143 respectively). 63 Journal of the Royal Society of Western Australia, 78(3), September 1995 2D, 4A). It shows veins clearly, as well as basiscopic lobes to the pinnae that are the hallmarks of this species, distinguishing it from the otherwise similar Dicroidium narrabeenense Jacob & Jacob (1950). Dicroidium eskense is another comparable leaf distinguished by asymmetric pinnae, but in this case due to a deep acroscopic sinus in the pinnae (Retallack 1977). Distribution. Dicroidium hughesii is well known from the Parsora Beds of India (Lele 1962a) and the Burgersdorp Formation of South Africa (Anderson & Anderson 1985). In both places, it is the dominant taxon, whereas only a few specimens were found at Culvida Soak [including material of White (1961) and White & Yeates (1976)]. Dicroidium narrabeenense (Walkom) Jacob & Jacob 1950 (Figs 2C-D,4B,G) Remarks. This species is represented by two specimens (F9149, F9153-4). It is here interpreted in the broad sense of Anderson & Anderson (1983), who included bipinnatifid remains referred to " Dicroidium australe" by Jacob & Jacob (1950) as well as smaller leaves (Fig 2G) that formerly would have been referred to " Dicroidium lancifolium” . However it has long been recognized that "D. lancifolium” is an extreme variant of D. odontopteroides (Retallack 1977), as well as of D. narrabeenense . The leaves similar to "D. lancifolium” associated with D. odontopteroides have pinna bases separated at the rachis, and are thin in texture, with clear venation, and thin cuticles with laterocytic stomata (Anderson & Anderson 1983), whereas leaves similar to ”D. lancifolium” associated with D. narrabeenense have confluent to overlapping pinna bases, are so thick that veins are difficult to see, and have thick cuticles with cyclocytic stomata (Jacob & Jacob 1950; Retallack 1973). Although cuticles cannot be prepared from the Culvida Soak specimen similar to ”D. lancifolium ”, it was thick with confluent pinna bases as in D. narrabeenense. Distribution. Dicroidium narrabeenense has been found in the Newport Formation near Sydney (Retallack 1973, 1980a) and the Camden Head Claystone near Laurieton, NSW (Holmes & Ash 1979), both late Early to early Middle Triassic (Scythian-Anisian). Comparable leaves also have been found from the Parsora Beds of India (as "Dicroidium odontoperoides” of Lele 1962a), of the same age (Anderson & Anderson 1983). Dicroidium zuberi (Szajnocha) Archangelsky 1968 (Figs 2F, 3B-D, 4E-F) Remarks. This is a common, widespread and polymor¬ phic species. One of the Culvida specimens is most like D. zuberi var papillatum (Townrow) Retallack (1977: F9145) which has rhomboidal pinnules, but most of them are like D. zuberi var sahnii (Seward) Retallack (1977: F9143-4, F9146-7, F9149-50, F9152) which has rounded pinnules that tend to coalesce into lobed outer pinnae. These varieties were intended as form-taxa. The frond size of the Culvida specimens is small and compact for this species (Retallack 1977, 1980a). Artabe (1990) has proposed that Frenguelli's genus Zuberia be resurrected for bipinnate Dicroidium leaves with rachis pinnules, so that the varieties of Retallack (1977) would become species of Zuberia. Many of these bipinnate Dicroidium leaves do have distinctive cyclocytic stomata (Townrow 1957; Anderson & Ander¬ son 1983), but some have laterocytic stomata and cu¬ ticles otherwise identical to those of Dicroidium odontopteroides (Townrow 1957). Until more information is available on cuticular and other distinctive characters of the type Argentinian specimens of Zuberia , that genus is maintained as a junior synonym of Dicroidium. Distribution. Very similar leaves to D. zuberi var papillatum have been found in the Newport Formation near Sydney, NSW (Retallack 1973, 1980a) and in the Erskine Sandstone at Derby, Western Australia (Antevs 1913; Townrow 1957). Dicroidium zuberi var sahnii is known from the Newport Formation near Sydney, NSW (Retallack 1973, 1980a) and from the Parsora Beds of the South Rewa Basin, India (Seward 1932; Lele 1962a). GENUS UMKOMAS1A (Thomas) emend Holmes 1987: 166. Umkomasia sp indet (Fig 2E) Remarks. This single cupule (F9147) is 8.3 x 6.2 mm in size, and appears identical to the more complete branching group of three cupules from Culvida Soak illustrated by White & Yeates (1976, PI 13, Fig 49). They appear to have three or four cupule lobes. They thus have fewer lobes and are smaller in size than early Triassic Umkomasia feistmantelii (Holmes & Ash) Holmes (1987) and larger in size with more cupule lobes than other described species of middle and late Triassic age (Thomas 1933; Holmes 1987). Distribution. The Culvida specimens are most like those associated with the type material of Dicroidium zuberi from the early Triassic Barreal Formation, near Barreal, Argentina (Frenguelli 1944a). FAMILY PELTASPERMACEAE Thomas 1933 GENUS LEPIDOPTERIS Schimper emend Townrow 1956; 4. Lepidopteris tnadagascariensis Carpentier 1935 (Fig 2A) Remarks. Only one specimen was found (F9148), but it shows the distinctive rachis-pinnules, rounded pinnule apices and a conspicuous median ridge to the rachis that is distinctive for this species. The current specimen lacks the rounded pinnules and large blisters on the rachis of the Permian species Lepidopteris stuttgardiensis and L. martinsii, but has less pointed and lobed pinnules than the middle Triassic species L. stormbergensis (Townrow 1966). Its pinnules are not so large or coalescent as in L. brownii or L. africana (Anderson & Anderson 1989). Con¬ stricted pinnule bases distinguish L. langlohensis (Ander¬ son & Anderson 1989). Distribution. This species has been found in the Hawkesbury Sandstone and Newport Formation near Sydney, NSW (Townrow 1966), the Camden Haven Claystone near Laurieton, NSW (Holmes & Ash 1979), and the Burgersdorp Formation near Aliwal North, South Africa (Anderson & Anderson 1985), all of late 64 Journal of the Royal Society of Western Australia, 78(3), September 1995 Early to early Middle (Scythian-Anisian) age. The type material was found at Amboriky in the early Triassic bed 3 of the Sakamena Group of Madagascar (Carpentier 1935). Acknowledgements : I thank R J Paten of Mines Adminstration Pty Ltd for sending these fossils and R E Gould for notifying me about them. A N Yeates, J E Gorter and M E White offered useful discussion on Canning Basin geology and paleontology, J H Lord of the Western Australian Geo¬ logical Survey provided a repository for the specimens. Research was funded by a Commonwealth Postgraduate Award and NSF grant OPP9315228. References Anderson J M & Anderson H M 1983 Palaeoflora of southern Africa Molteno Formation (Triassic). Vol 1 Part 1 Introduction. Part 2 Dicroidium. A A Balkema, Rotterdam. Anderson J M & Anderson H M 1985 Palaeoflora of southern Africa. Prodromus of South African megafloras, Devonian to Lower Cretaceous. A A Balkema, Rotterdam. Anderson J M & Anderson H M 1989 Palaeoflora of southern Africa. Vol 2 Gymnosperms (excluding Dicroidium). A A Balkema, Rotterdam. Antevs E 1913 Results of Dr E. Mojberg's Swedish scientific expeditions to Australia 1910-1913. V Some Mesozoic plants. Kungliga Svenska VetenskapsAkademien Handlingar 51:1-71. Archangelsky S 1960 "Chiropteris harrisii " nueva especies de la Serie La Golondrina, Provincia de Santa Cruz. Acta Geologica Lilloana 3:289-293. Archangelsky S 1968 Studies on Triassic fossil plants from Argentina. IV The leaf genus Dicroidium and its possible relation to Rhexoxylon stems. Palaeontology 11:500-512. Artabe A E 1990 Revalidacion del genero triasico Zuberia Frenguelli 1943, familia Corystospermaceae. Revista del Museo de La Plata 9 Paleontologia 55:145-157. Benton M J 1987 Mass extinction among families of non-marine tetrapods: the data. Memoire de la Societe Geologique de France 150:21-32. Bose M N 1974 Triassic floras. In: Aspects and appraisal of Indian paleobotany (eds K R Surange, R N Lakhanpal & P C Baradwaj). Birbal Sahm Institute, Lucknow, 285-293. Bourke D J, Gould R E, Helby R, Morgan R & Retallack G J 1977 Floral evidence for a Middle Triassic age of the Gunnee Beds and Gragin Conglomerate, near Delungra, New South Wales. Journal and Proceedings of the Royal Society of New South Wales 110:33-40. Brunnschweiler R O 1954 Mesozoic stratigraphy and history of the Canning Desert and Fitzroy Valley, Western Australia. Journal of the Geological Society of Australia 1:35-54. Burges N A 1935 Additions to our knowledge of the flora of the Narrabeen Stage of the Hawkesbury Series in New South Wales. Proceedings of the Linnaean Society of New South Wales 60:257-264. Carpentier A 1935 Etudes paleobotaniques sur le groupe de la Sakoa et le groupe Sakamena (Madagascar). Annales Geologiques du Service de Mines Republic du Malagasy 5:1-32. Dobruskina I A 1985 Some problems of the systematics of Triassic lepidophytes. Paleontological Journal 19:74-88. Erwin D H 1994 The Permo-Triassic extinction. Nature 367:231- 235. Fliigel E 1994 Pangean shelf carbonates: controls and paleoclimatic significance of Permian and Triassic reefs. In: Pangaea: paleoclimate, tectonics and sedimentation during accretion, zenith and breakup of a supercontinent (ed G de V Klein). Special Paper of the Geological Society of America 288:247-266. Foord A H 1890 Notes on the palaeontology of Western Austra¬ lia. Geological Magazine 7:97-105. Frenguelli J 1942 Contribuciones al concimiento de la flora del Gondwana Superior en la Argentina. X Chiropteris barrealensis n sp. Notas de la Museo de La Plata 7 Paleontologia 51:341-353. Frenguelli J 1944a Las especies del genero "Zuberia" en la Argen¬ tina. Anales de la Museo de La Plata Paleontologia B Paleobotanica 2:1-74. Frenguelli J 1944b Contribuciones al concimiento de la flora del Gondwana superior en la Argentina XV La florula de la base de la Serie de Cacheuta en el Cerro de los Banos, Mendoza. Notas de la Museo de La Plata 9:271-310. Frenguelli J 1947 El genero Cladophlebis y sus representantes en la Argentina. Anales de la Museo de La Plata 2 Paleontologia 2:74 p. Frenguelli J 1948 Estratigrafia y edad de llamado retico en la Argentina. Gaea 8:159-309. Germar E F 1852 Sigillaria sternbergi Munster aus dem bunten sandstein. Deutsch Geologische Gesellschaft Zeitschrift 4:183-189. Gorter J D 1978 Triassic environments in the Canning Basin, Western Australia. Bureau of Mineral Resources Journal of Australian Geology and Geophysics 3:25-33. Holmes W B K 1987 New corystosperm ovulate fructifications from the Middle Triassic of eastern Australia. Alcheringa 11:165-173. Holmes W B K & Ash S R 1979 An early Triassic megafossil flora from the Lome Basin, New South Wales. Proceedings of the Linnaean Society of New South Wales 103:47-70. Jacob K & Jacob C 1950 A preliminary account of the structure of the cuticles of Dicroidium ( Thinnfeldia ) from the Mesozoic of Australia. Proceedings of the National Institute of Sciences India 16:101-126. Kawaski S 1925 Some older Mesozoic plants in Korea. Bulletin of the Geological Survey of Chosen 4:71 p. Kimura T 1959 On the Tetori flora (Part II) Addition to the Meso¬ zoic plants from the Kuzuryu Sub-group, Tetori Group, Ja¬ pan. Bulletin of the Science High School Attached to Tokyo University of Education 2:104-120 Kon'no E 1939 A new Chiropteris and other fossil plants from the Heian System, Korea. Japanese Journal of Geology and Geo¬ physics 16:105-114. Lele K M 1962a Studies in the Middle Gondwana flora. 1 On Dicroidium from the South Rewa Gondwana Basin. Palaeobotanist 10:48-68. Lele K M 1962b Studies in the Indian Middle Gondwana flora. 2 Plant fossils from the South Rewa Gondwana Basin. Paleobotanist 10:69-83. Menendez C A 1957 Asterotheca hilariensis sp nov, del triasico superior de Hilario, San Juan. Ameghiniana 1:25-30. Parrish J T 1990 Gondwanan paleogeography and paleoclimatol- ogy. In: Antarctic paleobiology (eds T N Taylor & E L Taylor). Springer, New York, 15-26. Pole M S & Raine J I 1994 Triassic plant fossils from Pollock Road, Southland, New Zealand. Alcheringa 18:147-159. Rao A R & Lele K M 1963 On the cuticle of Dicroidium (Thinnfeldia) sahnii (Seward) with some observations on the genera Thinnfeldia and Dicroidium. Palaeobotanist 11:7-12. Retallack G J 1973 Stratigraphy, palaeobotany and environmen¬ tal analysis of an area around Pittwater, north of Sydney, NSW. BSc Hons Thesis, University of New England, Armidale, NSW. Retallack G J 1975 The life and times of a Triassic lycopod. Alcheringa 1:3-29. Retallack G J 1977 Reconstructing Triassic vegetation of eastern Australasia: a new approach to the biostratigraphy of Gondwanaland. Alcheringa 1:247-277. 65 Journal of the Royal Society of Western Australia, 78(3), September 1995 Retallack G J 1980a Late Carboniferous to Middle Triassic megafossil floras from the Sydney Basin. In: A guide to the Sydney Basin (eds R ) Helby & C Herbert). Bulletin of the Geological Survey of New South Wales 26:385-430. Retallack G J 1980b Middle Triassic megafossil plants and trace fossils from Tank Gully, Canterbury, New Zealand. Journal of the Royal Society of New Zealand 10:31-63. Retallack G J 1985 Triassic fossil plant fragments from shallow marine rocks of the Murihiku Supergroup, New Zealand. Journal of the Royal Society of New Zealand 15:1-26. Retallack G J 1995 Permian-Triassic life crisis on land. Science 267:77-80. Retallack G J, Renne P R & Kimbrough D L 1993 New radiomet¬ ric ages for Triassic floras of southeast Gondwana. In: The non-marine Triassic (eds S G Lucas & M Morales). Bulletin of the New Mexico Museum of Natural History and Science 3:415-418. Retallack G J, Veevers J J & Morante R 1995 Global coal gap between Permian-Triassic extinction and Middle Triassic re¬ covery of peat-forming plants. Bulletin of the Geological Soci¬ ety of America (in press). Riihle von Lilienstem H 1931 Uber Chiropteris. Palaontogische Zeitschrift 13:253-277. Sadovnikov G N 1982 The morphology, systematics and distribu¬ tion of Tomiostrobus. Paleontological Journal 16:100-109. Seward A C 1903 Fossil flora of Cape Colony. Annals of the South African Museum 4:1-122. Seward A C 1908 On a collection of fossil plants from South Africa. Quarterly Journal of the Geological Society of London 64:83-108. Seward A C 1932 On some fossil plants from the Parsora Stage, Rewa. Record of the Geological Survey of India 66:235-343. Stace C A 1965 Cuticular studies as an aid to plant taxonomy. Bulletin of the British Museum, Natural History, Botany 4:78 p. Thomas H H 1933 On some pteridospermous plants from the Mesozoic rocks of South Africa. Philosophical Transactions of the Royal Society of London B222: 193-265. Townrow J A 1956 The genus Lepidopteris and its southern hemisphere species. Avhandlingar Norske Videnskaps Akademian Oslo Mathematische-Naturwissenschaften Klasse 2:1-28. Townrow J A 1957 On Dicroidium probably a pteridospermous leaf, and other leaves now removed from this genus. Transactions of the Geological Society of South Africa 60:1-36. Townrow J A 1966 On Lepidopteris madagascariensis Carpentier (Peltaspermaceae). Journal and Proceedings of the Royal Society of New South Wales 98:203-214. White M E 1961 Appendix 6. In: The geology of the Canning Basin (eds J J Veevers & A T Wells). Bulletin of the Bureau of Mineral Resources, Geology and Geophysics, Canberra, Australia 60:291-310. White M E 1981 Cylorneia undulata (Burges) gen et comb nov, a lycopod of the early Triassic strata of New South Wales. Records of the Australian Museum 33:723-734. White M E & Yeates A N 1976 Plant fossils from the northeastern part of the Canning Basin, Western Australia. Records of the Bureau of Mineral Resources, Geology and Geophysics, Canberra, Australia 1976/18, 32 p. Yeates A N, Crowe R W A, Towner R R, Wybom L A I & Passmore V L 1975. Geology of the Gregory Sub-basin and adjacent areas of the Canning Basin, Western Australia. Records of the Bureau of Mineral Resources Geology and Geophysics, Canberra, Australia 1975/77, 47 p. 66 Journal of the Royal Society of Western Australia, 78:67-79, 1995 New Pleistocene and Holocene stratigraphic units in the Yalgorup Plain area, southern Swan Coastal Plain V Semeniuk 21 Glenmere Road, Warwick WA 6024 Received June 1995; accqjted November 1995 Abstract Within the area of Yalgorup Plain (amended from the Yoongarillup Plain) between Mandurah and Bunbury, there are a variety of fossiliferous limestone, aeolian limestone, and quartz sand units. Three new Pleistocene formations are recognised: the Tims Thicket Limestone, the Myalup Sand, and the Kooallup Limestone. These formations, previously part of the Tamala Limestone, underlie distinct Pleistocene landforms and comprise distinct shore-parallel systems formed by coastal progradation, interrupted by subaerial unconformities. They represent sedimentation during the Pleistocene, which was dominated by cuspate forelands, barriers, and wave-built platforms. The western edge of the Yalgorup Plain is bordered by Holocene barrier dunes which are a geomorphic extension of the Leschenault Peninsula barrier. Various Holocene formations underlie the barrier in the Yalgorup region. Generally, they can be assigned to the Safety Bay Sand, Becher Sand, Bridport Calcilutite, and Leschenault Formation, and to the Preston Beach Coquina, a new unit formally defined in this paper. Four types of large-scale standard Holocene stratigraphic sequences can be recognised in this region: Type 1, prograded bank, beach, beachridge sediments; Type 2, prograded shoreface to beachridge sediments; Type 3, retrograded barrier dune sediments overlying estuarine lagoonal sediment; and Type 4, retrograded barrier dune sediment unconformable on Pleistocene sediment. These standard sequences can be used to help unravel the Holocene history of the barrier dunes in the Yalgorup area, and specifically between Preston Beach and the Bouvard Reefs. Introduction The western part of the Swan Coastal Plain between Mandurah and Bunbury contains Pleistocene limestone (McArthur & Bettenay 1960; Playford et al. 1976), linear wetlands and estuarine lagoons, and a shore-parallel Holocene barrier dune system (Searle & Semeniuk 1985; Semeniuk 1985). Within this area is the Yalgorup Plain (amended here from the Yoongarillup Plain), a Pleistocene to Holocene surface developed on a variety of fossiliferous limestone, aeolian limestone, and quartz sand. Previous drilling in this area was conducted by Commander (1988). Recent research has delineated a range of new Pleistocene to Holocene stratigraphic units in this area, which are described in this paper. Recognition of these new units has provided a tool to unravel the local Pleistocene and Holocene coastal history, as described in Semeniuk (1995a, b). Stratigraphic and geomorphic patterns derived from the Pleistocene terrain in the region show that there are distinct shore-parallel tracts of Pleistocene landforms that developed by coastal progradation, interrupted by subaerial unconformities. Sedimentation was dominated by narrow beachridge plains and cuspate forelands (Semeniuk 1995a). Defining a new Holocene stratigraphic unit in the region has also aided in the recognition of four key stratigraphic sequences that help unravel coastal history. © Royal Society of Western Australia 1995 Stratigraphic data for this paper are from drill holes, road cuts, quarries and excavations. Description of stratigraphy and drilling transects, apart from those pro¬ vided for the type sections in this paper, are presented more fully by Semeniuk (1995a,b). Amendment of the Yoongarillup Plain to Yalgorup Plain There has been a problem with the concept, nomen¬ clature and application of the term Yoongarillup Plain in the Mandurah-Bunbury area (McArthur & Bartle 1980; Semeniuk 1990). Originally the term "Yoongarillup" was used to denote a quartz sand underlying a plain developed on fossiliferous limestone near Busselton, and named after locality in that area (McArthur & Bettenay 1958). Through application in mapping, the term "Yoongarillup" eventually came to encompass a wider range of landforms and soils. The term Yoongarillup Plain was first formally used as a landform/soil unit describing a melange of near-coastal plains and flats between Mandurah and Bunbury (McArthur & Bartle 1980) that included coastal plains underlain by fossiliferous limestone, supratidal to tidal estuarine flats, and geomorphically degraded Quindalup Dunes. An earlier attempt was made to rationalise the use of the term firstly, by excluding the terrain of Quindalup Dunes and estuarine-related flats from the definition, and secondly by restricting the term to refer to the genetically distinct fossiliferous limestone plains 67 Journal of the Royal Society of Western Australia, 78(3), September 1995 formed by Pleistocene marine progradation that are re¬ stricted palaeogeographically to the Mandurah-Bunbury area (Semeniuk 1990). At the time of this first amend¬ ment (Semeniuk 1990), the opportunity also existed to rectify the name, but this was not done. The term "Yoongarillup", deriving from the locality of Yoongarillup, south of Busselton, should refer to features in that region, and not to those restricted to the coastal tract between Bunbury and Mandurah. At present, following extended usage from the Busselton area into the Mandurah-Bunbury area by McArthur & Bartle (1980), and later amendment by Semeniuk (1990), the term "Yoongarillup Plain" now refers to a coastal plain quite removed from its origin. Following the results that clarify the Pleistocene palaeogeography in this region (Semeniuk 1995a), I propose that the situation be now rectified: firstly; that the term Yoongarillup Plain be restricted to the Busselton area; and secondly, that the term "Yalgorup Plain", derived from the local region, be applied to the set of coastal landforms that are palaeogeographically restricted to the Mandurah-Bunbury region. In this way, two genetically and geographically distinct plains will have different names. The Pleistocene palaeogeographic system, and stratigraphic units within the Yalgorup Plain and adjacent systems The Yalgorup Plain The Yalgorup Plain is long and narrow, some 60 km long and 5-6 km wide (Fig 1), and is underlain by fossiliferous limestone, aeolian limestone and quartz sand. Though generally of low relief and undulating, there is local relief of 5-10 m (and up to 15 m) in the form of either calcarenitic aeolianite ridges or quartz sand ridges. The Plain is bordered to the east by a prominent ancestral hinterland ridge (the Mandurah- Eaton ridge; Semeniuk 1995a) that is a linear, moderately high (on average 20-40 m high, locally up to 70 m high, and itself on average some 20 m higher than the Yalgorup Plain), and 3-4 km wide, extending in a north- south direction for 90 km from Mandurah to Eaton (Fig 1), with a slight concavity in its form on the western margin. The junction of this ridge with the Yalgorup Plain is sharp, with the ridge descending steeply down to the Plain. The Mandurah-Eaton ridge is composed of Pleistocene aeolian quartz sand and aeolian limestone but has variable stratigraphy from south to north. To the south, it is mainly quartz sand and lesser limestone (Eaton Sand, Semeniuk 1983), with limestone occurring as aeolianite lenses (Semeniuk & Glassford 1987). To the north, limestone is more common. Even where limestone dominates, the ridge is mantled by yellow quartz sand. The Plain is bordered to the west by a Holocene coastal barrier aeolian ridge (the Leschenault- Preston barrier), or by Holocene estuarine lagoonal deposits. Offshore from the Yalgorup Plain, the nearshore shelf is mostly an unconformity surface cut into Pleistocene limestone, with veneers of quartz sand or Holocene sediment (Semeniuk & Meagher 1981; Searle & Semeniuk 1985). Locally, particularly to the north of the study area, remnant shore-parallel ridges of Pleistocene limestone form discontinuous rocky reefs and islands (the Bouvard Reefs). Another line of former reefs, the Buffalo Reefs, occurs buried under Holocene dunes to the south (Semeniuk 1995a). Generally, there are no prominent limestone ridges or rocky reefs, buried or otherwise, in the central part of the area. The age of the Plain is concluded to be Pleistocene, for the following reasons; 1, the formations that underlie the plain abut or overlie other units in the area are gen¬ erally conceded to be Pleistocene in age ( e.g . the Bassendean Sand, and the yellow sand and limestone that underlies the Mandrah-Eaton Ridge); 2, radiocarbon ages for the limestones at various sites returned ages >30-40,000 years. Pleistocene lithologies Ten types of sedimentary rock and sediment in the Pleistocene sequences occur in this region. These are (limestone terminology follows Dunham 1962): 1. Shelly/bioturbated calcarenite: fine to medium grained, bioturbated to structureless, skeletal quartz grainstone, with abundant calcareous algae and invertebrate skeletons. Shell locally present and randomly oriented by bioturbation; 2. Bioturbated foraminiferal calcarenite: fine to medium sand grainstone of calcareous algae, invertebrate skeletons, quartz, shell grit, whole molluscs, and abundant granule-sized foraminifera ( Marginopora ); bioturbation consists of vertical burrows 3^4 cm diam, penetrating downwards for 50-75 cm, otherwise bioturbation shows general swirling; 3. Laminated and cross-laminated marine calcarenite: medium grained skeletal quartz grainstone, generally without shell beds; conspicuously laminated, cross-laminated and trough-bedded; 4. Laminated to cross-laminated beach calcarenite: skeletal quartz grainstones, with variable structure and texture, but forming a set vertical sequence, usually over 2. 0-2. 5 m. Mollusc shells present (Donax most common); sedimentary structures and two diagnostic cephalopods distinguish four subfacies (Semeniuk & Johnson 1982) - lower part is trough-bedded, medium to coarse grained, with local shells, middle zone is medium to coarse grained, with oriented shell, and low inclined cross¬ lamination, then there is medium grained calcarenite, with bubble-sand structures, and upper part is crudely layered to bioturbated to structureless, medium to coarse grained, with diagnostic cephalopods Sepia and Spirula ; 5. Cross-laminated to structureless calcarenitic aeolianite: large scale cross-laminated to structureless fine to medium grained skeletal quartz grainstone; cross-lamination sets 2-5 m high; calcrete rhizoconcretions and pipes common; 6. Calcreted limestone: this limestone is mainly calcrete in sheet-like form, 20-30 cm thick, or massive, within and on top of the parent limestone, or coating 68 Journal of the Royal Society of Western Australia, 78(3), September 1995 Figure 1. A, B, C & D: Location of study area in southwestern Australia in relationship to regional geology and geomorphology. E: Location of the Holocene Quindalup Dunes, Yalgorup Plain, and offshore reefs, and location of drill sites and quarries for type sections. F: Location of study transects (Transects 1^1 & 7 are presented in Figure 2; the full data including the un-numbered transects are presented in Semeniuk 1995a). G: Map of the Pleistocene geomorphic units (landforms) on the Yalgorup Plain. pipes; lateral and vertical relationships and grada¬ tions, and palimpsest grains and textures indicate parent lithology was aeolianite; 7. Indurated, bored limestone: limestone varies from aeolianite to shelly limestone, and may also be calcreted; key features are induration by calcite cements (fine grained calcite, sparry calcite. calcrete), borings and pot-holes; borings are 1 cm diameter or less; limestone surface may be fissured, with local pot-holes some 10-30 cm diameter, and veneered with rounded limestone gravel and shells of Ninella, Marmarstoma, Littorina, limpets, barnacles; 69 Journal of the Royal Society of Western Australia, 78(3), September 1995 8. Shelly calcilutite: structureless lime mudstone to skeletal lime mudstone; 9. Quartz sand: yellow, grey, or white, medium grained, and well to poorly sorted; grains mostly quartz, with some felspar and minor heavy minerals; quartz sand in this area is largely structureless (cf Glassford & Semeniuk 1990); and 10. Shelly terrigenous mud: dark grey, structureless terrigenous mud with estuarine shells. Some general features about these lithologies are noted here. All limestones are generally white, cream, tan to buff in colour. Grainstones are weakly to strongly cemented by sparry calcite, depending on location relative to the water table. Most limestones also are variably weakly indurated by calcrete, and additionally may be stained by iron oxides. Molluscs in the fossiliferous limestones (Table 1) are typically assemblages from seagrass bank environments (cf Logan et al 1970; Semeniuk & Searle 1985; Semeniuk 1995a). Calcreted limestone is a rock type best developed at unconformity surfaces, and indicates major subaerial exposure. Indurated, bored limestone (often with a gravel veneer of limestone lithoclasts and shells) is generally a micrite- or calcrete-indurated surface that had been exposed as a shore platform or hardground in the submarine shelf environment; this type of limestone has analogues in the Holocene and also indicates a major unconformity. Pleistocene formations Pleistocene limestones and quartz sand form distinct tracts of terrain on the Yalgorup Plain. Three new formations are described here to facilitate interpretation of the regional Quaternary history and palaeogeography. Previously, the limestone units in this area were referred to the Tamala Limestone, but they can be identified readily as distinct units by their lithology, stratigraphy and geography. They are lithologically distinct from the Tamala Limestone at its original location (Logan et al. 1970), at its type section (Playford & Low 1972), and from the calcarenitic aeolianite regarded as Tamala Limestone in the central Swan Coastal Plain of the Perth Regional area (Fairbridge 1953; Klenowski 1976; Gozzard 1983, 1986). Pleistocene limestones that com¬ prise the offshore limestone ridges and the deeper sec¬ tions of the Quaternary profile are left undifferentiated at present. The new Pleistocene formations are: 3. Kooallup Limestone: a Late Pleistocene shoaling sequence of submarine, beach and aeolian calcarenites; 2. Myalup Sand: quartz sand, mostly grey to white, generally sandwiched between the Tims Thicket Limestone and Kooallup Limestone; 1. Tims Thicket Limestone: a shoaling sequence of submarine, beach and aeolian calcarenites depos¬ ited earlier in the Pleistocene. The features of these formations are summarised in Table 2. The stratigraphy and relationships between the units are shown in Figures 2 & 3. Table 1 List of molluscs in the Tims Thicket & Kooallup Limestones BIVALVIA: Anodontia perplea Irus distans Brachidontes ustulatus Mactra australis Callucina lacteola Mactra matthexvi Chlamys asperrimus Mysella sp Chionery cardioides Saccostrea cucculata Divalucina cumin gi Paphies cuneata Dona franasensis Pinna sp Electroma georgiana Taivera coelata Eucrassatella sp Tawera lagopus Glycymeris strialularis Tellina tenuilirata Gomphina undulosa Thraciopsis subrecta Hemidona cliaptnani Wallucina cf jacksoniensis GASTROPODA: Acteocina sp Mitrella austrina Amalda monilifera Mitrella menkeana Amblychilepas oblonga Naccula punctata Astralium squamiferum Natica sp Bedeva paivae Notocochlis gualteriana Bittium granarium Notomella bajula Bulla quoyii Oliva australis Calyptraea calyptraeformis Parcanassa sp Cantharidus lepidus Phasianella australia Cantharidus irisodontes Phasianella solida Cantharidus sp Phasianella ventricosa Clanculus sp Phasiatrochus bellulus Collisella onychitis Polinices conicus Cominella tasmanica Proterato sulcerato Conus anemone Pyrene scripta Dicathais orbita Pyrenidae pseudomycla Drupa sp Syrtwla sp Ethminoiia vitiliginea Thalotia conica Gibbula lehmanni Thalotia lehmanni Gibbula preissana Thalotia pulcherrima Haminoea brevis Thalotia chlorostoma Hipponi conicus Turbo intercostalis Hipponi foliaceus Tanea sagittata Leiopyrga octona Veillum marroxvi Mangelia sp Zafra vercoi Most identifications based on standards identified by G W Kendrick (Western Australian Museum) as cited by Semeniuk & Searle (1985); supplementary identifications from Roberts & Wells (1981) and Wells & Bryce (1985). All molluscs listed, except for the gastropod Parcanassa are also present in Holocene seagrass assemblages, cf Semeniuk & Searle (1985) Tims Thicket Limestone Definition and characteristics: The Tims Thicket Lime¬ stone is a Pleistocene fossiliferous and non-fossiliferous calcarenite cropping out along the eastern Yalgorup Plain. Its lower to middle part is fossiliferous marine, and its uppermost part is aeolian. Derivation of name: Tims Thicket area in the northern Yalgorup National Park. Type section: A disused quarry at 32°39,06” 115°37'36", along Tims Thicket Road. Distribution: The formation is restricted to the Yalgorup Plain area, and crops out almost along the entire length of the eastern part of the Plain (Fig 1). It also is inter¬ sected in cores underlying Holocene units. 70 Journal of the Royal Society of Western Australia, 78(3), September 1995 Table 2 Limestone members in Pleistocene formations in the Yalgorup coastal area FORMATION MEMBER LITHOLOGY COMMENTS KOOALLUP LIMESTONE LAKESIDE MEMBER cross-laminated to structureless, fine and medium calcarenite aeolianites BELLEVUE MEMBER white to cream shelly/ bioturbated calcarenite, laminated cross-laminated shelly calcarenite/coquina, laminated to crosslaminated beach calcarenite submarine to beach zone facies SPRINGHILL MEMBER calcilutite and shelly calcilutite submarine basin facies TIMS THICKETT LIMESTONE WHITE HILL ROAD MEMBER cross-laminated to structureless, fine and medium calcarenite aeolianites CLIFTON DOWNS MEMBER white to cream shelly/ bioturbated calcarenite, bioturbated foraminiferal limestone, laminated cross- laminated marine calcarenit laminated to cross-laminate beach calcarenite submarine to beach zone facies Table 3 Description of Holocene stratigraphic units1 in the Yalgorup coastal area SEDIMENT UNIT STRUCTURE LITHOLOGY HOLOCENE DEPOSITTONAL ENVIRONMENT Safety Bay Sand (aeolian facies) large scale cross laminated to structureless white, cream, orange fine and medium sand dune Safety Bay Sand (beach facies) laminated to bedded; seaward sloping layers white, cream, yellowish medium and coarse sand and shelly sand beach Preston Beach Coquina bedded to crudely layered white, cream, yellowish, & tan shell beds, shell grit, shelly sand, and coarse to medium sand subtidal shoreface to beach Becher Sand structureless, bioturbated crudely layered grey sand and shelly sand; shells from seagrass assemblage subtidal seagrass bank Bridport Calcilutite structureless to bioturbated; shell layers grey calcilutite and shelly calcilutite deep subtidal basin Leschenault Formation structureless, bioturbated, crudely layered grey sand, mud, muddy sand, locally shelly; estuarine shells estuarine sand and mud flats 1 Data from Semeniuk (1983), Semeniuk & Searle (1985, 1987), and this paper. Thickness and geometry: At the type section, the for¬ mation has been logged as 9.1 m thick. Regionally, the unit is a ribbon to wedge, some 60 km long, up to 5 km wide and 5-10 m thick. Lithology: The sequence of limestone at the type section (from 1 at the base to 5 at the top) is: 5. cross-laminated to structureless calcarenitic aeolianite: 2.0 m thick; 4. laminated to cross laminated beach calcarenite: 2.0 m thick; 3. laminated and cross laminated marine calcarenite: 1.3 m thick; 71 Journal of the Royal Society of Western Australia, 78(3), September 1995 LEGEND HOLOCENE SEDIMENT YELLOW QUARTZ SAND REMOBILIZED ON PLEISTOCENE FORMATIONS LARGE SCALE CROSS-LAMINATED FINE/MEDIUM CALCARENITE LAMINATED TO CROSS - LAMINATED BEACH CALCARENITE SHELLY /BIOTURBATED CALCARENITE AND LAMINATED SHELLY CALCARENITE /COQUINA SHELLY AND GRAVELLY LIMESTONE (ON BORED INDURATED LIMESTONE) CALCILUTITE AND SHELLY CALCILUTITE LU Z g < f) LU 2 J Q. D J J < 0 0 * UNASSIGNED PLEISTOCENE TO HOLOCENE SEDIMENTS o O 0 ° o °o o° SHELLY CALCILUTITE, CALCILUTITE, SHELLY MUD AND MUD GREY/WHITE QUARTZ SAND CALCARENITE UNIT 31 P LARGE SCALE CROSS-LAMINATED FINE/MEDIUM CALCARENITE LAMINATED TO CROSS-LAMINATED BEACH CALCARENITE BIOTURBATED CALCARENITE SHELLY /BIOTURBATED CALCARENITE FORAMI NIFERAL CALCARENITE UJ Z g < / > LU 2 J h LU * 0 1 h 0 > 2 h CALCRETE HORIZON BORED INDURATED LIMESTONE UNCONFORMITY AEOLIAN LIMESTONE LENS YELLOW QUARTZ SAND OLDEST PLEISTOCENE LIMESTONE "BASEMENT" BOUNDARY BETWEEN SHELLY AND AEOLIAN LIMESTONE IN PLEISTOCENE "BASEMENT" 2 SITE NUMBER SITE FOR QUARRY ROAD CUT OR CLIFF FACE EXTENT OF QUARRY A SITE FOR CORE LENGTH OF CORE M S L POSITION OF PRESENT M S L KL KOOALLUP LIMESTONE MS MYALUP SAND TTL TIMS THICKET LIMESTONE Figure 2A. Stratigraphy of transects. Legend to transects. 72 Journal of the Royal Society of Western Australia, 78(3), September 1995 Figure 2B. The stratigraphy of the transects 1-4, 7. 73 Journal of the Royal Society of Western Australia, 78(3), September 1995 Figure 3. Summary of stratigraphic relationships between the Pleistocene formations (from Semeniuk 1995a). 74 Journal of the Royal Society of Western Australia, 78(3), September 1995 2. bioturbated foram ini feral calcarenite: 1.8 m thick; 1. shelly /bioturbated calcarenite with seagrass assemblage biota: 2.0 m thick; unconform able on yellow quartz sand. The bioturbated foraminiferal calcarenite is laterally equivalent to shelly/bioturbated calcarenite and lami¬ nated and cross-laminated marine calcarenite. The lime¬ stone types in the formation are in a shoaling-upward sequence, with seagrass bank lithofacies and the later¬ ally equivalent sand wave lithofacies and bioturbated foraminiferal calcarenite lithofacies passing up into a beach sequence and then into a beachridge/dune se¬ quence. Within the beach facies, the subfacies are also in a shoaling sequence: trough cross-bedded calcarenite passes up into laminated calcarenite with Donax (a beach zone indicator), into calcarenite with bubble structures, into structureless calcarenite with Sepia and Spirula , and then finally into large scale cross-laminated aeolian calcarenite. Stratigraphic relationships: The formation unconformably overlies older Pleistocene limestone at depth, abutting and pinching out unconformably to east against yellow sand and limestone that underlie the Mandurah-Eaton Ridge, and pinches out downdip (westwards) as a natural palaeogeographic synoptic sur¬ face. In its eastern parts, it is unconformably overlain by the Myalup Sand filling karst features on the limestone; in its western parts it is unconformably overlain by Ho¬ locene sediment. Fossils: Lower to middle parts of the formation contain molluscan and foraminiferan fossils, indicative of seagrass bank environments. The molluscs are listed and compared with the Holocene seagrass bank assemblage in Table 1. The upper part of the unit, the beach facies, contains Donax spp, Paphies sp, Sepia spp, and Spirula spirula. Members in the formation: The natural lithologic se¬ quence in the formation reflects a sequence of shoaling, with seagrass bank lithofacies and the laterally equiva lent sand wave lithofacies and bioturbated foraminiferal calcarenite lithofacies passing up into a beach sequence and then into a beachridge/dune sequence. In a broad view, the formation consists of a distinctive mid-lower part of fossiliferous limestones passing up into aeolian limestones. It is proposed that the broad separation of the calcarenites from fossiliferous submarine to littoral to non-fossiliferous aeolian be the basis of recognising two members in the formation, as follows (Table 2): White Hill Road Limestone Member : upper part of formation consisting of aeolian calcarenite. The area along White Hill Road, south of Mandurah, exposes good sequences of this member, with relict beachridge patterns evident (see figure 5 of Semeniuk 1995a). The type location for the member is the same as for the formation. Clifton Downs Member: lower to middle part of the formation consisting of fossiliferous marine calcarenite passing up into fossiliferous laminated beach calcarenite, exposed in the Clifton Downs area, east of Lake Clifton. The type location for the mem¬ ber is the same as for the formation. Additonal quarry reference sites are located along Tims Thicket Road in the Tims Thicket area. Myalup Sand Definition and characteristics: The Myalup Sand is a predominantly grey to white quartz sand unit between the Tims Thicket Limestone (below) and the Kooallup Limestone (above). There are local thin lenses of limestone in the formation. Derivation of name: Myalup Swamp, in the southern Yalgorup Plain area, east of Binningup. Type section: Sand ridge at 32°58'00" 115°42'57", south of the Yalgorup National Park. Distribution: The formation is restricted to the Yalgorup Plain, cropping out almost along the entire length of its eastern and middle part. It also is intersected in core in this area. Thickness and geometry: Where intersected in core and trenches, the formation has variable thickness from 5—15 m thick. Regionally, the unit is a ribbon to shoestring, some 60 km long, up to 5 km wide and 5-15 m thick. Lithology: The type section, 15 m thick, is generally structureless grey to white quartz sand, mainly fine to medium sand-sized, and poorly sorted. The formation becomes progressively more iron-stained in the upper 3-4 m. Though quartz rich, it also contains minor feldspar. The upper parts of the formation may be stained yellow. Locally, there are carbonate-rich lenses, <1 m thick, located 2-3 m below present mean sea level. It was not generally possible to differentiate facies in the quartz sand sections of the Myalup Sand. Stratigraphic relationships: Contact with the underlying Tims Thicket Limestone is sharp and uncon- formable, marked by prominent karst in the limestone. Contact with the overlying Kooallup Limestone is sharp and unconformable, with the base of the limestone trun¬ cating the ridge-and-depression sequence in the Myalup Sand which pinches out under the limestone. In the de¬ pressions on the Myalup Sand, there is an overlying undifferentiated suite of Pleistocene to Holocene shelly terrigenous mud, calcilutite and shelly calcilutite (wet¬ land and estuarine facies). The contact of Myalup Sand with the sediments that underlie the Mandurah-Eaton Ridge is difficult to differentiate. Fossils: No fossils has been retrieved from the formation, but the limestone lenses therein hold scope to be fossiliferous. Kooallup Limestone Definition and characteristics: The Kooallup Limestone is a Pleistocene fossiliferous and non-fossiliferous calcarenite that crops out along the western zone of the Yalgorup Plain. Subsurface sections may contain local calcilutite lenses. The lower to middle part of the forma¬ tion is fossiliferous, and the uppermost part is aeolian. Derivation of name: Kooallup Lagoon, east of Lake Preston and north of Myalup. Type section (type locality): Quarry, east of Lake Preston, in the southern part of Yalgorup Plain area, at 32°02,46" 115°42,36”. Distribution: The formation is restricted to the Yalgorup Plain area, cropping out almost along the entire length of the western part of the Yalgorup Plain. Additionally, 75 Journal of the Royal Society of Western Australia, 78(3), September 1995 it has been intersected in core underlying Holocene units in this region. Thickness and geometry: At the type section, the formation is 10.8 m thick. Regionally, the unit is a ribbon to wedge, some 60 km long, up to 5 km wide and 5-16 m thick. Lithology: At the type section, there are four types of limestone; in stratigraphic order, from 1 at the base to 3 at the top, these are: 3. cross-laminated to structureless calcarenitic aeolianite: 1.8 m thick; 2. laminated to cross laminated beach calcarenite: 3.0 m thick; 1. laminated and cross laminated marine calcarenite and shelly/bioturbated calcarenite with seagrass assemblage biota: 6 m. In a section further south from the type section, the same sequence of limestones occur but are underlain by calcilutite: 3. laminated to cross laminated beach calcarenite: 2 m thick; 2. laminated and cross laminated marine calcarenite: 9 m thick; 1. shelly calcilutite (5 m thick); unconformable on an older un-named shelly marine limestone. In general, the Kooallup Limestone sequences exhibit shoaling: seagrass bank facies, and the laterally equiva¬ lent sand wave facies, are overlain by beach and then beachridge/dune facies. The beach sequence progresses from subtidal to supratidal, with preservation of bubble structures and shells of Donax, Sepia, and Spirilla. Lo¬ cally, in former inter-ridge marine depressions, there are lenses of calcilutite. Stratigraphic relationships: The formation unconformably overlies an older as yet un-named Pleistocene limestone at depth, and pinches out unconformably to the east against the Myalup Sand. To the west, it pinches out downdip as a natural palaeogeographic surface, or abuts a buried ridge that is the extension of the Bouvard Reefs system. Fossils: The fossil within the Kooalup Limestone is similar to that in the Tims Thicket Limestone. The lower to middle parts of this formation contain molluscs (Table 1) and foraminifera, derived from seagrass bank environments. The upper part of the unit, the beach facies, contains Donax spp, Paphies sp. Sepia spp, and Spirilla spirilla. Members in the formation: As for the Tims Thicket Limestone, the sequence in the Kooallup Limestone re¬ flects shoaling, with (local basin calcilutites and) seagrass bank lithofacies passing up into beach facies and then into a beachridge/ dune facies. The sequences essentially form a fossiliferous lower to middle part, and a non-fossiliferous (aeolian) upper part. It is proposed that this broad separation of limestones be the basis of recognising three members in the formation, as follows (Table 2): Lakeside Limestone Member: upper part of formation consist¬ ing of aeolian calcarenite. The type location for the member is the same as for the formation; additionally, the pits and quar¬ ries around the Lakeside property east of Lake Preston, and the eastern shore of Lake Preston exposes good sequences of this member. Bellevue Limestone Member: lower to middle part of forma¬ tion consisting of fossiliferous marine limestones and fossilif¬ erous laminated beach calcarenite. The type location for the member is the same as for the formation; additionally, local pits and quarries in the area of the Bellevue property, along the northeastern shore of Leschenault Inlet, expose good sec¬ tions of this member. Springhill Limestone Member: lower part of the formation consisting of calcilutite and shelly calcilutite. A local lens of this calcilutite occurs at depth under the Springhill property (located 2 km SE of Binningup). The type section for the mem¬ ber is Core Site 2, Transect 7 (Fig 2). [Note that use of the Springhill Limestone Member is to be distinguished from the previous use of the term Spring Hill, as two words, for the Palaeozoic Spring Hill Limestone in the Bonaparte Basin; though not formally defined to date in the Bonaparte Basin, the Spring Hill Limestone was first used by Druce 1963]. Distribution: An additional feature in the Kooallup Limestone is a south to north facies change; to the south, beachridge and dune facies above the seagrass bank and beach facies are quartz sand rich; to the north, they are more carbonate-grain rich. This transition is related to major inputs of quartz sand from two sources in the south: erosion of the Mandurah-Eaton ridge (with con¬ comitant net northwards longshore drift), and the Col¬ lie, Preston and Wellesley rivers transporting quartz from the dunes further east. Pleistocene palaeogeographic units The Pleistocene formations are restricted to distinct, mappable tracts of country that represent discrete phases of coastal sedimentation, progradation, and geomorphic history' in this region. In effect, their distribution reflects accretionary stages of the Swan Coastal Plain between Mandurah and Bunbury. In this context, they have been assigned palaeogeographic names to highlight their discrete palaeogeographic character (Semeniuk 1995a). The distribution of the Pleistocene formations in relationship to the Pleistocene landform units are as follows (Semeniuk 1995a): 1. Youdaland: the most eastern and oldest Pleistocene land- form unit on the Yalgorup Plain, underlain by the Tims Thicket Limestone; the name derives from Youda, between Tims Thicket and Mandurah; 2. Myalup Sand Shelf and Myalup Sand Ridge: a sand shelf system that separates the Mandurah-Eaton Ridge from Kooallupland, and a linear ridge system generally sand¬ wiched between the Youdaland and Kooallupland, respec¬ tively, and underlain by the Myalup Sand; the name derives from Myalup Swamp, between Myalup and Binningup; and 3. Kooallupland: the most western and youngest Pleistocene landform unit on the Yalgorup Plain, underlain by Kooallup Limestone; the name derives from Kooallup Lagoon, located to the east of Lake Preston. The evolution of the Yalgorup (coastal) Plain is interpreted to have taken place in several stages related to sealevel still-stands in the Pleistocene (Semeniuk 1995a): 1. formation of an older Pleistocene limestone beachridge plain (Youdaland), within which there was shoaling from 76 Journal of the Royal Society of Western Australia, 78(3), September 1995 marine seagrass carbonate sedimentation to beach to beachridges/ dunes; 2. accumulation of quartz rich coastal sand barriers (Myalup Sand Shelf and Myalup Sand Ridge); 3. formation of a younger Pleistocene limestone beachridge plain (Kooallupland) within which there was, again, shoaling from marine seagrass carbonate sedimentation to beach to beachridges/dunes. Thus the overall progressive accretion of the Yalgorup Plain records, with subaerial interruptions (Fig 3); 1, sedimentation and progradation in a coastal setting partly behind offshore rocky reefs; 2, changes in style from cuspate foreland and shoreface accretion to coastal barriers; and 3, alternation in sedimentation from car¬ bonate-rich to quartz-rich. The results also provide sev¬ eral insights into the Quaternary history of the Perth Basin in southwestern Australia in regards to the alternations of carbonate /siliciclastic sedimentation in general, the control of the geometry of coastal sediment bodies by ancestral topography such as the offshore limestone ridges, the longevity of the limestone ridge ancestral topography, and the age structure of the Pleistocene coastal plains (Semeniuk 1995a). Holocene stratigraphy Much of the western edge of the Yalgorup Plain is bordered by Holocene barrier dunes. Though these are a geomorphic extension of the Leschenault Peninsula barrier (Semeniuk 1985), stratigraphically they present a different Holocene history to elsewhere. Various Holo¬ cene sediment types occur in this area and, generally, they can be assigned to previously defined formations. However, one unit remains undescribed, and is formally defined in this paper as the Preston Beach Coquina. The sediment types and their assigned formations are summarised in Table 3. Preston Beach Coquina Definition and characteristics: The Preston Beach Coquina is the name proposed for light-coloured, bedded, laminated, cross-laminated shell gravel and shelly sand along the shoreface of the Leschenault- Preston Sector, and in the subsurface. Derivation of name: Preston Beach, south of Mandurah. Type section: Preston Beach at 32°52’5 7" 115°38'40". Distribution: The unit is widespread along the coastal zone of the Leschenault-Preston Sector, and additionally forms isolated units in the subsurface in other coastal sectors further north. Thickness and geometry: The unit is up to 6-7 m thick. Regionally along the Leschenault-Preston Sector, the unit will appear as a discontinuous ribbon, some tens of kilometres long, up to 100-200 m wide and 6-7 m thick. Elsewhere it forms lenses, some 1-3 m thick, and hundreds of metres wide. Lithology: At the type section, the formation is a light- coloured (yellowish, tan, buff, to cream), bedded, crudely layered, to locally laminated, to cross-laminated deposit of shell gravel, shell grit, shelly sand, and sand. Sand grains are quartz, or quartz, bioclasts and lime¬ stone lithoclasts. The upper 1 m of the unit also has a distinctive deposit of storm debris, with Sepia spp and Spirula spirilla, mixed shell, and local horizons of pebbles of pumice. Stratigraphic relationships: The formation overlies a variety of units, depending on locality and the extent of erosion along its base: it has an erosional contact with both the Becher Sand and Leschenault Formation, and an unconformable contact with Pleistocene sediment and sedimentary rock. The formation is overlain conformably and gradationally by dune deposits of the Safety Bay Sand. Laterally, the unit may pass into shelly sand and sand of the beach facies of the Safety Bay Sand. It may also pass laterally into Becher Sand. Fossils: Molluscan shell remains in the formation include: Donax spp, Glycymeris sp, Mactra sp, and Donacilla sp, and in the uppermost part, Spirula spirula and Sepia spp. Age: Radiocarbon ages show that the unit is wholly Holocene. In the Leschenault-Preston Sector, its age is 5455 14C yrs BP to present (Semeniuk 1995b). Elsewhere, such as at Quinns Rock, it is older, but still Holocene. Distribution: The Preston Beach Coquina is a distinctive, widespread shell gravel unit formed in high energy beach to shoreface settings. The dominant, and consistent shell gravel content serves to separate it from other marine and strand units such as the Becher Sand and the beach facies of the Safety Bay Sand, respectively. Holocene sedimentary sequences In coastal southwestern Australia, there are four main large-scale standard Holocene stratigraphic sequences (Semeniuk 1995b, and Fig 4): Type 1, prograded bank, beach, beachridge system composed of a shoaling sequence of subtidal basin sediment, seagrass bank sediment, beach sand, and dune sand; Type 2, prograded shoreface to beachridge system composed of a shoaling sequence of shoreface shell /sand, beach sand/ shell, and dune sand; Type 3, retrograded barrier dune system composed of dune sand overlying estuarine lagoonal sediment; and Type 4, retrograded barrier dune system composed of dune sand unconformable on Pleistocene sediment. Each type tends to be localised in a specific sector of coast. Type 1 occurs in settings of cuspate forelands be¬ hind barrier limestone reefs, ridges and islands, such as the Rockingham- Becher system (Searle et al. 1988) and the Whitfords Cusp (Semeniuk & Searle 1986). Type 2 occurs along the modern shore face of the Leschenault- Preston barrier (e.g. Preston Beach). Type 3 occurs in the Leschenault Peninsula area (Semeniuk 1985), and Type 4 occurs in the northern part of the Leschenault-Preston Sector and in coastal sectors further north (Semeniuk et al. 1989). Barrier dune stratigraphy, Yalgorup area The standard Holocene stratigraphic sequences have been used to help unravel the Holocene history of the barrier dunes in the Yalgorup area, and specifically be¬ tween Preston Beach and the Bouvard Reefs. The Holo¬ cene stratigraphy under the barrier dunes here is complex; 77 Journal of the Royal Society of Western Australia, 78(3), September 1995 o 1 - 2km E o X o CL CL CO Figure 4. The four standard Holocene stratigraphic sequences in the southern Perth Basin (from Semeniuk 1995b). 78 Journal of the Royal Society of Western Australia, 78(3), September 1995 the marine Holocene record begins with a Type 1 se¬ quence, which is sharply overlain by a Type 3 sequence, with a final capping of Type 2 sequence (Semeniuk 1995b). The sedimentary sequence records dramatic coastal changes associated with rising Holocene sealevels; simple seagrass bank, beach, beachridge sedi¬ mentation/progradation was abruptly terminated, and succeeded by development of a barrier dune with its associated barred lagoon. The sequence developed is in¬ terpreted to be a response to seas rising into a bathy- metrically complex area. The area between Preston Beach and Bouvard Reefs, being in a transition zone be¬ tween a coastal sector of complex bathymetry to the north and one of simple bathymetry to the south, had the potential to record differing styles of Holocene sedi¬ mentation in response to varying sealevel (Semeniuk 1995b). References Commander D P 1988 Geology and hydrogeology of the superficial formations and coastal lakes between Harvey and Leschenaut inlets (Lake Clifton Project). Geological Survey of Western Australia Professional Papers, Report 23, 37-50. Druce E C 1963 Devonian and Carboniferous conodonts from the Bonaparte Gulf Basin, northern Australia, and their use in international correlation. Bureau of Mineral Resources Bulletin 98. Dunham R J 1962 Classification of carbonate rocks according to depositional texture. American Association of Petroleum Geologists Memoir 1: 108-121. Fairbridge R W 1953 Western Australian Stratigraphy. Text Book Board Publication, University of Western Australia, Perth. Glassford D K & Semeniuk V 1990 Stratification and disconformities in yellow sands of the Bassendean and Spearwood Dunes, Swan Coastal Plain, southwestern Australia. Journal of the Royal Society of Western Australia 72:75-93 Gozzard J R 1983 Fremantle Part Sheets 2033 I and 2033 IV, Perth Metropolitan Region, 1:50,000 Environmental Geology Series, Geological Survey of Western Australia, Perth. Gozzard J R 1986 Perth, Sheet 2034 II and part 2034 III and 2134 III, Perth Metropolitan Region, 1:50,000 Environmental Geology Series, Geological Survey of Western Australia, Perth. Klenowski G 1976 Geotechnical properties of the Coastal Limestone in the Perth Metropolitan area. Geological Survey of Western Australia, Annual Report for 1975, 42-46. Logan B W, Read J F & Davies G R 1970 History of carbonate sedimentation, Quaternary epoch, Shark Bay, Western Australia. American Association of Petroleum Geologists Memoir 13:38-84. McArthur W M & Bartle G A 1980 Soils and Land Use Planning. In: The Mandurah Bunbury Coastal Zone, Western Australia. Land Resources Management Series No 6, CSIRO, Melbourne. McArthur W M & Bettenay E 1958 The soils of the Bussleton area, Western Australia. CSIRO Divisional Report 3/58. McArthur W M & Bettenay E 1960 The development and distribution of the soils of the Swan Coastal Plain, W.A. Soil Publication No 16, CSIRO, Melbourne. Playford P E & Low G H 1972 Definitions of some new and revised rock units in the Perth Basin. Geological Survey of Western Australia, Annual Report for 1971, Perth, 44-46. Playford P E, Cockbain A E & Low G H 1976 Geology of the Perth Basin, Western Australia. Western Australia Geological Survey Bulletin 124, Perth. Roberts D & Wells F 1981. Seashells of Southwestern Australia. Creative Research, Perth. Searle D J & Semeniuk V 1985 The natural sectors of the inner Rottnest Shelf coast adjoining the Swan Coastal Plain. Journal of the Royal Society of Western Australia 67:116-136 . Searle D J, Semeniuk V & Woods PJ 1988 Geomorphology, stratigraphy and Holocene history of the Rockingham - Becher plain. Journal of the Royal Society of Western Aus¬ tralia 70:89-109 Semeniuk V 1983 The Quaternary stratigraphy and geological history of the Auslralind-Leschenault area. Journal of the Royal Society of Western Australia 66:71-83. Semeniuk V 1985 The Age Structure of a Holocene Barrier Dune System and its implication for sealevel history reconstructions in Southwestern Australia. Marine Geology 67:197-212. Semeniuk V 1990 The geomorphology and soils of Yoongarillup Plain, in the Mandurah-Bunbury coastal zone, southwestern Australia: a critical appraisal. Journal of the Royal Society of Western Australia 73:1-7. Semeniuk V 1995a Pleistocene coastal palaeogeography in south¬ western Australia - carbonate and quartz sand sedimentation in cuspate forelands, barriers and ribbon shoreline deposits Journal of Coastal Research (in press). Semeniuk V 1995b An early Holocene record of rising sealevel along a bathymetrically complex coast in southwestern Australia Marine Geology (in press). Semeniuk V & Glassford D K 1987 Origin of limestone lenses in Perth Basin yellow sand. Journal of the Royal Society of Western Australia 70:35-47 Semeniuk V & Johnson D P 1982 Recent and Pleistocene beach and dune sequences, Western Australia. Sedimentary Geol¬ ogy 32:301-328. Semeniuk V & Meagher T D 1981 The geomorphology and surface processes of the Australind - Leschenault Inlet coastal area. Journal of the Royal Society of Western Australia 64:33-51. Semeniuk V & Searle D J 1985 The Becher Sand, a new stratigraphic unit for Holocene sequences of the Perth Basin. Journal of the Royal Society of Western Australia 67:109-115. Semeniuk V & Searle D J 1986 The Whitfords Cusp - its geomorphology, stratigraphy and age structure. Journal of the Royal Society of Western Australia 68:29-36. Semeniuk V & Searle D J 1987 The Bridport Calcilutite. Journal of the Royal Society of Western Australia 70:25-27. Semeniuk V, Cresswell I D & Wurm PAW 1989 The Quindalup Dunes: the regional system, physical framework and vegetation habitats. Journal of the Royal Society of Western Australia 71:23-47 Wells F E & Bryce C W 1985. Seashells of Western Australia. Western Australian Museum, Perth. 79 Journal of the Royal Society of Western Australia, 78:81-87, 1995 Seagrass communities in Exmouth Gulf, Western Australia: A preliminary survey L J McCook, D W Klumpp & A D McKinnon Australian Institute of Marine Science, Cape Ferguson, PMB 3, Townsville MC QLD 4810 Manuscript received July 1995; accepted October 1995 Abstract A preliminary survey of seagrass communities in Exmouth Gulf, Western Australia, found very low abundances of seagrasses. This seems surprising given the abundance of seagrass beds elsewhere in northern and western Australia, and the highly productive prawn fishery in the gulf; prawn fisheries are usually associated with seagrass systems. Quantitative and qualitative survey of 64 sites, mainly in the inaccessible south and east of the gulf, in September 1994, indicated that seagrasses were neither extensive nor abundant. Percent covers were rarely over 5-10%. Predominant seagrasses were species of Cymodocea Konig, at depths of 0-5 m, and Halodule Endlicher in intertidal areas. Species recorded were largely in accordance with published distribution ranges. Subjective assessments indicate that epiphytic and epilithic ephemeral macroalgae contribute significant amounts of production, compared to seagrasses. The lack of extensive seagrass beds is considered in terms of the physical environment of the gulf, and in terms of the carbon source for the highly productive prawn fishery. Despite the ecological and economic importance of seagrasses, this survey is only the second published account of seagrasses for the coast between Shark Bay, Western Australia, and the Gulf of Carpentaria. Introduction Seagrass beds are of major ecological and economic importance to coastal ecosystems, particularly in tropical Australia (e.g. Larkum et al. 1989; Poiner el al. 1989; Pollard et al. 1993); yet very little is known about seagrasses throughout the north-west of Australia. There has only been one published survey (Walker & Prince 1987) of seagrass communities between the Gulf of Carpentaria and Shark Bay (Western Australia), roughly a quarter of Australia's coastline. In north-east and western Australia, seagrass beds are important in terms of area (Kirkman & Walker 1989; Lee Long et al. 1993; Coles et al. 1989) and primary production /standing crop (Walker et al. 1988; Walker 1989; more generally Klumpp et al. 1989; Hillman et al. 1989) , and as nursery habitats for commercial and recre¬ ational fisheries. Seagrass beds in Shark Bay, about 450 km south of Exmouth, are amongst the most extensive, pro¬ ductive and diverse described, with very high standing crops (Walker et al 1988; Walker 1989; Kendrick et al. 1990) . In the Gulf of Carpentaria, Torres Strait and Great Barrier Reef regions, considerable work has emphasized the extent of seagrass beds and their importance as prawn and fish nurseries (Coles et al. 1987, 1989, 1993a,b; Poiner et al. 1989; Lee Long et al. 1993). In north-west Australia, Walker & Prince (1987; also Prince 1986, pers. comm.) qualitatively surveyed seagrass species distributions at a wide range of locations. Seagrass diversity was found to be high, with an overlap of southern temperate and tropical Indo-Pacific species. These distributions were critical to interpretations of a centre of seagrass diversity and speciation in New © Royal Society of Western Australia 1995 Guinea (Walker & Prince 1987; Kirkman & Walker 1989). More detailed surveys of some north-west shelf marine habitats have been carried out for the oil and gas indus¬ tries, as far south as Exmouth Gulf, but this information is proprietary (R Hilliard, LeProvost, Dames & Moore Ltd, pers. comm.). Poiner et al. (1989) identified an urgent need for information on north-west Australian seagrasses. The present study describes seagrass community composition in Exmouth Gulf, north-west Australia, with an emphasis on the relatively inaccessible and shallow east coast. Exmouth Gulf is a large shallow basin set in a remote, arid tropical area (22°S), enclosed by the Cape Range on the west, and by extremely arid plains to the east (Fig 1). The west coast is largely sand beaches and most of the east coast has a narrow fringe of mangroves bordering extensive salt flats (up to 10 km wide). Rainfall and river runoff in the area are extremely low, and depend on rare floods resulting from cyclones. There is very little published information on marine habitats in Exmouth Gulf, despite its area (~ 3,000 km2) and the presence of a highly productive prawn fishery (> 1,000 tonnes p.a., R Watson, Fisheries Department of WA 1991 and pers. comm.) and several oil exploration projects in the area (WA Environmental Protection Agency 1991a, b). There is very little information on the east coast of the Gulf, which is inaccessible by road (Start & McKenzie 1992). Methods The survey, which took place between 24 and 30 Sep¬ tember 1994, focused on five areas approximately regu¬ larly spaced on the eastern and southern coasts of the Exmouth Gulf, and one area on the west coast (Fig 1). 81 Journal of the Royal Society of Western Australia, 78(3), September 1995 Tent Island Simpson Island Whalebone. Island Islam*6 Islets Giralia't Bay Jg Saltflats t14"45’E Figure 1. Map of Exmouth Gulf, Western Australia, showing survey sites (•). Numbers indicate sites for quantitative estimates of community composition made on 50 m transects. On the inset, SB and GC indicate Shark Bay and the Gulf of Carpentaria, respectively. 82 Journal of the Royal Society of Western Australia, 78(3), September 1995 We combined qualitative observations, intended to de¬ scribe type and extent of benthic vegetation, with sys¬ tematic quantitative sampling in areas where seagrasses were most abundant. This approach was an adaptation of planned systematic sampling in response to low abundances. Where seagrasses were absent or extremely sparse, quantitative data were not collected. Qualitative spot checks also indicated the representivity of the quan¬ titative data. The west coast site was immediately oppo¬ site Norcape Lodge, Exmouth township. The survey fo¬ cused on the extensive shallow and tidal areas of the southern and eastern gulf, which have little or no road access. Lack of time and navigation information re¬ stricted the survey to south of Tent Island. The quantitative sampling involved estimating abundance of biota at 5 replicate, randomly placed quadrats on a 50 m transect haphazardly laid in areas where spot checks indicated seagrasses to be present in measurable amounts. At each point, we estimated: (i) % cover of seagrasses (by genera, or species w'here easily distinguished in situ) and algae (in functional groups) using 100 points on a 1 m2 string grid; and (ii) density of seagrasses and any abundant macro-invertebrates using a (50 cm)2 quadrat (density estimates at site 7, a Halodule uninervis bed, used a (25 cm)2 quadrat, with data scaled accordingly). Density was counted as numbers of erect shoots (leaf groups /clusters) per quadrat. At three sites (4, 8 & 9), we did two transects separated by about 1-200 metres, to indicate variability in abundance (patchiness) at that scale. At sites 2-6, we also collected all plant matter in (50 cm)2 quadrats for biomass estimates. This material was frozen and later thawed, blotted, separated into above ground and below ground fractions, and wet weighed. Qualitative surveys involved spot checks of benthic community composition, water depth and sub¬ strate composition. Where seagrass fronds were not seen, their absence was confirmed by checking for rhizomes in the sediment. Depths were estimated from diving depth gauges and corrected relative to tidal datum by interpolation. Latitude and longitude were determined using a hand-held GPS (Global Positioning System, Motorola Trx 61000A5). Salinity (portable salinometer) and turbidity (secchi disk) were measured as part of another study (McKinnon & Ayukai in press). Since this survey focussed on seagrasses, detailed information on algae was not collected. Seagrass specimens were pressed and lodged with the Herbarium of the University of Western Australia. Results Seagrasses were neither extensive nor abundant throughout the areas surveyed. Seagrasses were rare or absent below 5 m (below datum), and even in areas with the greatest abundances, percent cover was rarely over 5% (Fig 2). Extensive areas of shallower water, espe¬ cially in the southern Gulf (Gales Bay & inshore of site 9, Fig 1) had little or no vegetation, but bare sandy or grav¬ elly substrate, often with a fine film of cyanobacteria. In the shallow mid-east of the gulf, sparse beds of Cymodocea semdata (R. Br.) Aschers. and Magnus and Cymodocea angustata Ostenfeld were common between low tide level and 5 m, but these were generally very low in biomass and often patchy. Leaf blade lengths were generally about 5 cm and always less than 10 cm. We found two dense beds of Halodule uninervis (Forsk.) Aschers. in Boissier in shallow intertidal areas near Is¬ lam Islets (site 7) and Simpson Islet (east of Burnside Islet) and a small bed of Syringodium isoetifolium (Aschers.) Dandy also near Burnside Islet. However, these were limited in extent, relative to the area of ap¬ parently similar habitat. Specific searches for roots or rhizomes in the sediments never indicated greater abun¬ dances than indicated by above ground shoots. Quantitative data on community composition (Fig 2) show that even the areas with most abundant Cymodocea beds were rarely over 5% cover (maximum cover mea¬ sured was 6%, but higher cover was noted at the north end of Tent Island). Small amounts of Halophila ovalis (R. Br.) Hook F, Halophila spmulosa (density <16 nr2), Syringodium isoetifolium (R. Br.) Aschers. (density < 144 nr2) and occasionally Halodule were present in Cymodocea beds, as were species of the algal genera Caulerpa, Halimeda, Udotea , and Penicillus, all Chlorophyta. Halodule beds had a higher % cover than Cymodocea beds, as well as a high % cover of fairly fine ephemeral algae (Fig 2). However, the mean cover (site 7) was only 21% and these beds were apparently quite limited in extent. Importantly, there were often large quantities of algae, attached to and/or entangled with the seagrass or attached to shells or subsurface rock. These algae included a wide range of groups, from fine filamentous turfs, through ephemeral epiphytes such as Hydroclathrus, Padina , Sporochnus, Dictyota, Asparagopsis, Laurencia, Dictymenia tridens, Gracilaria and Hypnea to perennial macrophytes, notably Sargassum decurrens (R Brown ex Turner) C Agardh and Sargassum spp. The ephemeral phaeophyte Hydroclathrus was particularly abundant. In some transects, the cover and biomass of these algae was greater than that of the seagrasses, despite the selection of sites on seagrass beds (Fig 2). Indeed, there were extensive, shallow subtidal areas with thick beds of these ephemeral and perennial algae, attached either epiphytically on very sparse seagrass rhizomes, or on hard substrate. Of the areas surveyed, algal beds were most abundant from Tent Island to Whalebone Island /Islam Islets. Large amounts of these species were common throughout the Gulf, both as unattached surface drift, and on the bottom as drift still attached to small fragments of shell etc. Benthic invertebrates were apparently generally sparse, although we did not sample in detail for burrowing species. Bivalves, starfish, holothurians, small gastropods, decapods and hermit crabs were observed, as well as many small corals, particularly mussids. In one area (site 1, depth 5 m) we found abundant cockles ( Anadara scapha mean density 38.2 per 0.25 m2, standard deviation 7.95), but these cockle beds were not found further south. Amphipods were common on collected plants. Small gastropods and hermit crabs were observed amongst seagrasses, but were not abundant. Plant matter was often observed trailing from burrows in the sediment. Qualitative observations indicate that the quantitative data are generally representative of areas with most abundant seagrasses. In the area from Tent Island to 83 Journal of the Royal Society of Western Australia, 78(3), September 1995 A. % Cover B. Frond Density per C. Wet Weight in grams 40 30 20 10 40 30 Cymodocea 1 2 Halophila aJa 3 4 5 6 tS * — h!h a b a b 8 9 .* - * m — * — * — * * fi9L* - *. aJa aJa 1234567 8 9 1234567 8 9 10 0 [ 1 _ g ifr-dja _ aJa 2 3 4 5 6 7 a b a b 8 9 H Caulerpa ED Ha limed a etc Q "Ephemeral" algae §!| Turf Algae | Sargassum D. Depth in m. a b a b a b 1234567 8 9 Figure 2. % Cover (A), density (B) and biomass (C) of seagrass and algal groups at a range of sites in eastern Exmouth Gulf. Depths below chart datum are shown in part (D). Site numbers (horizontal axes) refer to Fig 1. Data are mean ± se of 5 replicates. % covers of algal groups are shown as cumulative to reduce the number of graphs. "Ephemeral" algae includes non-filamentous but simple, fast growing taxa such as Hydroclathrus , Padina, Sporochnus , Asparagopsis and Dictyota. Densities were measured in 0.25 m2 quadrats, scaled to log (density m 7 + 0.4). ""Asterisks indicate taxa present but < 1% cover or density < 1 frond nv2. Biomass estimates are wet weights, divided into below ground (dark shading) and above ground (lighter shading) plant fractions. Wet weight error bars are se of below ground and total weights. Note that biomass estimates are only available for sites 2-6. Samples were not collected at the other sites. As discussed in the text, these data are for the areas with most abundant seagrasses, since quantitative data were not collected where seagrasses were even more sparse. Islam Islets and Whalebone Island (Fig 1) there were extensive algal beds in the shallow subtidal zone, often attached to hard pavement. Further south, spot checks inshore of site 9 found largely bare substrate with no seagrass, although Sargassum was present on hard substrate. Gales Bay, in the south-west, had a largely bare sandy substrate from 7 m to the intertidal, with occasional sponges, bivalves, tunicates, soft corals and small, filamentous algae and cyanobacteria. Cymodocea and Halophila were recorded, but in very low abundances. Large numbers of turtles were observed throughout shallow areas surveyed. 84 Journal of the Royal Society of Western Australia, 78(3), September 1995 On the west coast of the Gulf, near Exmouth town beach, we swam a depth profile from 10 m to the inter¬ tidal. At 10 m, seagrasses were absent but sponges, tuni- cates and seawhips, and some brown algae were present. By 8 m, brown algae (especially Sporochnus sp) were abundant, as well as invertebrates and small amounts of Cymodocea sp, Halophila ovalis and H. spinulosa. Around 5 metres, Halophila species were patchy but common, particularly in sheltered positions behind coral bommies etc. In the shallow subtidal and low intertidal, there were thick beds of Sargassitm , along with some beds of Cymodocea and a small patch of very thick Thalassodendron ciliatum (Forsk) den Hartog on rocky pavement (100% cover, but < 100 m2). Most of the intertidal zone was bare sand beach. The shallow waters of Exmouth Gulf were very turbid, with large amounts of suspended material, due to rough sea conditions and strong tidal currents. Apart from a midday lull, the Gulf was very choppy, with chop over 1 m for much of each day, due to strong land and sea breezes. Tidal currents were strong (up to 0.7 knots). Although in deeper areas (> 5 m) the bottom was generally fine sand or silt and mud, in shallow areas there was often only a thin veneer of sand or gravel over hard rocky pavement. The bottom sediments appeared to be highly mobile, particularly in shallower areas, and we observed several seagrass beds with exposed rhizomes and roots, where sediments appeared to have been washed away. The highest salinity measured in the Gulf was 38.5 %o. Discussion Seagrass species recorded in this survey were consistent with patterns found in a previous larger scale survey (Walker & Prince 1987, Table II; also W'alker pers. comm, and Prince pers. comm.), except that our record of C. serrulata at Tent Island (22°S) considerably extends the southern limit of this tropical species (Walker 1991). The only other published Western Australian record is from Sunday Island, King Sound (16°S, also recorded as drift at Carnarvon 24°45'; Walker & Prince 1987). All the species recorded have tropical affinities. This is in con¬ trast to Shark Bay where the southern temperate species Posidonia australis Hook f. and Amphibolis antarctica (Labill) Sonder & Aschers ex Ashers are common, the latter species in extensive and dense meadows (Walker 1989). Seagrass abundance in this survey was low. Cover was generally less than 5%, mean shoot densities were always less than 1,000 nr2 and often less than 100 m2, and biomass was generally less than 60-100 g wet wt m 2. In contrast, in Shark Bay, 450 km south of Exmouth, Posidonia australis and Amphibolis antarctica are extremely abundant; A. antarctica reaches biomasses of 2 kg dry wt m 2 and densities of 300-500 shoots nr2, each being up to 2 m long (Walker et al. 1988), but southerly seagrasses are often more abundant (Coles et al. 1989). Also in Shark Bay, Walker et al. (1988) recorded 600 leaves m 2 and 60-100 g dry wt m*2 of Halodule uninervis, and 30% cover of Cymodocea angustata as understorey. Just north of Exmouth Gulf, at 21° 16’ S, Walker & Prince (1987) found several hundred hectares of C. angustata at 30-50% cover. Maximum densities and dry weights of Halodule uninervis in the Great Barrier Reef region and New Guinea are also higher than those in Exmouth Gulf (Brouns 1987; Lee Long et al. 1993; also Mellors et al. 1993). Density and biomass of Cymodocea in our study were within the ranges reported for the Great Barrier Reef region (Lee Long et al. 1993), but leaves were longer in the Great Barrier Reef region (Coles et al. 1987). Bio¬ mass records for C. serrulata in New Guinea and south¬ ern Queensland (148 and >400 g dry wt nr2 respectively; Brouns 1987; Coles et al. 1989) are much higher than our values. Ratios of above and below ground biomass in Exmouth Gulf were similar to those in the literature (Brouns 1987; Moriarty & Boon 1989). The apparent lack of extensive or abundant seagrass beds in Exmouth Gulf is surprising, given the extent of apparently suitable shallow substrate, the productivity of the gulf, and the contrast with other areas of northern and western Australia. Thus, the low abundance of seagrass recorded in our survey raises three issues: (a) the representivity of the survey; (b) the reasons for the low abundance of seagrass; and (c) the source of carbon production for the Gulfs abundant prawns, turtles and dugongs. Our results should represent typical seagrass abun¬ dances throughout the southern and eastern Gulf. The survey is limited by the locations and particularly the time of sampling, and it remains possible that seagrasses are much more abundant in other areas, or perhaps at other times of year. However, it is very unlikely that all 64 sites fell in areas of particularly low seagrass abun¬ dance, especially given our bias to areas of relatively high abundance. Given the turbidity of the water and the clear decrease in abundance between 5 and 10 m, we presume seagrasses to be restricted to shallow areas of the Gulf. It is possible that the unsurveyed northern ar¬ eas of the Gulf have more seagrass, especially given the trend to very low abundances in the southern Gulf. Lack of time and navigational information precluded survey of the north-east Gulf. Qualitative surveys over the 6 months following our study by M Forde (Bowman, Bishaw & Gorham Ltd, pers comm ) largely confirmed the patterns of abundance and species composition reported here, except for seasonally abundant Halophila spinulosa in deeper (15 m) and less turbid water north of the Gulf. Within the Gulf, there is no indication of higher seagrass abundances at other seasons. We found no sign of dormant or decaying rhizome /root tissues, and July- September 1994 sampling with a grapple dragged on the bottom did not indicate higher abundance or longer fronds (D Pont, M G Kailis Fisheries, pers. comm.). Even if seagrass leaf length were to increase at other seasons, cover would remain low without a many-fold increase in density (Fig 2) and root mass. Seasonal changes in tropical seagrasses rarely exceed two-fold changes in standing crop (Mellors et al. 1993 for H. uninervis on the Great Barrier Reef; Hillman et al. 1989; Lanyon et al. 1989) and density (Brouns 1987 for H. uninervis in New Guinea). In New Guinea C. serrulata varied seasonally by only 30% (Brouns 1987). However, just north of Exmouth, Forde (Bowman, Bishaw & Gorham Ltd, pers. comm.) reported large changes in H. spinulosa above¬ ground biomass over 6 months, so seasonal differences remain possible. Longer term declines in seagrass abun¬ dance may have occurred in Exmouth Gulf (e.g. Poiner 85 Journal of the Royal Society of Western Australia, 78(3), September 1995 et al. 1989), but there has not been any parallel decline in prawn catches (R Watson, Fisheries Department of WA, 1991 & pers comm). The low abundance of seagrasses in the areas sur¬ veyed probably results from the lack of suitable sub¬ strate. Much of the bottom is either hard substrate or highly mobile coarse sediments, with very little fine silt and clay, and sediments appear to be eroding rather than accumulating (K Woolfe, James Cook University of North Qld, pers. comm.). The seagrasses may not be able to accumulate sufficient root mass to stabilise these sedi¬ ments. Neither salinity nor nutrient levels are likely to be limiting seagrasses. Salinity in the well-flushed Gulf is moderate compared to the reported tolerances (64 %o for H. unineruis, 50%o for C. angustata (Walker 1989); 45% for H. ovalis, 75%o for Halodule (Hillman et al. 1989). Freshwater input to the Gulf is negligible. The abun¬ dance of algae on hard substrates suggests that nutrients are not limiting, and the Gulf showed no sign of eu¬ trophication. Trawl disturbance is not responsible for low seagrass abundance, since the areas surveyed are designated a nursery zone (D Pont, M G Kailis Fisheries, pers. comm.), and are largely too shallow for trawling. Assuming that our results do represent seagrass abundance throughout the Gulf, it would seem surpris¬ ing that the Gulf is a highly productive region. The yearly catch of prawns ( P emeus latisulcatus, Penaeus esculentus, Metapenaeus endeavouri) has been over 1,000 tonnes for 8 of the last 10 years and is not apparently declining (R Watson, Fisheries Department of WA, 1991 & pers. comm.). Seagrass beds are important elsewhere to prawn and other fisheries, as nursery areas (Pollard 1984; Coles et ah 1987, 1989, 1993a,b; Poiner et al. 1989; Bell & Pollard 1989; Lee Long et al. 1993; Watson et al. 1993; Hill & Wassenberg 1993; less true for P. latisulcatus) and probably make major trophic contribu¬ tions to prawn production via detrital pathways (Klumpp et al. 1989). Exmouth Gulf has an estimated dugong population of 1,000 and turtles are very abun¬ dant (Prince et al. 1981; Prince 1986; Preen et ah, James Cook University of North Qld, pers. comm.; pers. obs.). Both dugongs and turtles are major seagrass grazers. Our observations suggest that the extensive algal beds are probably major primary producers in this ecosys¬ tem. Several previous studies of dense seagrass beds have found that both macroalgae and micro-algae make important contributions to biomass and productivity (Orth & Van Montfrans 1984; Kendrick et al. 1990; Klumpp et al. 1989; Borowitzka & Lethbridge 1989; Pol¬ lard Kogure 1993). Whilst we did not quantify their contribution, these algal beds apparently have very high turnover and export to deeper water. Most of the species observed are structurally simple, rapid growing, and fragile (especially Hydroclathrus). We suggest that rapid growth, high proportional breakage and losses (due to rough conditions), rapid transport by wind and tidal currents, and fast break down would make these algae major sources of detritus throughout the Gulf. Other major primary producers may include phytoplankton and salt-flat cyanobacteria. Phytoplankton production is probably not high, since standing stocks of chlorophyll a are low (McKinnon & Ayukai, in press). The extensive salt flats (5-10 km wide. Fig 1) in the high intertidal of the southern and eastern Gulf have a thick mat of cyanobacteria which may contribute significant amounts of carbon during spring tides. Algal, phytoplankton and salt flat production may thus contribute to the high yield of prawns, via detrital food chains (Klumpp et al. 1989). Algal beds may also serve as prawn nursery grounds in this area. Similarly, abundance of suitable algae could sustain large populations of green turtles, which eat both seagrass and algae. However, dugongs are believed to feed on seagrass almost exclusively (Lanyon et al. 1989; Erftemeijer et al. 1993). It is difficult to explain how an area with low seagrass abundances could apparently sustain a large dugong population. Intertidal Halodule beds, such as we observed, are preferred feeding grounds (Lanyon et al. 1989). Such beds may be more extensive than our survey indicates, or may be extensive outside our survey area. Exmouth dugongs may also consume unusual amounts of algae, since Lanyon et al. (1989) note that dugongs do eat algae when seagrasses are scarce. Given the vulnerable status of dugongs (IUCN 1990), the issue warrants further attention. In summary, our results suggest that much of the extensive shallow east and south coast of Exmouth Gulf does not have abundant seagrasses, perhaps because of a paucity of suitable substrate. This raises questions about primary production in a region with high prawn production and herbivore populations. We suggest that beds of ephemeral and perennial algae may be impor¬ tant primary producers in the region. Acknowledgements : We are very grateful to D Walker, H Kirkman and G Kendrick for assistance with identifications, advice, support and hospital¬ ity. We thank D Pont for preliminary data, local knowledge and advice. We greatly appreciate the cheerful field assistance of D Gaughan. We thank R Christie for technical assistance, and the skipper and crew of the AIMS research vessel RV Lady Basten. Helpful comments on the manu¬ script were provided by R Coles, T Done, and the reviewers. This is ATMS Publication 764. References Bell J D & Pollard D A 1989 Ecology of fish assemblages and fisheries associated with seagrasses. In: Biology of seagrasses (eds A W D Larkum, A J McComb & S A Shepherd) Elsevier, Amsterdam, 565-609. Borowitzka M A & Lethbridge R C 1989 Seagrass epiphytes. In: Biology of seagrasses (eds A W D Larkum, A J McComb & S A Shepherd) Elsevier, Amsterdam, 458-499. Brouns J J W 1987 Aspects of production and biomass of four seagrass species (Cymodoceae) from Papua New Guinea. Aquatic Botany 27:333-362. Coles R G, Lee Long W J, Squire B A & Bibby J M 1987 Distribution of seagrasses and associated juvenile commercial penaeid prawns in north-eastern Queensland waters. Australian Journal of Marine and Freshwater Research 38:103- 19. Coles R G, Lee Long W J, Watson R A, Derbyshire K J & Done T 1993a Seagrass research in tropical north eastern Australia. In: International Workshop on Seagrass Biology, Komato 1993 (ed K Aioi). Ocean Research Institute, University of Tokyo, Tokyo, 14-27. Coles R G, Lee Long W J, Watson R A & Derbyshire K J 1993b Distribution of seagrasses, and their fish and penaeid prawn communities, in Caims Harbour, a tropical estuary, northern Queensland, Australia. Australian Journal of Marine and Freshwater Research 44:193-210. 86 Journal of the Royal Society of Western Australia, 78(3), September 1995 Coles R G, Poiner I R & Kirkman H 1989 Regional studies- seagrasses of north-eastern Australia. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shep¬ herd). Elsevier, Amsterdam, 261-278. Erftemeijer P L A, Moka D & Moka W 1993 Stomach content analysis of a dugong ( Dugong dugon ) from South Sulawesi, Indonesia. Australian Journal of Marine and Freshwater Research 44:229-233. Hill B J & Wassenberg T J 1993 Why are some prawns found in seagrass? An experimental study of brown ( Penaeus esculent us) and grooved (P. semisulcatus ) tiger prawns. Australian Journal of Marine and Freshwater Research 44: 221-228. Hillman K, Walker D I, Larkum A W D & McComb A J 1989 Productivity and nutrient limitation. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 635-685. IUCN 1990 IUCN Red List of Threatened Animals. International Union for Conservation of Nature and Natural Resources- World Conservation Union, Gland. Kendrick G A, Huisman J M & Walker D I 1990 Benthic macroalgae of Shark Bay, Western Australia. Botanica Marina 33:47-54. Kirkman H & Walker D I 1989 Regional studies- Western Australian seagrass. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 157-181. Klurnpp D W, Howard R K & Pollard D A 1989 Trophodynamics and nutritional ecology of seagrass communities. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 394-457. Lanyon J, Limpus C J & Marsh H 1989 Dugongs and turtles; Grazers in the seagrass system. In: Biology of Seagrasses (ed A W D Larkum, A J McComb & S A Shepherd) Elsevier, Amsterdam, 610-634. Larkum A W D, McComb A J & Shepherd S A 1989 Biology of Seagrasses. Elsevier, Amsterdam. Lee Long W J, Mellors J E & Coles R G 1993 Seagrasses between Cape York and Hervey Bay, Queensland, Australia. Austra¬ lian Journal of Marine and Freshwater Research 44: 19-31. McKinnon A D, & Ayukai (in press) Copepod egg production and food resources in Exmouth Gulf, Western Australia. Journal of Marine and Freshwater Research, in press. Mellors J E, Marsh H & Coles R G 1993 Intra-annual changes in seagrass standing crop, Green Island, northern Queensland. Australian Journal of Marine and Freshwater Research 44: 33- 42. Moriarty D J W & Boon P I 1989 Interactions of seagrasses with sediments and water. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 500-535. Orth R J & Van Montfrans J 1984 Epiphyte-seagrass relationships with an emphasis on the role of micrograzing: A review. Aquatic Botany 18:43-69. Poiner I R, Walker D I & Coles R G 1989 Regional studies- seagrasses of tropical Australia. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 279-303. Pollard P C 1984 A review of ecological studies on seagrass-fish communities, with particular reference to recent studies in Australia. Aquatic Botany 18:3-42. Pollard P C & Kogure K 1993 The role of epiphytic and epibenthic algal productivity in a tropical seagrass, Syringodium isoetifolium (Aschers) Dandy, community. Australian Journal of Marine and Freshwater Research 44:141- 154. Pollard P C, Kioke I, Mukai H & Robertson A 1993 Tropical seagrass ecosystems; Structure and dynamics in the Indo- West Pacific. Australian Journal of Marine Inland Freshwater Research 44. CSIRO, Australia. Prince R I T 1986 Dugong in northern waters of Western Australia 1984. Department of Conservation and Land Management, WA. Technical Report 7. Prince R I T, Anderson P K & Blackman D 1981 Status and distri¬ bution of dugongs in Western Australia. In: The Dugong. Workshop Proceedings (ed H Marsh). James Cook University of North Queensland, Townsville, 67-87. Start T & McKenzie N 1992 East of the Gulf. Landscope 8(2): 41- 46. WA Environmental Protection Agency 1991a Proposed exploration drilling on the Cody, Spider and Cassidy projects, Exmouth Gulf. Western Australia. Environmental Protection Agency, Perth. WA Environmental Protection Agency 1991b Proposed offshore pretroleum exploration drilling in EP325, North West Shelf. Western Australia. Bulletin No 582. Environmental Protection Agency, Perth. Walker D I 1989 Regional studies- seagrass in Shark Bay, the foundations of an ecosystem. In: Biology of Seagrasses (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 182-210. Walker D I 1991 The effect of sea temperature on seagrasses and algae on the Western Australian coastline. Journal of the Royal Society of Western Australia 74:71-77. Walker D I & Prince R I T 1987 Distribution and biogeography of seagrass species on the northwest coast of Australia. Aquatic Botany 29:19-32. Walker D I, Kendrick G A & McComb A J 1988 The distribution of seagrass species in Shark Bay, Western Australia, with notes on their ecology. Aquatic Botany 30:305-317. Watson R 1991 The state of the fisheries, December 1991 Report. Fisheries Department of Western Australia, Perth. Watson R A, Coles R G & Lee Long W J 1993 Simulation estimates of annual yield and landed value for commercial penaeid prawns from a tropical seagrass habitat, northern Queensland, Australia. Australian Journal of Marine and Freshwater Research 44:211-220. 87 Journal of the Royal Society of Western Australia, 78:89-90, 1995 The Royal Society of Western Australia Medallist, 1995 Emeritus Professor A R (Bert) Main A R (Bert) Main, BSc(Hons), PhD, DSc, CBE, FAA The Medallist for 1995, Emeritus Professor A R (Bert) Main, was elected by the Council of The Royal Society of Western Australia because of his contributions to the scientific knowledge of the natural history of Western Australia. The Medal was presented at the Royal Society of Western Australia Medal Lecture, held in the "old" Zoology Building at The University of Western Austra¬ lia, by Professor Alan Robson, Deputy Vice-Chancellor of The University of Western Australia, on Monday 20 November, 1995. Professor Albert Russell (Bert) Main was born in Perth in 1919. He grew up on a vineyard in the Swan Valley and attended Midland Junction Central School. Following a period of employment in the State Public Service, during which he studied accountancy and later towards matriculation at night school, he joined the CMF in 1939 and later AIF and saw service within Aus¬ tralia before joining the RAAF. After qualifying, he served as a navigator of Lancaster bombers in Europe, seeing out the last part of the war as a POW in Ger¬ many. Upon returning to Australia he matriculated and be¬ gan study at UWA financed by the Commonwealth Re¬ construction Training Scheme (CRTS). He graduated in Geology and with 1st class Honours in Zoology. In 1950, he was awarded a Fulbright Scholarship to the Univer¬ sity of Chicago. He returned from Chicago to take up an appointment in the Department of Zoology at the Uni¬ versity of Western Australia in 1952, where he received © Royal Society of Western Australia 1995 his Ph. D. degree in 1956. He has an Honorary Doctorate of Science from the University of Western Australia. In 1958, Professor Main was awarded a Carnegie Travelling fellowship to the University of California at Los Angeles, in the United States of America. He was elected a fellow of the Australian Academy of Science in 1969. He received the Britannica Australia Award with Professor Harry Waring in 1970. He was elected an ANZAAS Fellow and later awarded the Mueller Medal of ANZAAS and Gold Medal of Australian Ecological Society. He was made Commander Civil Division of the Order of the British Empire CBE in 1981. Currently, Professor Main is a member of Society for Conservation Biology; Ecological Society of Australia; Geological Society of Australia; Australian & New Zealand Association for the Advancement of Science and is an Honorary Foreign Member of the American Society of Ichthyologists & Herpetologists. Professor Main is broadly interested in natural his¬ tory, with a special interest in the ecology of animals in arid situations. He has had a long standing interest in environmental management and conservation. He served on the Fauna Advisory Committee and later the West Australian Wildlife Authority and was a founda¬ tion member and later Chairman of the Western Austra¬ lian Environmental Protection Authority. He was also president of the Zoological Gardens Board. He has been a member of the Council of the Australian Institute of Marine Science in Townsville, and President of the National Parks Authority, and was a member of the Australian Universities Commission 1971-1977. 89 journal of the Royal Society of Western Australia, 78(4), December 1995 The Royal Society of Western Australia Medal Recipients 1924-1995 The Medal of The Royal Society of Western Australia was instituted in 1924 to mark the centenary of the birth of Lord Kelvin (26 June, 1825). The Royal Society Medal (originally referred to as the "Gold Medal" and then subsequently sometimes as the "Kelvin Medal' due to the association of the inaugural award with the centen¬ nial Kelvin celebration and because the medal bears in relief on its obverse side the head of Kelvin) is awarded approximately every four years for distinguished work in science connected with Western Australia. The original dye for the medal, first struck in 1924, remains in the safe-keeping of The Society. The first three medals were struck in gold, and all subsequent medals in silver. The first medallist of The Royal Society was Dr Wil¬ liam J. Hancock, Government Electrical Engineer and Honorary Medical Radiographer at Perth Hospital, who in 1924 was the recipient of the Medal of The Royal Society of Western Australia, in recognition of his pio¬ neering work in the medical application of X-rays. The most recent medallist. Emeritus Professor A. R. (Bert) Main, was presented with the eighteenth Medal of The Royal Society of Western Australia, in recognition of his contributions to zoology. Recipients of the Medal of The Royal Society of West¬ ern Australia are: (reference to Journal of the Royal Society notice of medal award in parentheses) 1924 Dr W J Hancock: radiography; medical application of x-rays (10:xvii) 1929 Dr E S Simpson: mineralogy and geology of Western Australia (15:iv) 1933 Mr W M Came: plant pathology; the bitter pit of apples (19:xi) 1937 Mr A Gibb Maitland: Pilbara survey and artesian water supplies (23:xi) 1941 Prof E de C Clarke: geology of Western Australia (27:v) 1945 Mr L Glauert: natural sciences (31:vi) 1949 Mr C A Gardner: botany, the flora of Western Australia (35:v) 1955 Dr H W Bennetts: veterinary science; live stock diseases (40:1) 1959 Prof E J Underwood: animal nutrition and husbandry (43:67) 1966 Mr C F H Jenkins: agricultural entomology and natural history (49:91) 1970 Prof R T Prider: geology; petrology and mineralogy (53:95) 1979 Prof R M Berndt: anthropology; aboriginal studies (63:29) 1979 Emer Prof B J Grieve: botany; ecophysiology and the flora of WA (63:29) 1979 Dr D L Serventy: zoology; ornithology and nature conservation 1983 Dr J S Beard: botany; vegetation classification and mapping (65:93) 1986 Prof C A Parker: soil biology 1993 Prof J R de Laeter: geophysics and geochronology (77:4) 1995 Emer Prof A R Main: zoology; ecology and nature conservation (78:89) 90 Journal of the Royal Society of Western Australia, 78:91-98, 1995 The Royal Society of Western Australia Medallist Lecture, 1995 The study of nature - A seamless tapestry A R Main Department of Zoolology, University of Western Australia, Nedlands WA 6907 Manuscript received December 1995 I would like to thank the Royal Society of Western Australia for their recognition of my research contributions. In a sense I am astonished that these intellectually enjoyable activities, undertaken out of my own curiosity, should be acknowledged by the award of the prestigious Royal Society Medal. I also thank Professor Alan Robson, Deputy Vice Chancellor of the University of Western Australia for personally presenting the medal. The ready and active participation of graduate students and colleagues ensured the success of all projects and I thank them sincerely. Introduction From time to time there have been inferences that my interests were too broad, my application undirected or that I was excessively opportunistic in my research. For example, when I was still involved with frog ecol¬ ogy and beginning studies of macropods, I was told by one worker that I could only expect one good idea in my life and I should stick with it. On a later occasion, another suggested that I had so many ideas that I must have stolen them. I take my cue for this lecture from these comments and briefly review and interpret the findings on the topics investigated by myself and co¬ workers to show that I did indeed pursue a theme of adaptations to aridity which led along many unexpected pathways. My research has been devoted to answering questions posed by problems arising from field observations. These were initiated and conditioned by my experiences as a child when I lived within walking distance of the bush. From the time I could walk a reasonable distance, I was taken to the bush on most weekends by my grand¬ father and natural history mentor. I have an abiding memory of puzzlement at the contrasts between sum¬ mer and winter and of wondering how so many animals managed to persist under such conditions. Briefly I was, and still am, interested in understanding the world about me. Natural History observations exposed the changes that took place but were unhelpful, except in so far as they posed testable hypotheses, in resolving the underlying mechanisms of persistence of animals or how such mechanisms may have evolved. In retrospect I now see these twin topics as motivating my curiosity about species, their adaptive traits and evolution. It was only after war service and the benefits of the Commonwealth Reconstruction Training Scheme that I was able to gain the necessary technical skills with which to pursue my childhood interests. Parentheti¬ cally, I may comment that upon my appointment to the Zoology Department I was asked by Professor Waring what I intended to do as research. My reply embraced the points mentioned above; as soon as I mentioned evo¬ lution his comment was 'It's all known, there's nothing © Royal Society of Western Australia 1995 to research in that'. Fortunately he did not attempt to stop me trying. In what follows, I set out the groups of animals on which work was conducted to show the scope of the studies and to illustrate two things; 1. The interdependence of studies when interpreting observations. 2. The relevance of these studies to a wider understand¬ ing of local biology, and how the studies led into societal issues such as con¬ servation. I devote much space to discussing work on identification of species of frogs, the group I had chosen for comparative studies of their adaptations to aridity. But, comparative studies depend on objective identifica¬ tion of the species being compared, and at the start of the investigation it was not clear that systematists used objective criteria to identify Western Australian species of frogs. I present the sorting out of the problems of species' identification in some detail because it illustrates the difficulties associated with obtaining an adequate, objective, and biologically-meaningful systematic and taxonomic knowledge on which to base comparative studies of adaptation to aridity. It also explains why I have in many quarters been seen as a herpetologist rather than as an ecologist. Frogs Harry Waring had come to Western Australia to study marsupials, especially their reproductive physiol¬ ogy, and he chose to work with the quokka Setonix brachyurus . This species was common on Rottnest Is¬ land, and it and other marsupials that Waring and his students studied were morphologically distinct, readily identified, and posed no problems when establishing the species status of the population upon which the work was being done. This was not necessarily the case with other easily collected animals. Prior to coming to Western Australia, Waring had worked on colour change using the South African clawed toad (Xcnopus lacvis) as an experimental animal and he was naturally favourably-inclined to frogs as ex¬ perimental animals. Moreover, while there were many named species of frogs occurring around Perth, it was 91 Journal of the Royal Society of Western Australia, 78(4), December 1995 not apparent that the named kinds had the same status as species that was enjoyed by marsupials. In particu¬ lar, the last two workers who had reviewed Australian frog taxonomy (Loveridge 1935, and Parker 1940) held markedly different views. This was unfortunate because I had formed the view that because of the unprotected nature of their skin, frogs should be useful organisms in evolutionary studies as it was believed, at the time, that speciation in some of the biota had been promoted by a great aridity in the geologically recent past (Crocker & Wood 1947). If this were so, then adaptations to drought and their evolution might be linked to the same phe¬ nomena. But, it was necessary to first be certain that named species had a biological reality in the field and this was certainly not so as the following illustrates In his revision of the Australian frogs of the family Leptodactylidae Parker (1940) used the flash colours in the groin to distinguished the two Crinia species, georgiana and signifera. These flash colours were said to be strongly developed and fast in preserved specimens of C. georgiana or absent or fugitive in C. signifera. How¬ ever, when preserving specimens I had noticed that colours remained fast in either formalin or alcohol when live specimens were preserved but faded when speci¬ mens were dead before preservation. On the other hand, Loveridge (1935) believed absence of colour was indica¬ tive of C. georgiana whose dorsal pattern was rugose or with a few scattered wrarts and raised undulating glan¬ dular folds. In contrast, he classified C. signifera into two sub-species, C. signifera affinis when the back pat¬ tern was perfectly smooth without the lyre shape or C. signifera ignita in which a lyre-shaped dorsal fold was present. Clearly this suggested that the names assigned by museum taxonomists to species were unlikely to re¬ flect the presence of field entities and gave no confi¬ dence that specimens agreeing with literature descrip¬ tions could be assigned names which represented bio¬ logical entities which may show adaptive traits to the local environment. Fortunately, the biological species concept had re¬ cently been proposed (Mayr 1949) and Moore's work showed the genetic inviability and hence the presence of cryptic or sibling species, within what was formerly re¬ garded as Ram pipiens (Moore 1946, 1949). I already knew from observations of the many distinctive call types among local frogs that if distinct calls were as¬ sumed to reflect a specific mate identification, then pu¬ tative species could be identified by collecting calling males which would represent different breeding, and hence genetically distinct, entities. The genetic distinct¬ ness of these entities could be confirmed by making crosses using Moore's technique of in vitro fertilisation, with appropriate controls. Normal development would be interpreted as being indicative of a within species cross, while abnormal development in the experimental cross and normal development in the control would in¬ dicate a cross between species. The sorts of abnormali¬ ties which occurred with crosses between different call types of Crinia were failure to develop beyond early cleavage stages, or embryos that hatched but were headless, haploid or with extruded guts. Clearly, populations used in the crosses and identified by their characteristic calls were not capable of sharing a common gene pool and hence could not be considered as biological species in an evolutionary sense. However, before an experimental programme could be undertaken, an objective method of recording and measuring of calls was needed. Fortunately, the Austra¬ lian Broadcasting Commission co-operated by loan of equipment and running a programme on the distinctive¬ ness of frog calls (Anon 1953). I was able to use these tapes to show by means of oscillograph analysis that different calls were distinctive in duration, repetition fre¬ quency, and pulse rate. It was now only necessary to build a portable tape recorder and the research could commence. The construction was undertaken by Murray Littlejohn, then a doctoral student, and he was able to show that a number of distinct call types could be identified in the genus Crinia and I was able to con¬ firm the species status of these by means of in vitro crosses. At the same time Tony Lee confirmed the spe¬ cies status of different call types (Littlejohn & Main 1959) within the genus Heleioporus , which are all large bur¬ rowing frogs. A new species was also identified within the dry land burrowing frog genus Neobatrachus. At this stage, the conflict between Parker and Loveridge on the characters to be used in distinguishing C. signifera and C. georgiana had been resolved, the species status of C. glauerti had been confirmed, and sibling species of Crinia , Heleioporus and Neobatrachus identified (see Main 1968 for review and literature citations), I might comment here that the recognition and nam¬ ing of species using biological characteristics was not readily accepted. The substance of the objections could be reduced to two elements, species were those entities identified by recognised taxonomists, or that species should be only diagnosed by characters that were avail¬ able to museum curators with access only to preserved specimens. Once the species status of specimens could be estab¬ lished, it was possible to readily identify calling species in the field and so be certain of the status of collected specimens. Only then could meaningful experimental comparisons be made in the adaptive responses of the species to aridity and drought. This work was done in collaboration with Peter Bentley (Bentley et al. 1958; Main & Bentley 1964; Bentley & Main 1972a, b). In the midst of this work on water balance in frogs, lizards, birds and macropods, a physiologist suggested to me that it was quite improper to prostitute physiology as a field study in order to interpret the ecology of species. The progeny arising from the controls from the in vitro crosses of the various call types of Crinia were reared to metamorphosis and these showed that the dor¬ sal pattern so emphasised as a taxonomic character by Loveridge (1935) was an inherited trait (Main 1965). The early work using the mating calls of males to identify species was the subject of much comment, of which the essentials were; we all know that frogs are cold-blooded vertebrates and cannot regulate their body temperature; moreover, call characteristics such as pulse rate and repetition frequency will be dependent, as in all cold-blooded animals, on body temperature; since body temperature cannot be regulated, then calls cannot be as constant and characteristic as is claimed. However, whenever calling males were collected for use in experi- 92 Journal of the Royal Society of Western Australia, 7 8(4), December 1995 mental crosses, body temperatures were recorded and when collated these records showed specific seasons of calling and body temperature ranges for each species (Main et al. 1959, figs 2 and 3 ). These doubts about my field-based approach pro¬ duced benefits for me because Waring induced J A Moore of Columbia, then a visiting Fullbright Scholar in Australia, and later Th Dobzhansky, also of the same institution, to visit and spend time with me in the field. Waring was also instrumental in having Julian Huxley visit and accompany me in the field. I benefited im¬ mensely from these occasions and the opportunities I had to discuss evolutionary problems with them in the field. I doubt that a higher level or more intense peer review could be achieved. I was funded in this research by University of West¬ ern Australia Research Grants, and administrators then, as now, were interested in accountability as the follow¬ ing illustrates. Many stores and fuel outlets in the coun¬ try would not accept University purchase orders because of the perceived paper work and the delays in obtaining cash recoups. Thus I frequently had to pay cash and obtain a recoup upon my return. On one occasion, after an extensive field trip to the goldfields and pastoral ar¬ eas, I was told that there were plenty of frogs around Perth and there was no need to travel, as I had done, to distant places to get them. I answered this criticism and was told not to be impertinent. When it comes to rebuk¬ ing administrators who desire to limit or set boundaries of research, I hope I have not mellowed over time. The frog work demonstrated the significance of behavioural avoidance of stressful situations as a complement to physiological adaptations. Clearly, the basis of the adaptations had been revealed and further work would involve laboratory analysis and skills which I lacked. The frog work was then reviewed and summarised by me (Main 1968). Other Species (Reptiles and Birds) Field work with frogs and macropods is primarily a nocturnal activity, and during the days 1 took every op¬ portunity to measure the body temperatures of lizards caught in the field. During the 1960's W R Dawson, accompanied by P Licht and V Shoemaker, spent a year in Western Australia studying the temperature toler¬ ances and biology of lizards. The data which I had already collected was assembled, and added to by them, to reveal that each species exhibited an activity pattern related to a specific temperature range (Licht et al. 1966). Prior to this S D Bradshaw had begun a comparative study on Amphibolurus ornatus { Ctenophorus ornatus ) which showed how intricate relationships between diet, water metabolism, electrolyte balance and behaviour led to individual survival and population persistence (Bradshaw & Main 1968). In a comparative study of three species of chat (. Ephthianura ) which live in progressively drier habitats, G K Williams showed that behavioural and physiologi¬ cal traits changed in a way that paralleled the environ¬ mental changes associated with the different habitats oc¬ cupied (Williams & Main 1976, 1977). Macropods While the foregoing research on frogs was proceed¬ ing, Professor Waring was in receipt of supporting funds from CSIRO for his studies of reproduction of marsupi¬ als. These studies concentrated on the reproductive physiology of the quokka (Setonix brachy tints) and the Tammar wallaby ( Macropits eugenii), and it was during this time that studies on the ecology of marsupials was begun (Main, Shield & Waring 1959). These ecological studies focused on the ability of kangaroos and walla¬ bies to survive periods of drought when water was scarce and available food was of poor nutritional qual¬ ity. A preliminary study by me showed that the nitro¬ gen levels in the gut contents of the quokka on Rottnest exceeded what could be expected even with the most favourable selection of dietary items, and suggested that the quokka might have a ruminant-like digestive system and thus would be capable of recycling urea. This sup¬ position was confirmed by Moir et al (1956). These observations led to further studies of dietary composi¬ tion, water metabolism and urea recycling in kangaroos and wallabies, Analysis of dietary composition was undertaken by G Storr, who used leaf epidermis to identify plant spe¬ cies eaten by the quokka, red kangaroo and the euro (Storr 1964, 1968) E H M Ealey, then of CSIRO Wildlife Section, had begun a study of the euro (Macropus robustus), then re¬ garded as a pest of the pastoral areas in the Pilbara. I was asked for assistance and advice based primarily on knowledge of the ecology of the quokka on Rottnest Island, which 1 regarded as a model of how a macropod adapted to aridity. I believed this to be so because, in its typical range, the quokka occupied mesic situations whereas on Rottnest its habitat was considerably more arid. Later, as the work on the euro developed, I was quizzed by the head of the Wildlife Section on what I was doing and where it might lead. I explained to him that I believed that an understanding of the water and dietary needs of the euro were basic to any understand¬ ing of the ecology of the species in the arid area in which it occurred. From his comments I was left in no doubt that what I was proposing was a nonsensical vision, certainly lacking in any basis in classical ecology and with no hope of contributing to an understanding of field observations. Moreover, it was likely to be unpro¬ ductive in the sense that it would not lead to an under¬ standing of how kangaroos could live in such an inhos¬ pitable environment and particularly how the species could achieve such a pest status when available forage could not support sheep. In the event financial support continued and led to papers on field ecology of the euro (Ealey & Main 1967), diet (Storr 1968), water metabolism (Ealey et al. 1965) and nitrogen requirements (Brown & Main 1967). In essence, this work showed that the euro could supplement the low nitrogen of its diet by recy¬ cling urea. Meanwhile, work continued on other macropods, es¬ pecially the tammar wallaby whose ability to persist and thrive in dry habitats is well exemplified by its occur¬ rence on East and West Wallabi Islands of the Abrolhos. These studies dealt with the ability to drink sea water 93 Journal of the Royal Society of Western Australia, 78(4), December 1995 (Kinnear et al 1968; Purohit 1974), recycle urea (Kinnear & Main 1975), and water and electrolyte metabolism (Bakker et al 1983). A general interpretation of this work in terms of nutritional biology of ruminant and ruminant-like mammals was developed (Kinnear et al 1979) and related to niche theory (Kinnear & Main 1979). These studies were reviewed in the context of native animals as resources (Main 1969), measures of well-be¬ ing (Main 1971; Bakker & Main 1980) and adaptations to desert conditions (Main 1975; Main & Bakker 1981), and as a factor in the Late Pleistocene extinctions of marsupials (Main 1978). A wider more generalised review (Main 1983) indicated a possible integration of macropod eco- physiology with ecological and evolutionary studies in the context of either the fundamental or realised Hutchinsonian niche (Hutchinson 1957). The fundamental niche is an n-dimensional space which encloses all the environmental states that a spe¬ cies can tolerate and persist in indefinitely. In the pres¬ ence of other species with which there are biological interactions, a species occupies a realised niche in which the dimensions are less than those of the fundamental niche. In the context of ecophysiology of macropods, these dimensions relate to heat tolerance (including re¬ flective pelage), water balance, electrolyte metabolism, ability to recycle urea (supplement dietary nitrogen), and behavioural traits related to environmental struc¬ ture (shelter and cover) which permit trade-offs between attributes of the niche and persistence in the unpredict¬ able Australian environment. In the absence of other interacting species, the realised niche may approximate the fundamental niche. This is important in cases of single species on islands or in reservations where spe¬ cies may persist in quite atypical habitats because they still satisfy' the dimensions of the fundamental niche e.g. quokka on Rottnest. Furthermore, when species are ob¬ served under field conditions, we see only the dimen¬ sions of the realised niche which may be a poor basis for prognosis of persistence or success in reservations where the ecosystem is changed and perhaps simplified. The findings arising from the studies on macropods, birds and reptiles have converged to show that nutri¬ tion, dietary composition, water and electrolyte balance coupled with behaviour conspire to provide a suite of labile adaptations which permit survival when the envi¬ ronment in which the animals found themselves offered the scope to display and exploit the evolved traits. These findings are relevant to conservation and raised questions of what could be retained in reservations. A question I had earlier (see below) approached from the point of view of adequacy of reserve size using macropod species retained on continental islands as a surrogate measure. However, the studies mentioned above raised questions other than simple size as contrib¬ uting to persistence. For example, habitat structure, composition, senescence and regeneration appeared to be critical to maintain the habitats which gave scope for adaptive traits to be deployed. While these studies indi¬ cated requirements necessary for persistence of macropods, they also hinted at another problem, namely, whether management for the successful retention of nominated species (flagship or umbrella species) would ensure that other elements of the biota were also be retained. Conservation Like my other biological interests my concern with conservation dates back to my youth and as a member of the West Australian Naturalists Club. While in the Army I had been stationed on the sand plains west of Dandarragan and had become very aware of the immense floral and invertebrate species richness of these areas. In 1947 I attempted to have extensive sandplain areas reserved but without success. The con¬ sensus opinion at the time was that these areas could not be used for agriculture and so would always remain as vacant crown land, and having them reserved would merely preclude other better areas from being consid¬ ered for reservation. The fallacy of this reasoning be¬ came apparent as soon as it was realized that the agri¬ cultural potential could be enhanced simply by applying trace elements to the land. My next direct involvement in conservation occurred when I set a student project on the Western Swamp Tor¬ toise (Pseudemydura umbrina). This was taken up by An¬ drew Burbidge as the topic of a doctoral dissertation. Again, taxonomic relationships were crucial in obtain¬ ing a perspective for its conservation (Burbidge et al 1974). Many subsequent observations by Burbidge have led to an understanding of the difficulties associated with conserving small populations of vertebrates. In the 1950's, I had become associated with the Fauna Advisory Committee, later to become the Western Aus¬ tralian Wildlife Authority, which had the statutory re¬ sponsibility for nature reserves under the Fauna Act. Initially the committee and the Authority advocated that no interference with reserves equated to good manage¬ ment of them i.e. leave Nature alone. To me, reservations posed several problems; 1 there was no evidence that the size of the reserves being set up would fulfil the purpose for which they were being created especially whether they were ad¬ equate in area to maintain the biota included; 2 whether the disturbance resulting from fire and con¬ sequent plant successional pattern would permit the persistence of animals for which the reserves were being set aside; 3 whether the changed biotic environment e.g. within reserves, would permit the display of evolved traits and behaviour which laboratory and field analyses had demonstrated were a pivotal part of adaptation to the present environment. At this time, even as at present, the species around which conservation efforts centred were vertebrates, usually marsupials and birds. In the absence of a long record of successful conservation within reservations, there was no way of directly establishing how large an area would retain one or more species of mammal. However, numerous islands around the Western Aus¬ tralian coast retained macropod marsupials since their isolation from the mainland by rising sea levels at the beginning of the Holocene. These islands offered a sur¬ rogate measure of the number of species of macropods which could exist on islands of various sizes for up to 11 000 years. Main (1961) and Main & Yadav (1972) argued that these data provided the only evidence for 94 Journal of the Royal Society of Western Australia, 78(4), December 1995 predicting what could be retained in reservations of re¬ stricted size, assuming they were the analogue of is¬ lands. The quokka on Rottnest and the tammar on the Abrolhos could persist in harsher, much more simpli¬ fied environments than that usually occupied on the mainland, provided that the animals could exploit their behavioural and physiological repertoire of adaptive traits i.e. the environment remained within the dimen¬ sions of the species' fundamental niche. Thus, the prog¬ nosis was that the populations would persist in reserves of equivalent size. Unfortunately this first approxima¬ tion underestimated the significance of the fox. When setting aside land for reservation, it was hoped that marsupials, with their readily identified public ap¬ peal, would act as umbrella species beneath which other elements of the biota (plants and other animals, espe¬ cially invertebrates) would be sheltered and so retained along with the high profile or principal species. Notwithstanding the persistence of macropods on is¬ lands, it was clear from personal familiarity that the habitats available on an island are quite unlike mainland habitats in which the species now occurs and which were presumably like the ancestral ones. It is conceiv¬ able that proximity to the sea, salt spray, and exposure to high winds and storms may have changed the com¬ position of communities. However, reserves in the wheatbelt landscape are also isolated and exposed, and retention of macropods in them was no guarantee that other elements of the biota initially included with them in a reserve would be also retained. This in turn sug¬ gested that a reserve management policy based on 'leav¬ ing Nature alone' was unlikely to be successful, particu¬ larly if macropods, or other charismatic species, were used as umbrella species in the hope of retaining other elements of the biota. Despite these quibbles, the islands around the Western Australian coast do indicate that retention of the original biota is dependent on island area. Large islands such as Bernier, Dorre and Barrow have not only retained macropods but also most re¬ semble the mainland with respect to composition of the vertebrate biota prior to the arrival of the red fox. The foregoing highlighted the need for understand¬ ing the dynamics of populations, communities and eco¬ systems, and particularly the role of disturbance, espe¬ cially fire and herbivory, in their maintenance and re¬ generation. In the absence of the formerly wide-spread burrowing marsupials, rabbits and fire now tend to be the commonest agents causing disturbance and inhibit¬ ing plant regeneration. However, invertebrates such as grasshoppers may also exert subtle influences such as inhibiting the post-fire regeneration of jarrah trees by eating seedlings at the cotyledon stage (Whelan & Main 1979). This result emphasised how much the judgment that, 'it is reasonable to assume that insects are much less significant than grazing mammals in determining floristic composition', depends on our level of igno¬ rance. An approach to the management of reservations in the Wheat belt of Western Australia was proposed by Main (1987). In the early 1970's, I accepted appointments to the Universities Commission, the Council of the Australian Institute of Marine Science, and I was one of the initial three members of the Western Australian Environmental Protection Authority. These appointments precluded ex¬ tensive field work and my research emphasis changed as other issues emerged. Important among the emerging issues were: (a) the potential for climatic change, under the influ¬ ence of the greenhouse effect as atmospheric com¬ position changed, to alter the structure and com¬ position of biota retained within reserves; (b) an acceleration of salt encroachment in the wheatbelt landscape; (c) a developing appreciation of the significance of patches of remnant vegetation in maintaining a hydrological balance (many of these patches may be reservations for conservation purposes); (d) an appreciation of the necessity to develop con¬ cepts which would guide management practices rather than offer universal prescriptions for ac¬ tion. In a very important sense, these issues all have a bear¬ ing on what can be retained within or outside reserva¬ tions because they determine components of the realized niche of those species which are of concern for conser¬ vation. But the problem of establishing what species can persist is experimentally intractable over the period available for meaningful action to ensure retention. The fundamental niche will only change slowly un¬ der the influence of selection; however, the realized niche will change whenever environmental conditions alter. Thus, should changes occur it is unlikely that the present composition will be retained because each spe¬ cies is likely to have a new realized niche resulting in the reorganisation of ecosystems. Notwithstanding these limitations, it seems likely that the insights gained from the studies reviewed above offer useful guidance in each of the fields listed. These have been applied in relation to ecosystem management (Main 1981a, b, c), and management of remnants of native vegetation (Main 1987), retention, significance and problems in retaining of rare species (Main 1982, 1984), fire tolerance of ani¬ mals (Main 1981b), the role of biodiversity (Main 1992), climatic change and its likely effect on nature conserva¬ tion (Main 1988), and restoration ecology (Main 1993). Concluding Remarks In the introduction, I set out how the lecture was to develop i.e. it was to be an illustration of how my natu¬ ral curiosity posed the way in which questions were asked and how these led to matters of conservation. It is now possible to draw the lessons together and in do¬ ing this I wish to emphasise the close relationship of biology, politics and society. The core of science is; recognise entities; distinguish them from similar entities; classify; explain i.e. have an hypothesis and then test it. The first step is to distin¬ guish and describe entities so that others can recognise them. Should recognition not be precise then the use of the entities in experiments e.g. taking sibling or cryptic species together as one entity will mean that compara¬ tive studies will be handicapped. 95 Journal of the Royal Society of Western Australia, 78(4), December 1995 But there is a more general point to be made with regard to classification. I have been, or am regarded (classified) as a herpetologist (I admit that being an elected Honorary Life Member of the American Society of Ichthyologists and Herpetologists must colour the in¬ terpretation) but the initial choice of frogs for study was my familiarity with them and a belief that they would be a satisfactory experimental animal for study of adap¬ tations to aridity. This last theme is the key to my work on frogs. Moreover, the pursuit of a problem or theme across disciplinary (taxonomic) boundaries i.e. a com¬ parative approach, is the only way to test the generality of an interpretative hypothesis which in the present case was that physiology, behaviour in an environment which permitted the display of evolved physiological and behavioural responses provided the basis for sur¬ vival of individuals and persistence of populations. The survival of organisms during periods of drought and aridity experienced in the past and their capacity to per¬ sist in restricted areas and as small populations during such times (including post-fire situations) is likely to have selected traits that are adaptive under such condi¬ tions. Study of such past events and population conse¬ quences is relevant to conservation where persistence as small populations in restricted areas is the central prob¬ lem to be solved. But this is an example of seamless Nature; the call now is for teams to undertake interdisci¬ plinary studies in order to solve a problem. The prob¬ lem may be well defined and pressing; however, there is often an absence of theory on which testable hypotheses can be erected. Additionally, it is often not clear that a need for theory is appreciated, the vision of what is to be studied is purely practical, applied and constrained by utilitarian goals. The current funding sees the study of Nature as being that of patches of fabric which can be stitched or cobbled together to serve utilitarian purposes. This result is the logical end point of the comments re¬ ceived by me and referred to above regarding my disre¬ spect for, and transgression of, the currently accepted disciplinary boundaries, namely that knowledge of the world about us is hopelessly fragmented. Of course the initial choice of questions to be ad¬ dressed in my research was not a scientific one. It may be conceived as morally proper to know and understand the surrounding world but the decision of what to study and how to proceed was a personal choice, only the methodology was scientific. Yet, as mentioned earlier, the crossing of perceived boundaries did receive some comment, which can be interpreted in two ways; 1 that the accepted boundaries should be taken as given and not transgressed; 2 the decision by me to try and understand the sur¬ rounding world was morally wrong. A proper ap¬ proach was to stay within boundaries and/or to ad¬ dress problems posed by society i.e. deal with applied problems where a need for cross disciplinary work may be demonstrated. Thus the current approach of assembling teams formalises, in an administrative sense, the personal prejudices so frequently expressed in the past i.e. syn¬ thesis is only needed when practical solutions demand it. The broad understanding and interpretation is seen as unnecessary; it is only needed to satisfy the pragmatic requirements of cost effective economic progress e.g. to ensure The sustainable utilisation of ecosystems and spe¬ cies' (Anon 1987). More broadly, the foregoing is often taken to mean retaining biodiversity (species richness) and ecosystem functions. Taking the ecosystem functions to embrace responses to competition, disturbance, and stress (Grime 1974) and recognising the physiological and behavioural requirements of species for their survival poses the ques¬ tion of the nature of the assumption being made about environmental stability. At its simplest, the assumption is that the present is like the past and the future will be likewise. Historical geological considerations suggest that this is unlikely- Moreover, conditions are changing rapidly with development which is leading to more radi¬ cal changes than occurred in the past; it is unlikely that the future will be like the present. Thus wre are left with the question of whether the physiological, nutritional, and life history and behavioural traits of species (their fundamental niches) can be satisfied in the new environ¬ ments. These are management problems and successful outcomes depend on decisions being made promptly, but with recognition that adaptive management in the sense used by Walters (1986) may be called for. This is a condition that is rarely met because there is a perception that there should be complete understanding before ac¬ tion is taken or decisions made. This desire for certitude is, however, often eclipsed by a desire on the part of administrators to achieve or maintain status or exercise sheer power which would be entirely appropriate in the jungles of the natural world. The general difficulties experience by ecologists when dealing with decision¬ makers have been discussed by me (Main 1996). De¬ spite the traumas and difficulties associated with in¬ volvement in arriving at, or implementing, decisions it behoves ecologists to become involved in the way their knowledge is used, without their presence when deci¬ sions are made ecological knowledge will be miscon¬ strued, misinterpreted or ignored in both conservation problems and environmental issues. I regard discovery, interpretation and application of information as the re¬ sult of a conscious decision as being the political mani¬ festation of the seamlessness of Nature. Acknowledgments: It is a pleasure to acknowledge; financial support from the University Research Grants committee, The Australian Research Grants Committee, arid from CSIRO funds to Professor H Waring; the co-opera¬ tion and tolerance of my co-workers; the generous assistance and help from many in organisations outside the University which made it all pos¬ sible. I am grateful for the Honorary Research Fellowship in the Zoology Department, University of Western Australia. To the extent that 1 have succeeded in the pursuit of my vision I have been favoured by luck in being in the right place at the right time. To all those who made it possible, especially my wife Barbara York Main, I am extremely grateful. It has been wonderful for me to pursue my hobby. I am sure I could not do so in the present economic and funding climate. References Anon. 1953 Love croak of the frog recorded at Bassendean. Broadcaster 20. No. 1002:1-4. Anon. 1987 A State Conservation Strategy for Western Austra¬ lia: a sense of direction. Department of Conservation and En¬ vironment, Perth, Western Australia. Bulletin 270. 96 Journal of the Royal Society of Western Australia, 78(4), December 1995 Bakker H R, Bradshaw S D & Main A R 1983 Water and electro¬ lyte metabolism of the tammar wallaby ( Macropus eugenii). Physiological Zoology 55:209-219 Bakker H R & Main A R 1980 Condition, body composition and total body water estimation in the quokka (Setonix brachyurus). Australian Journal of Zoology 28:395-406. Bentley P J & Main A R 1972a Zonal differences in permeability of the skin of some anuran amphibia. American Journal of Physiology 223:361-363. Bentley P J & Main A R 1972b Effect of vasotocin on cutaneous water uptake by an Australian frog, Crinia georgiana. Copeia 1972:885-886. Bentley P J, Lee A K & Main A R 1958 Comparison of dehydra¬ tion and hydration of two genera of frogs (Heleioporus and Neobatrachus) that live in areas of varying aridity. Journal of experimental Biology 35:677-684. Bradshaw S D & Main A R 1968 Behavioural attitudes and regu¬ lation of temperature in Amphibolous lizards. Journal of Zo¬ ology 154:193-221. Brown C D & Main A R 1967 Studies on marsupial nutrition. V. The nitrogen requirements of the euro Macropus robustus. Aus¬ tralian Journal of Zoology 15:7-27. Burbidge A A, Kirsch J A W & Main A R 1974 Relationships within the Chelidae (Testudines: Pleurodira) of Australia and New Guinea. Copeia 1974:392-409. Crocker R L & Wood J G 1947 Some historical influences on the development of South Australian vegetation communities and their bearing on concepts of classification in ecology. Transactions of the Royal Society of South Australia 71:91- 136. Ealey E H M & Main A R 1967 Ecology of the euro, Macropus robustus (Gould) in north-western Australia. CSIRO Wildlife Research 12:53-65. Ealey E H M, Bentley P J & Main A R 1965 Studies on water metabolism of the hill kangaroo Macropus robustus (Gould) in Western Australia. Ecology 46:473-479. Grime J P 1974 Vegetation classification by reference to strate¬ gies Nature 250:26-31. Hutchinson G E 1957 Concluding remarks. Cold Springs Harbour Symposium on Quantitative Biology 222:415-427. Kinnear J E & Main A R 1975 The recycling of urea nitrogen in the wild tammar wallaby ( Macropus eugenii), a ruminant-like marsupial. Comparative Biochemistry & Physiology. 51A:793-810. Kinnear J E & Main A R 1979 Niche theory and macropod nutri¬ tion. Journal of the Royal Society of Western Australia 62:65- 74. Kinnear J E, Purohit G & Main A R 1968 Ability of the tammar wallaby ( Macropus eugenii : Marsupialia) to drink seawater. Comparative Biochemistry and Physiology 25:261-282. Kinnear J E, Cockson A, Christensen P & Main A R 1979 The nutritional biology of ruminants and ruminant-like mammals - a new approach. Comparative Biochemistry & Physiology 64A-.357-365. Licht P, Dawson W R, Shoemaker V H & Main A R 1966 Heat resistance of some Australian lizards. Copeia 1966:162-169. Littlejohn M J & Main A R 1959 Call structure in two genera of Australian burrowing frogs. Copeia 1959:266-270. Loveridge A 1935 Australian Amphibia in the Museum of Com¬ parative Zoology, Cambridge, Massachusetts. Bulletin of the Museum of Comparative Zoology 78:1-60. Main A R 1961 The occurrence of Macropodidae on islands and its climatic and ecological implications. Journal of the Royal Society of Western Australia 44:84 - 89. Main A R 1965 The inheritance of dorsal pattern in Crinia species (Anura: Leptodactylidae). Journal of the Royal Society of Western Australia 48:60-64. Main A R 1968 Ecology, Systematics and Evolution of Australian Frogs. In: Advances in Ecological Research (ed J B Cragg). Academic Press, London, 5:37-86 Main A R 1969 Native animal resources. In: Arid Lands of Aus¬ tralia (eds R O Slayter & R A Berry). Australian University Press, Canberra, 93-104. Main A R 1971 Measures of wellbeing in populations of herbivo¬ rous macropod marsupials. In: Dynamics of Populations (eds P J den Boer & G R Gradwell). Centre for Agricultural Pub¬ lishing and Documentation, Wageningen, The Netherlands, 159-173. Main A R 1975 Adaptations of Australian vertebrates to desert conditions. In: Evolution of desert biota (ed D W Goodall). University of Texas Press, Austin, 101-131. Main A R 1978 Ecophysiology: towards an understanding of late Pleistocene extinctions. In: Biology and Quaternary Environ¬ ments (ed D Walker & J C Guppy). Australian Academy of Science, Canberra, 169 - 183. Main A R 1981a Ecosystem theory and management. Journal of the Royal Society of Western Australia 64:1-4 Main A R 1981b Fire tolerance of heathland animals. In: Ecosys¬ tems of the World 9 B: Heathlands and Related Shrublands (ed R L Specht). Elsevier Scientific Publishing, Amsterdam, 85-90. Main A R 1981c Plants as animal food. In: Biology of Native Australian Plants (eds J S Pate & A J McComb). University of Western Australia Press, Perth, 342-360. Main A R 1982 Rare species: precious or dross? In: Species at Risk: Research in Australia (eds R H Groves & W D L Ride). Australian Academy of Science, Canberra, 163 - 174. Main A R 1983 Macropod ecophysiology: a possible integration with ecological and evolutionary studies. In: Research on Rottnest Island (ed S D Bradshaw). Journal of the Royal Soci¬ ety of Western Australia 66:5-9. Main A R 1984 Rare species, problems of conservation. Search 15:94-97. Main A R 1986 Resilience at the level of the individual animal. In: Resilience in Mediterranean-type Ecosystems (ed B Dell, A J M Hopkins & B B Lamont). Dr W Junk, Dordrecht, 83-94. Main A R 1987 Management of remnants of native vegetation - a review of the problem and the development of an approach with reference to the wheatbelt of Western Australia. In: Na¬ ture Conservation: The Role of Remnants of Native Vegeta¬ tion (eds D A Saunders, G W Arnold, A A Burbidge and A J M Hopkins). Surrey Beatty and Sons, Chipping Norton, 1-13. Main A R 1988 Climatic change and its impact on nature conser¬ vation in Australia. In: Greenhouse: Planning for Climatic Change (eds G I Pearman). CSIRO Division of Atmospheric Physics, Melbourne, & E J Brill, Leiden, 361-374. Main A R 1992 The role of biodiversity in ecosystem function: an overview. In: Biodiversity in Mediterranean Ecosystems in Australia (ed R J Hobbs). Surrey Beatty and Sons, Chipping Norton, 77-93. Main A R 1993 Restoration ecology and climatic change. In: The Reconstruction of Fragmented Ecosystems (eds D A Saunders, R J Hobbs & P R Ehrlich). Surrey Beatty and Sons, Chipping Norton, 27-32. Main A R 1996 Decision makers and ecologists: networks and difficulties. In: Nature Conservation 4: The Role of Networks (eds D A Saunders, J L Craig, & E M Mattiske). Surrey Beatty & Sons, Chipping Norton, 141-147. Main A R & Bakker H R 1981 Adaptation of macropod marsupi¬ als to aridity. In: Ecological Biogeography of Australia (ed A Keast). Dr W Junk, The Hague, 1491-1519. Main A R & Bentley P J 1964 Water relations of Australian bur¬ rowing frogs and tree frogs. Ecology 45:379-382. Main A R, Littlejohn M J & Lee A K 1959 Ecology of Australian frogs. In: Biogeography and Ecology in Australia (eds A Keast, R L Crocker & C S Christian). Dr W Junk, Den Haag, 396-411. Main A R, Shield J S & Waring H 1959 Recent studies in marsu¬ pial ecology. In: Biogeography and Ecology in Australia (eds 97 Journal of the Royal Society of Western Australia, 78(4), December 1995 A Keast, R L Crocker & C S Christian). Dr W Junk, Den Haag, Netherlands, 315-331. Main A R &Yadav M 1971 Conservation of macropods in re¬ serves in Western Australia. Biological Conservation 3:123- 133. Mayr E 1949 Systematics and the Origin of Species. Columbia University Press, New York. Moir R J, Somers M & Waring H 1956 Studies on marsupial nutrition. 1 Ruminant-like nutrition in a herbivorous marsu¬ pial Setonix brachyurus (Quoy and Gaimard) Australian Jour¬ nal of Biological Sciences 9:293 - 304. Moore J A 1946 Incipient intraspecific isolating mechanisms in Rana pipie?is. Genetics 31:304-326. Moore J A 1949 Patterns of Evolution in the genus Rana. In: Genetics and Evolution (eds G L Jepson, E Mayr and G G Simpson). Princeton University Press, Princeton, New Jersey, 315-338. Parker H W 1940 The Australian frogs of the Family Leptodactylidae. Novitates Zoologicae 42:1-106. Purohit K G 1974 Observations on size and relative medullary thickness in kidneys of some Australian mammals and their physiological appraisal. Zeitschrift fur Angewandte Zoologie 4:495-505. Storr G M 1964 Studies on marsupial nutrition, iv. Diet of the quokka Setonix brachyurus (Quoy & Gaimard) on Rottnest Is¬ land, Western Australia. Australian Journal of Biological Sci¬ ence 17:469-481. Storr G M 1968 Studies on marsupial nutrition. The diet of kangaroos ( Megaleia rufa and Macropus robustus) and the me¬ rino sheep near Port Hedland, Western Australia. Journal of the Royal Society of Western Australia 51:25-32. Walters C 1986 Adaptve Management of Renewable Resources. Macmillan, New York. Whelan R J & Main A R 1979 Insect grazing and post-fire plant succession in south-west Australian woodland. Australian Journal of Ecology 4:387-398. Williams CK & Main A R 1976 Ecology of Australian chats (Epthianura Gould): seasonal movements, metabolism and evaporative water loss. Australian Journal of Zoology 24:397- 416. Williams C K & Main A R 1977 Ecology of Australian chats ( Epthianura Gould): aridity, electrolytes and water economy. Australian Journal of Zoology 25:673-691. 98 Journal of the Royal Society of Western Australia, 78:99-101, 1995 Diurnal stratification of Lake Jandabup, a coloured wetland on the Swan Coastal Plain, Western Australia D S Ryder & P Horwitz Department of Environmental Management, Edith Cowan University, Joondalup WA 6027 Manuscript received June 1995 ; accepted February 1996 Abstract Variations in temperature and dissolved oxygen gradients were examined for three microhabi¬ tats in Lake Jandabup over a 24 hour period in 1993; each microhabitat was thermally stratified between 1300 and 1900 hrs. The presence of a shallow water column (< 0.5m) that is shown here to stratify thermally under a known set of conditions has implications for sediment nutrient release, insufficient oxygen supply for aerobic organisms and an accumulation of organic sedi¬ ments. Introduction The presence of dissolved humic substances in aquatic systems can influence the physical and chemical characteristics of the standing water (Kuuppo-Leinikki & Salonen 1992), causing the rapid attenuation of solar radiation at shallow depths (Bowling et al. 1986) and producing steep thermal gradients that are resistant to mixing (Bowling 1990). Literature regarding the stratifi¬ cation of Australian inland waters is dominated by studies of large and/or deep lakes and reservoirs (e.g. Bowling 1990) with Western Australian literature also having these emphases (e.g. Imberger 1985). Recent work has highlighted the stratification of coloured or saline waters in southwestern Australia. For instance, Edward et al. (1994) found many coastal wetlands in the extreme southwest to be thermally stratified, while Burke & Knott (1989) demonstrated that saline Lake Hayward in Yalgorup National Park was monomictic. Schmidt & Rosich (1993) examined stratification and thermal stabilities of Swan Coastal Plain wetlands, concluding that only deep (>3m) or highly coloured wetlands would stably stratify. However, Burke & Knott (1989) showed that stratification can be achieved in shallow lakes, provided that there was sufficient salinity difference between the upper and lower water layers. The aim of this study was to examine diurnal fluctua¬ tions in temperature and dissolved oxygen gradients over a 24 hour period during autumn in Lake Jandabup. Methods Lake Jandabup occupies a shallow (<1.5 m), oval¬ shaped basin, 22 km north of Perth on the Swan Coastal Plain (Allen 1979). The littoral zone is dominated by Baumea articulata (R Br) ST Blake (jointed twig rush) and other rushes with Typha orientalis C Presl (bulrush) lim¬ ited to isolated pockets within the lake (Froend et al. 1993). The study sites 1, 2 and 3 (see Ryder & Horwitz 1995) were within vegetation communities dominated by B. © Royal Society of Western Australia 1995 articulata, T. orientalis and a community with no emergent or submerged vegetation respectively. Dissolved oxygen (DO; % saturation) and temperature (°C) were recorded at two hourly intervals in situ on March 20 to 21, 1993, over a 24 hour period using a portable Wissenshaftlich Technischc Werkstdtten Oximeter. Measurements were made at the top (5 cm below water surface) and bottom (epibenthic) of the water column. Ambient air tempera¬ ture and wind data were obtained from the Bureau of Meteorology. Results Water depths at sites 1, 2 and 3 were 0.449 m, 0.404 m and 0.453 m respectively. The maximum air temperature on site was 26.5°C at 1500 hours and the minimum of 13.1°C was recorded at 0500 hours. Wind data are shown in Fig 1. A thermal gradient was present between the top and bottom of the water column at all sites between 1300 and 1900 hours. Sites 1 (B. articulata community), and 2 (T. orientalis community) exhibited the greatest differ¬ ences between surface and epibenthic temperatures, of 6.6°C and 6°C respectively (Fig 1). Site 3 (open water) had the highest surface temperature of 24.1°C; however, only a 3.4°C maximum temperature difference was evi¬ dent. All sites displayed low epibenthic oxygen satura¬ tion while surface water remained well oxygenated. Site 1 had the lowest epibenthic DO level of 0 % saturation. Site 2 displayed similar, although slightly higher values. Site 3, which was exposed to the prevailing winds more so than the other sites, had a surface oxygen level reach¬ ing 128 % saturation at 1700 hours, with epibenthic DO levels of 11 % saturation. Discussion This study has demonstrated the potential for the daily thermal stratification of Lake Jandabup, despite its shallow water depth and relatively high surface area. The climatic conditions, particularly the wind pattern, were typical of summer and autumn conditions experi¬ enced on the Swan Coastal Plain (Gentilli 1972). This 99 Journal of the Royal Society of Western Australia, 78(4), December 1995 SITE 1 20 t Wind Direction and Time of Day Figure 1. Temperature (°C) and DO saturation (%) readings for the top and bottom of the water column at sites 1 2 and 3 and DO bottom m* direCtl°n OVer the 24 h°Ur Study peri°d from 20 to 21 March 1993‘ Temperature top 0, temperature bottom ♦ , DO top □, 100 Journal of the Royal Society of Western Australia, 78(4), December 1995 pattern of daily stratification could therefore be expected to occur regularly over a six month period. Surface and epibenthic temperature differences and epibenthic deoxygenation were most pronounced in the dense B. articulata community and least evident in open water. As the major mixing force in wetlands is pro¬ duced by the shearing effect of the wind at the water surface (Schmidt & Rosich 1993), it appears that the emergent vegetation may be limiting the input of atmo¬ spheric oxygen as well as providing shelter and resis¬ tance to mixing. Wetlands such as Lake Jandabup, with large areas of emergent vegetation and coloured water may therefore have an exacerbated daily and seasonal cycle of thermal stratification. Based on traditional limnology (e.g. Wetzel 1983), wetlands such as Lake Jandabup would be classified as warm polymictic, based on the temperature, climate and basin geomorphology. However, despite the obvious shallowness of Lake Jandabup, the presence of a water column that has been shown to stratify thermally under a known set of conditions requires the wetland to be classified as warm continuous monomictic. The presence of a stratified water column based on thermal profiles is supported by differences in surface and epibenthic DO levels. These DO differences are most pronounced during the period of thermal stratification, but persist throughout the 24 hour period due to the atmospheric diffusion of oxygen through the water col¬ umn being unable to supply the high oxygen demand of the organic soils present in the lake. This has implications for sediment nutrient release, insufficient oxygen supply for the metabolic requirements of aerobic organisms, and an accumulation of organic sediments in Lake Jandabup and other similar wetlands on the Swan Coastal Plain. Acknowledgments: We would like to thank Mr Doug Ryder and Mr Allan Ellies for their help in the field. References Allen A D 1979 The hydrogeology of Lake Jandabup Swan Coastal Plain, W.A. In: Western Australian Geological Sur¬ vey, Annual Report 1979, 32-40. Bowling L C 1990 Heat contents, thermal stabilities and birgean windwork in dystrophic Tasmanian lakes and reservoirs. Australian Journal of Marine and Freshwater Research 41:429- 441. Bowling L C, Steane M S & Tyler P A 1986 The spectral distribu¬ tion and attenuation of underwater irradiance in Tasmanian inland waters. Freshwater Biology 16:313-335. Burke C M & Knott B 1989 Limnology of four groundwater-fed saline lakes in southwestern Australia. Australian Journal of Marine and Freshwater Research 40:55-68. Edward D H D, Gazey P & Davies P M 1994 Invertebrate com¬ munity structure related to physico-chemical parameters of permanent lakes of the south coast of Western Australia. Jour¬ nal of the Royal Society of Western Australia 77:51-63 Froend R H, Farrell R C, Wilkins C F, Wilson C C & McComb A J 1993 Wetlands of the Swan Coastal Plain. Volume 4 - The Effect of Altered Water Levels on Wetland Plants. Water Au¬ thority of Western Australia and Environmental Protection Authority, Perth. Gentilli J 1972 Australian Climate Patterns. Nelson Australia Pa¬ perbacks, Sydney. Imberger J 1985 The diurnal mixed layer. Limnology and Ocean¬ ography 30:737-770. Kuuppo-Leinikki P & Salonen L 1992 Bacterioplankton in a small polyhumic lake with an anoxic hypolimnion. Hydrobiologia 229:159-168. Ryder D S & Horwitz P 1995 Seasonal water regimes and leaf litter processing in Lake Jandabup, a seasonal wetland on the Swan Coastal Plain, Western Australia. Marine and Freshwa¬ ter Research 46:1077-1084. Schmidt L G & Rosich R S 1993 Physical and chemical changes and processes. In: Wetlands of the Swan Coastal Plain. Vol¬ ume 6 - Wetland Classification on the Basis of Water Quality and Invertebrate Community Data (eds J A Davis, R S Rosich, J S Bradley, J E Growns, L G Schmidt & F Cheal). Water Au¬ thority of Western Australia and Environmental Protection Authority, Perth, 29-80. Wetzel R G 1983 Limnology. Saunders Company, London. 101 Journal of the Royal Society of Western Australia, 78:103-106, 1995 Cocoon formation by the treefrog Litoria albognttata (Amphibia: Hylidae): A 'waterproof' taxonomic tool? P C Withers1 & S J Richards2 1 Department of Zoology, The University of Western Australia, Nedlands, WA 6907 2 Department of Zoology, James Cook University, Townsville, QLD 4811 Manuscript received June 1995; accepted January 1996 Abstract The hylid frog Litoria albognttata (formerly Cyclorana alboguttatus ) forms a cocoon during aestivation. After 21 days of water deprivation a thin, transparent cocoon is formed; the cocoon covers and closely adheres to the entire body surface, except the external nares. The cocoon consisted after 21 days of about 24 layers of squamous epithelial cells, about 14-18 p thick. It reduced markedly the rate of evaporative water loss (measured at 22 °C) from 39.3 mg g 1 h 1 (non-cocooned frogs) to 2.1 mg g'1 h l (cocooned frogs). The formation of a cocoon by Litoria albognttata raises the question of the possible phylogenetic significance of cocoon formation amongst the Australopapuan frogs, because no other Litoria has yet been reported to form a cocoon but cocoon formation is common amongst both Cyclorana and Neobat rachus . Cocoon for¬ mation may have independently arisen at least three times amongst Australopapuan frogs ( Litoria , Cyclorana , Neobatrachns). Alternatively, Litoria albognttata may be more closely allied with Cyclorana than Litoria , or cocoon formation was a primitive capability of the frogs ancestral to both Litoria and Cyclorana, and so cocoon formation independently evolved only twice in Australopapuan frogs. Introduction This study was prompted by the report of Lee & Mer¬ cer (1967) that Litoria albognttata (called Cyclorana alboguttatus by them) form a cocoon; they also reported that Cyclorana platycephala, C. australis, Neobatrachns pictns and Limnodynastes spenceri form a cocoon. The cocoon of these frogs, and of a number of other non- Australian frogs ( Pyxicephalns , Leptopelis, Lepidobatrachns, Ceratophrys, Pternohyla, Smiliscns) is a multi-layered cov¬ ering of sloughed skin formed during aestivation; it markedly reduces evaporative water loss (Loveridge & Craye 1979; McClanahan el al 1976; Ruibal & Hillman 1981; McDiarmid & Foster 1987). Since the pioneering study of Lee & Mercer (1967), cocoon formation has been described for a variety of species of Cyclorana (van Beurden 1982; Withers 1995; Richards, unpublished ob¬ servations) and Neobatrachns (Withers 1995), but cocoon formation has not been observed for Limnodynastes spenceri (Withers, unpublished observations) and there have been no further studies of cocoon formation by Litoria albognttata. Cocoon formation by Litoria albognttata is of particu¬ lar interest because the taxonomic position of Litoria albognttata within the monophyletic Cyclorana or Litoria anrea species group is problematical (Tyler & Davies 1993). The Australopapuan hylid frogs (Hylidae, Pelodryadinae: Cyclorana , Litoria and Nyctimystes ) ap¬ pear to be a monophyletic group (Tyler 1979; Hutchinson & Maxson 1987; Tyler & Davies 1978, 1993), but only relatively few of the species have been considered in phylogenetic analyses within the taxon. Litoria albognttata was described as Chiroleptes alboguttatus Gunther (1867); it was moved to Mitrolysis by Cope © Royal Society of Western Australia 1995 (1889) and then to Cyclorana by Parker (1940). Tyler (1973) considered it to be a species of Litoria and a mem¬ ber of the Litoria anrea complex, as L. albognttata. The objectives of this study were to confirm that Litoria albognttata formed a cocoon during aestivation, to examine the structure of the cocoon, and to measure the rate of evaporative water loss of normal and cocooned frogs. Methods Twelve specimens of Litoria albognttata were collected at Townsville, Queensland, and transported to Perth for study (two frogs are deposited in the WA Museum as voucher specimens R119533, R119534). Initially, the frogs were maintained individually, in the dark, with access to free water, in plastic containers that had a small hole in the lid for gas exchange. After initial measure¬ ments were made for hydrated frogs, no free water was provided and the frogs were slowly dehydrated to induce cocoon formation. Frogs were maintained, and all ex¬ periments conducted, at room temperature ( approx . 22 °C). The rate of evaporative water loss (EWL) was deter¬ mined by flow-through hygrometry. EWL was first measured at the start of the study for control, hydrated frogs, and then at the end of 21 days of dehydration for cocooned, aestivating frogs. Each frog was weighed to ± 0.001 g and then placed in a vertically-oriented glass tube (5 cm diameter) on a plastic mesh platform. Com¬ pressed dry air (dewpoint = -10 °C) was passed at a flow rate of 2000 ml min 1 through the tube and then a General Eastern model 1100A dewpoint hygrometer. The analog voltage output of the hygrometer was moni¬ tored by a Thurlby 1905a digital multimeter, and its RS232 output was interfaced to a PC. The excurrent 103 Journal of the Royal Society of Western Australia, 78(4), December 1995 dewpoint was monitored at 30 sec intervals, and con¬ verted to relative and absolute humidity using the equa¬ tions of Parrish & Putnam (1977). The absolute evaporative water loss rate (EWL; mg min-1) and mass-specific evaporative water loss (MSEWL; mg g'1 h1) were calcu¬ lated from the air flow rate, incurrent and excurrent absolute humidity, and body mass. The surface-area- specific evaporative water loss (SAEWL; mg cm'2 h'1) was calculated assuming a surface area (cm2) of 9.9 grams0567 (McClanahan & Baldwin 1969). Total resistance to evaporative water loss (R=AC/SAEWL; sec cm1) was calculated from SAEWL (converted to pg cm'2 sec1), assuming the difference in water vapor concentration (AC; pg cm'3) driving evaporation was the difference be¬ tween absolute humidity for saturated air at 22 °C and the incurrent absolute humidity. Samples of cocoon were removed from aestivating frogs, and examined by scanning and transmission elec¬ tron microscopy. For scanning electron microscopy, air- dried samples and glutaraldehyde-fixed samples of skin were mounted on an aluminium stub using double¬ sided adhesive tape, and sputter-coated with gold-palla¬ dium. The thickness of the cocoon was determined us¬ ing the air-dried specimens, and the number of layers counted using the glutaraldehyde-fixed specimens. Specimens were examined using a Phillips 505 scanning electron microscope. Samples of air-dry cocoon were prepared for trans¬ mission electron microscopy by fume fixation with os¬ mium tetroxide and direct embedding in araldite. Ultrathin sections were cut and stained with uranyl ac¬ etate and lead citrate. Specimens were examined using a JOEL FX2000 transmission electron microscope. All experiments were conducted with the approval of the Animal Ethics Committee, University of Western Australia. Results All frogs were hydrated and healthy at the start of the study, when EWL was determined for all individuals, and then were water-deprived. Most specimens of Litoria alboguttata became quiescent, adopted the water- conserving posture, and commenced cocoon formation (Fig 1) within 7 days of the start of water deprivation. A well-developed, transparent cocoon was apparent after 21 days, when EWL was redetermined and the cocoon was removed for microscopical examination. At this time the cocoon was a thin, transparent sheet that cov¬ ered and closely adhered to the entire body surface, in¬ cluding the closed eyes, mouth and cloaca. Only the external nares were free of the cocoon, to allow pulmo¬ nary ventilation. The cocoon was easily peeled from the skin; the freshly-exposed skin appeared and felt moist. The piece of cocoon examined by scanning electron microscopy consisted of about 24 discrete layers (Fig 2A) forming a compact sheet (Fig 2B) approximately 14-18 p thick; each layer is consequently calculated to be about 0.6-0.7 p thick. A transmission electron micrograph of the cocoon (Fig 3) more clearly shows the individual electron-dense cell layers, about 0.7 to 1.0 p thick, sepa¬ rated by 0.1 to 0.2 p thick inter-cellular spaces. The outer surface of the cells is more crenulated than the inner surface. Inter-cellular junctions are evident in some of the layers. The mean body mass of L. alboguttata was 20.6 ± se 0.7 grams (n=9). The rate of evaporative water loss de¬ clined markedly from 13.4 mg min'1 for non-cocooned L. alboguttata (39.3 mg g'1 h1; 4.07 mg cm'2 h1) to 0.74 mg min'1 for cocooned frogs (2.1 mg g'1 h'1, 0.22 mg cm'1 h1; Table 1). The resistance was considerably higher for cocooned frogs (89.4 sec cm1) than non-cocooned frogs (3.1 sec cm1; Table 1). Table 1. Evaporative water loss for hydrated and cocooned Litoria alboguttata. Values are mean ± standard error; n is the sample size. All values for cocooned frogs are significantly dif¬ ferent from the values for control frogs, by t-test (P<0.05). Hydra ted(n=9) Cocooned (n=6) Absolute EWL (mg min1) 13.4±0.6 0.74 ± 0.12 Mass-specific EWL (mg g1 h'1) 39.3±2.0 2.1 ± 0.3 Surface-area-specific EWL (mg g1 h'1) 4.1±0.2 0.22 ± 0.03 Resistance (sec cm-1) 3.1 ±0.3 89.4 ±14.9 Figure 1. An aestivating Litoria alboguttata in the water-conserving posture and covered with a transpar¬ ent cocoon that covers the entire outer body surface (including eyes, mouth and cloaca) except for the openings of the nares. 104 Journal of the Royal Society of Western Australia, 78(4), December 1995 A B Figure 2. Scanning electronmicrographs of the cocoon of Litoria alboguttata. A. glutaraldehyde-fixed specimen showing individual layers B. an air-dried specimen showing the compact in vivo structure of the cocoon. Scale bars are 10 p. Figure 3. Transmission electronmicrograph of an air-dried speci¬ men of Litoria alboguttata cocoon, showing detail of squamous cells and intercellular space. Imbricate intercellular junctions are indicated by arrow heads. Scale bar is 500 pm. Discussion It is clear from Lee & Mercer (1967) and this study that Litoria alboguttata forms a cocoon during aestivation; however, cocoon formation has not yet been described for any other species of Litoria. The general appearance and structure of the cocoon of L. alboguttata is similar to that of other Australian cocoon-forming frogs (. Neobatrachus and Cyclorana ; Withers 1995), and also Pyxicephalus adspersus (Parry & Cavill 1978; Loveridge & Craye 1979), Lepidobatrachus llanensis (McClanahan el al. 1976), Pternohyla fodiens (Ruibal & Hillman 1981) and Smilisca baudinii (McDiarmid & Foster 1987). The thick¬ ness of individual layers of the cocoon of L. alboguttata , 0.6-0. 7 p, is more similar to that of Neobatrachus spp (0.57-0.62 p) than Cyclorana rnaini (0.39 p; Withers 1995) but the rapid rate of cocoon formation for L. alboguttata (at least 1.1 layers d*1) is more similar to that of C. maini (0.57 d*1) than Neobatrachus spp (0.22-0.35 d'1; Withers 1995). The cocoon of Litoria alboguttata significantly reduces its rate of evaporative water loss, as has been observed for other cocoon-forming Cyclorana spp, Neobatrachus spp, and other genera. The resistance to water loss (3.1 sec cm*1) of non-cocooned L. albogutta is slightly higher than that for typical 'non-waterproof' frogs and a free water surface (about 1 sec cm1). This might reflect an underestimation of evaporative surface area, a substan¬ tial resistance of the cutaneous boundary layer, or initial stages of cocoon-formation in some of the "non- cocooned" individuals (resistance ranged for non- cocooned individuals from 2.4 to 4.9 sec cm1). In contrast, the resistance to water loss of cocooned frogs was much higher, at about 90 sec cm*1; such resistance values are typical for other species of cocooned frogs (Loveridge & Withers 1981; Withers, unpublished observations for Cyclorana and Neobatrachus spp). Cocoon formation has presumably evolved indepen¬ dently at least three times in the Australopapuan frogs i.e. in both hylid genera Cyclorana and Litoria and the myobatrachid genus Neobatrachus. If Litoria alboguttata is ascribed to the Cyclorana australis species group as suggested by Maxson et al. (1982, 1985), then cocoon formation need only have evolved independently twice in Australopapuan frogs, once in Cyclorana (there is at least one cocoon-forming frog in each of the three Cyclorana species groups listed by Tyler & Davies 1993), and once in Neobatrachus. Or, if cocoon formation was a primitive capability of the common ancestor of both Litoria and Cyclorana , but has been subsequently lost in most Litoria spp, then cocoon formation need only have 105 Journal of the Royal Society of Western Australia, 78(4), December 1995 evolved twice in Australopapuan frogs. However, the analyses of Maxson et al. (1982, 1985) suggest that C. platycephala is remote from its congeners (and L. alboguttata) and is allied with the Litoria aiirea complex; if so, then cocoon-formation might have evolved inde¬ pendently in C. platycephala (or other members of the L. aurea complex also form cocoons but this has not yet been recorded, or they have lost the capacity to form a cocoon). The ability to form a cocoon, although it cannot be considered to be a taxonomic tool per se, provides suggestive evidence of a close phylogenetic relationship between Cyclorana and Litoria alboguttata , and it will be of interest to see if such a relationship holds when the patterns of cocoon formation and phylogenetic relation¬ ships are clearer for Australopapuan frogs. Acknowledge merits: We thank Tom Stewart for his major contributions to the histological work in this study, the Electron Microscopy Centre of the University of Western Australia for providing SEM and TEM facilities, and the Department of Conservation and Land Management for permission to import frogs into Western Australia. We are grateful to M Tyler and W Buttemer for constructive comments on the manuscript. References Cope E D 1889 The Batrachia of North America. Bulletin of the United States National Museum 34:7-525. Gunther A 1867 Additions to the knowledge of Australian reptiles and and fishes. Annals and Magazine of Natural History (3) 20:45-68. Hutchinson M N & Maxson L R 1987 Phylogenetic relationships amongst Australian tree frogs (Anura: Hylidae: Pelodryadinae). Australian Journal of Zoology 35:61-74. Lee A K & Mercer E H 1967 Cocoon surrounding desert-dwelling frogs. Science 157:87-88. Loveridge J P & Craye G 1979 Cocoon formation in two species of Southern African frogs. South African Journal of Science 75:18-20, Loveridge J P & Withers P C 1981 Metabolism and water balance of active and cocooned African bullfrogs Pyxicephalus adspersus. Physiological Zoology 54:203-214. Maxson L R, Tyler M J & Maxson R D 1982 Phylogenetic relation¬ ships of Cyclorana and the Litoria aurea species-group (Anura: Hylidae): a molecular perspective. Australian Journal of Zoology 30:643-651. Maxson L R, Ondrula D P & Tyler M J 1985 An immunological perspective on evolutionary relationships in Australian frogs of the hylid genus Cyclorana. Australian Journal of Zoology 33:17-22. McClanahan L L, Shoemaker V H & Ruibal R 1976 Structure and function of the cocoon of a ceratophryid frog. Copeia 1976:179-185. McClanahan L L & Baldwin R 1969 Rate of water uptake through the integument of the desert toad, Bufo punctatus. Compara¬ tive Biochemistry and Physiology 28:381-389. McDiarmid R W & Foster M S 1987 Cocoon formation in another hylid frog, Smilisca baudinii. Journal of Herpetology 21:352- 355. Parker H W 1940 The Australasian frogs of the family Leptodactylidae. Novitates Zoologicae 42:1-106. Parrish O O & Putnam T W 1977 Equations for the determination of humidity from dewpoint and psychrometric data. NASA Technical Note D-8401. Parry C R & Cavill R 1978 A note on cocoon formation and struc¬ ture in Pyxicephalus adspersus Tschudi (Anura: Ranidae). Transactions of the Rhodesian Scientific Association 58:55-58. Ruibal R & Hillman S S 1981 Cocoon structure and function in the burrowing hylid frog, Ptenwhyla fodiens. Journal of Herpe¬ tology 15:403-408. Tyler M J 1973 The systematic position and geographic distribution of the Australian frog Chiroleptes alboguttatus Gunther. Pro¬ ceedings of the Royal Society of Queensland 85:27-32. Tyler M J 1979. Herpetofaunal relationships of South America with Australia. In: The South American herpetofauna: its origin, evolution and dispersal (ed W E Duellman). The University of Kansas, Museum of Natural History, Lawrence, 73-106. Tyler M J & Davies M 1978 Species groups within the Australo- Papuan hylid frog genus Litoria Tschudi. Australian Journal of Zoology Supplementary Series 63:1-47. Tyler M J & Davies M 1993 Family Hylidae. In: Fauna of Australia. Vol. 2A Amphibia & Reptilia (eds C J Glasby, G J B Ross & P L Beesley). Australian Government Publishing Sendee, Canberra, 58-63. van Beurden E 1982 Desert adaptations of Cyclorana platycephala : a holistic approach to desert-adaptation in frogs. In: Evolution of the Flora and Fauna of Arid Australia (eds W R Barker & P J M Greenslade). Peacock Publications, South Australia, 235- 240. Withers P C 1995 Cocoon formation and structure in the aestivating Australian desert frogs, Cyclorana and Neobatrachus. Austra¬ lian Journal of Zoology 43:429-441. 106 Journal of the Royal Society of Western Australia, 78:107-114, 1995 Foraging patterns and behaviours, body postures and movement speed for goannas, Varanus gouldii (Reptilia: Varanidae), in a semi-urban environment G G Thompson Edith Cowan University, Joondalup Drive, Joondalup, WA 6027 Manuscript received July 1995; accepted February 1996 Abstract Two Gould's goannas ( Varanus gouldii) were intensively observed in the semi-urban environ¬ ment of Karrakatta Cemetery, Perth, Western Australia. After emerging and basking to increase their body temperature, they spent most of their time out of their burrows foraging, primarily in leaves between grave covers, and under trees and shrubs. Mean speed of movement between specific foraging sites was 27.6 m min1, whereas the overall mean speed while active was only 2.6 m min'1 because of their slower speeds while foraging. A number of specific body postures were observed, including; vigilance, walking, erect, and tail swipes. Specific feeding and avoidance behaviours were also recorded, along with the influence that two species of birds had on their selection of foraging sites. Introduction Our knowledge of foraging habits, patterns, home range and activity area size, posture, and behaviour for large goannas has been extended since the early work of Cowles (1930) with V. niloticus, and Green & King (1978) with V. rosenbergi. Recent comprehensive studies include those by Auffenberg (1981a, b; 1988; 1994) for V. komodoensis , V. bengalensis and V. olivaceus, Auffenberg et al. (1991) for V. bengalensis , Daltry (1991) for V. salvatorf and Weavers (1993) for V. varius. General de¬ scriptions of varanid locomotion, postures and foraging behaviour are given by King & Green (1993 a,b). How¬ ever, there remains a paucity of data concerning the behaviours, body postures and movement patterns of small and medium sized varanid lizards. Pianka (1994) reports that Australian desert goannas, such as V. gouldii , are exceedingly wary, unapproachable, and unobservable; this consequently makes it very diffi¬ cult to study their use of time and space, foraging behaviour and body postures. Because V. gouldii at Karrakatta Cemetery have learned to accommodate to the presence of people, this site provides a unique op¬ portunity for a study of their behaviour and foraging patterns that is not possible in more remote locations. This study is the third in a series (Thompson 1992, 1994) on the movements, behaviours and ecology of V. gouldii at Karrakatta Cemetery, Perth, Western Austra¬ lia. The earlier papers report on the daily distance trav¬ elled, foraging areas and the size of the activity area during the breeding season. The primary objective of this study was to examine in detail the behaviours, pos¬ tures and preferred micro-habitat foraging sites for two particular individual V. gouldii. © Royal Society of Western Australia 1995 Methods Study Site Karrakatta Cemetery (115° 47' E, 31° 55' S) is located within the Perth metropolitan area, approximately 4 km west-south-west of the city centre. It has 53 ha allotted to burial plots and another 53 ha to roads, ornamental gardens and buildings. The cemetery is planted with a range of exotic shrubs and trees, with many of the north¬ ern, central and eastern areas being grassed. There are numerous rose gardens to the south and east of the main entrance, which are located on the northern boundary (see Thompson 1992, 1994). A nature reserve of ap¬ proximately 7 ha is located on the south-eastern bound¬ ary. The study site was located in the southern section of the cemetery (Fig 1). Monitoring procedures A miniature radio-transmitter with a battery life of approximately 140 days (11 g mass; TX1 1C tempera¬ ture-pulse, Bio-Telemetry Tracking) was attached to the lateral aspect of the base of the tail for seven V. goiddii in early November, 1993. The radio transmitter was sewn into a denim harness that was glued with Selleys 'kwik- grip' to the skin of the goanna's tail to encircle the tail for a length of approximately 50 mm. Each goanna was located prior to observations commencing each day with a Bio-tel RX3 receiver operating in the 150-151 MHz band in conjunction with a 3EY directional antenna. The behaviour and daily movements of five of these V. gouldii were initially monitored to choose two goannas that readily adapted to being observed. Behavioural ob¬ servations over five weeks commenced on these two V. gouldii [#1, a female, with a body mass of 330 g, snout- to-vent length (SVL) of 314 mm; and #2, a female, with a body mass of 348 g and SVL of 300 mm] on 1 December 1993 and concluded on 18 January 1994. This study was conducted in conjunction with a study that continuously 107 Journal of the Royal Society of Western Australia, 78(4), December 1995 200 metres I Tree canopy with a ground cover of leaves Shrubs with a ground cover of leaves Ground cover of leaves Grassed area Open ground with almost no leaf litter Buildings Figure 1. Karrakatta Cemetery, showing the study site vegetation and V. gouldii activity areas. monitoring of the body temperature of these V. gouldii over a number of months using temperature sensitive transmitters attached to the lizards and a data-logger connected to a large stationary antenna. During the first week of observation, each goanna was initially monitored from a distance of approxi¬ mately 50 m. After this adjustment period, most obser¬ vations were then carried out from a distance of 15 to 35 m. On numerous occasions, the observed V. gouldii moved toward the observer, reducing the distance to observer to less than 5 m, and on a fewT occasions to less than 2 m. This suggests that the presence of an observer had a minimal influence on their behaviour. The data from the first week of observation of the two V. gouldii were not analysed, as this period allowed these two V. gouldii to adjust to the continued presence of an ob¬ server. Two well trained observers spent a total of 131 hours watching and recording data for these two V. gouldii or waiting for them to emerge from their burrows, but they never spent more than half a day observing one goanna, to minimise any impact that these observations might have had on their normal behaviour. The recording pro¬ cedures of the two observers were checked every couple of days during the first two weeks to ensure consistency of observations and recording of data. Foraging, movement and posture recording procedures Observations for each goanna on the choice of forag¬ ing site, general habitat, extent of exposure to the sun and distance moved were recorded for the previous nine minute interval in the categories showm in Table 1. Movement behaviour for each nine minute period was classified into seven categories (Table 1). Data reported for the micro-habitat of foraging sites, general habitat type selected, extent of exposure to the sun, and dis¬ tance travelled, include only the time between when the goanna commenced foraging to when it either returned to a burrow or observations ceased (but excludes the time between when the goannas emerged from their overnight burrow and when they were first observed basking in the sun to increase their body temperature Tb). Goannas often moved through various sites during the nine minute period. The predominant environment for the period is the one reported. The number of times that V . gouldii #1 was seen to capture a prey item was also recorded, and is reported as the mean number of minutes between successful strikes. In addition, observ¬ ers recorded the time when the goanna produced a scat, its behaviour and when the two V. gouldii were harassed by birds. Observers also noted the feeding and digging behaviours for these two goannas. In late December, 1993, both goannas were recorded by video camera for a total of 6 hours, (four hours for 108 Journal of the Royal Society of Western Australia, 78(4), December 1995 Table 1. Alternatives used to classify the foraging sites, general habitat, extent of exposure to the sun and movement behavior. Micro-habitat foraging sites General habitat type selected Extent of exposure to the sun Movement behaviour Between grave covers Leaves under trees or shrubs Under grave covers Leaves in an open area Grass in an open area Grass under trees or shrubs Between graves On top or the side of graves Under grave covers Grassed areas On the road verge Treed areas In the yards of adjacent houses Up a tree In the nature reserve (heavily grassed with 50% tree canopy) Total sun 3A sun Vi sun M sun All of the time in the shade Under grave cover or in a burrow Emerging from an overnight burrow Basking but not moving Basking and moving Moving about outside a grave cover Moving under a grave cover Stationary and looking Avoidance (people, dogs, birds, vehicles) one and two hours for the other), over a four day period and body postures were drawn from these video im¬ ages. Reference to this visual record was also used to clarify the written descriptions of each observer for pos¬ tural, feeding and digging behaviours. The linear distance that each of the two V . gouldii moved during each nine minute observation period was estimated by recording the number of grave lengths and widths (3.6 m and 1.8 m respectively) that the goanna had passed during the period. This distance is a slight underestimate of the total distance moved due to the meanderings of the goannas while they foraged. How¬ ever, as the goannas mostly moved between grave covers or between foraging sites, their actual movement path was generally a series of near linear movements. These were summed for each period. For example, if a goanna was to circumambulate a grave cover the total linear distance would be approximately 10.8 m presuming it moved midway between adjacent grave covers. From these data, the mean distance travelled per minute was calculated. The average speed of movement between foraging sites was estimated by recording the time, to the nearest 0.1 second, that it took for a V. gouldii to cover a known distance (e.g. length of a grave plot) over thirteen trials. A trial score was only used if the goanna was moving in a straight line past two points whose Figure 2. Body postures adopted by V. gouldii at Karrakatta Cemetery. A: lying on a grave cover with the posterior abdomen flattened onto the grey concrete surface and the head and neck raised; B: on the side of a grave cover with the abdomen dorso-ventrally flattened and directed towards the sun; C: cloaca and anterior proportion of the tail being wiped on the ground after extruding a scat; D: lateral sigmoidal trotting action shown while walking; E: standing erect by balancing on the hind feet and tail; F: walking with only the distant end of the tail dragging on the ground; G, H, & I: tail swipes with non-aggressive posture. 109 Journal of the Royal Society of Western Australia, 78(4), December 1995 distance apart could be measured, it did not stop to for¬ age or view the surrounds, had not been obviously dis¬ turbed, and was not avoiding being seen or attacked. The ambient temperature (Ta), one metre above the ground, in the shade of a tree, was recorded every nine minutes, for the duration of the study using a Data-flow logger and a calibrated temperature sensitive probe at¬ tached to a transmitter. Occasional measurements of the temperature in the shade about one metre above the ground, surface soil in full sunlight, under a grave cover and on top of a grave cover in the sun were recorded to examine the range of thermal environments available to a goanna at any one time. Results and Discussion Emergence and responses to ambient temperature The goannas were seen to emerge from their over¬ night retreat on eight occasions between 0700 hours and 0945 hours. The mean Ta on emerging was 23.4 (± se 0.98) °C. On 15 occasions, the V. gouldii had emerged unseen before 0900 hours, and on another 11 occasions the goannas had not emerged by 0945 hours. As has been reported for other varanids (Sokolov et al 1975; King 1980; Weavers 1983), these two goannas would most often poke their heads out of their burrows for a while before their whole bodies emerged in the sun, sug¬ gesting that they are initially warming their heads possi¬ bly to increase the functioning of their nervous system (King & Green, 1993a). Once the V. gouldii had moved from under the grave cover and onto the top of it, they would dorso-ventrally flatten and lower their abdomen onto the grave cover so as to expose the greatest body surface area for both radiative and conductive heat gains (Fig 2A). They would remain constantly vigilant during this period (2 - 35 minutes) with head and neck held aloft, until they moved off to forage. After emerging and basking, foraging most often commenced in an area of full sunlight, and the goannas would often seek partially-shaded or fully-shaded areas to forage as the ambient temperature increased. A mean ambient temperature of 25.8 °C (± se 0.75, n = 14) was recorded for the first nine minute period that the goanna spent in either in 34 shade or a full shade. V. komodoensis are reported to adopted a similar strategy of moving from sunny areas into the shade as the ambient tem¬ perature increases (Auffenberg 1981a). One V. gouldii (#1), while vigorously digging out a spider's hole in full sun-light (ambient Ta of 27.3 °C, 10:30 hours), stopped on three occasions to open its mouth to gular pant. When it finally had caught the spider, it moved quickly to the closest shade. The gular panting by V. gouldii in this situation was probably an indication that the go¬ anna was approaching its critical maximum Tb in re¬ sponse to solar radiative heat gain and perhaps meta¬ bolic heat production from the vigorous activity. Hyoid breathing (or gular panting) occurs at a Tb of approxi¬ mately 39 °C for V. rosenbergi (King 1977, p.126) and 41 °C for V. griseus (Vernet et al. 1988b). Auffenberg (1981a) reports that gular panting indicates that V. komodoensis is approaching its critical maximum Tb. Measurements of the temperature in the shade about one metre above the ground, surface soil in full sun¬ light, under a grave cover and on top of a grave cover in the sun were occasionally recorded to examine the range of thermal environments available to a goanna at any one time. After mid-morning when the surface soil and grave covers had heated up, a wide variation in thermal environments were most often recorded (e.g. on Decem¬ ber 24 at 1015 hours, air temperature was 27 °C, under grave cover temperature was 24.1 °C, surface soil tem¬ perature was 39.6 °C and on top of a grave cover it as 46.4 °C) enabling the goanna to effectively regulate its body temperature by choosing a particular thermal en¬ vironment. Foraging site selection The general habitats and foraging micro-habitats se¬ lected by the two goannas were determined for periods of 41.05 and 26.55 hours respectively (Table 2). The activity area for each goanna (Fig 1) was determined by the smallest convex polygon that incorporated all of the locations that each lizard was found. The activity area of V. gouldii #1 (7.8 ha, excluding the bitumen road and the area outside the cemetery boundary) was approxi¬ mately 61% open ground with almost no leaf litter, 16% tree canopy with a ground cover of leaves, 10% grassed area, 11% leaf litter on open ground, and 2% shrubs with a leaf litter ground cover (grave covers were in¬ cluded in the total space calculations according to the immediate surrounds). The activity area of V. gouldii #2 (8.5 ha) was 42% open ground with almost no leaf litter, 34% grassed area, 7% a tree canopy and ground cover of leaves, 12% leaf litter on open ground and 5% shrubs with a ground cover of leaves (Fig 1). The activity areas for these two goannas were substantially larger than were recorded for six other female V. gouldii (2.29 ± se 0.23 ha, range 1.6 - 3.28 ha, mean mass = 380 g) mea¬ sured during the previous breeding season (Thompson, 1994). It is not known if there was a real difference in activity area sizes, or whether the presence of an Table 2. The percentage of time spent by two U. gouldii range of habitat types and selected microhabitat foraging sites. The data reported comes from 65 hours of observations and only includes the time between when V. gouldii commenced foraging until the goannas either returned to a burrow or observations ceased. Habitat type Percentage of time V. gouldii #1 V. gouldii #2 Between graves 35.6 39.7 On top or the side of graves 27.5 35.0 Under grave covers 12.2 9.9 Grassed areas 9.9 5.5 On the road verge 4.8 4.2 Treed areas 3.6 4.7 In yards of adjacent houses 3.2 0.0 In nature reserve 3.1 0.0 Up a tree 0.0 1.0 Micro-habitat foraging sites within activity area Between grave covers 47.7 77.9 Leaves under trees or shrubs 19.7 9.9 Under grave covers 13.7 4.2 Leaves in an open area 9.8 5.6 Grass in an open area 8.1 2.4 Grass under trees or shrubs 0.9 0.0 110 Journal of the Royal Society of Western Australia, 78(4), December 1995 observer increased the size of these activity areas, or the difference was due to the different nature of the data used to calculate the activity areas. In the earlier study, only one location was recorded for each goanna each day, and the total activity area was calculated from the smallest convex polygon that included all recorded loca¬ tions. In this study, the location of each goanna was known for up to half a day over a period of approxi¬ mately six weeks. Continuous observations are more likely to determine the real boundary of a goanna's ac¬ tivity area than would single daily observations, a factor not taken into account by the 'direct' polygon method used by Thompson (1994). Perhaps a better comparison might be with the activity area sizes calculated using the Jennrich & Turner (1969; females mean activity area of 3.91 ha ± se 0.36) approach, or the weighted ellipse (Samuel & Garton 1985: females mean activity area of 3.1 ha ± se 0.33). Both goannas spent the highest proportion of their foraging time (47.7% for #1 and 77.9% for #2) in the leaf litter between grave covers, without a tree or shrub canopy. These areas were estimated to be only 11 and 12% respectively of their total activity areas (Fig 1). The leaf litter between grave covers had often accumulated to a depth greater than 50 mm and many of the grave covers contained a hole under the edge, into which the goanna could retreat. It could be speculated that these two goannas selectively chose these area to forage be¬ cause the leaf litter provided a good food source (Mar¬ ginal Value Theorem; Charnov 1976) and the grave cov¬ ers were suitable protection from predators. The two other most-foraged habitats were the leaf litter under trees and shrubs (#1, 19.7% and #2, 9.9%), and under grave covers (#1, 13.7% and #2, 4.2%). Trees and shrubs with a ground cover of leaf litter under the canopy com¬ prised approximately 18 and 12% respectively of the ac¬ tivity areas of these two goannas. The shrubs and trees provided both shade in the hotter parts of the day, which enable the varanids to continuing foraging within their eccritic body temperature range, camouflage pro¬ tection and a source of food in the leaves. On one occasion, V. gouldii #2 climbed up in to a shrub that was approximately 1500 mm high and wide, and was observed to search the foliage, for a period of over 23 minutes. During this time, it climbed onto and searched almost every branch in the shrub that would support its body mass. Pianka (1970) reported that V . gouldii sought refuge in trees, but Thompson & Homer (1963) suggest that V. g. flauirufus was not arboreal. Thompson (1994) reports finding four V. gouldii up trees in Karrakatta Cemetery with 'no evidence to suggest that they had been chased into this situation'. He speculated that they might have been thermoregulating or foraging. In this study, the arboreal V. gouldii was actively forag¬ ing. Foraging patterns and behaviours V. gouldii #1 was seen to catch 64 prey items (spiders, insects and skinks) during a period of 35.5 hours of observation over 17 days. This was equal to one prey item being captured every 33.3 minutes. One scat was normally produced by each goanna each morning within an hour of commencing to forage. At the completion of excreting the scat, the V. gouldii would raise the middle and distal end of its tail, and drag its cloaca and the proximal end of the tail on the ground for 8-12 cm (Fig 2C). The two V. gouldii appeared to use both visual and auditory cues to detect their prey from a distance. When a goanna approached a patch of leaves, it would slow its forward speed of movement, move its head and neck slowly from side-to-side, using the snout to shift leaf litter while flicking the tongue in and out. The front feet were often used to scratch leaves away or dig into the ground for prey items. The use of the flicking forked tongue was taken as an indication that the lizard was using olfactory cues to find prey (Pianka 1982; Auffenberg 1984; Cooper 1989; Garrett & Card 1993). Once a prey item had been located, the V. gouldii would normally persist searching until the prey was captured, or in excess of five minutes. On one occasion, V. gouldii #2 was monitored while foraging in an exposed location (full sun-light), on a grassed area adjacent to a water pipe; having detected a prey item, it searched a small area for over 20 minutes before catching a small skink, then quickly moved to a shaded and more protected area. V. gouldii #1, on two occasions, was observed to jump approximately 300 mm into the air to try to catch an insect flying over is head; on both occasions it was unsuccessful. On five occasions, V. gouldii were observed to dig a triangular shaped hole, with the apex adjacent to a spi¬ der hole, to excavate the spider. The goanna used its front feet to dig the sand from the hole and the back feet to remove the small pile of sand that accumulated at the top of the hole. The back feet were never observed to assist in digging the hole itself. All prey that were caught were quickly oriented in the mouth and swal¬ lowed, without signs of chewing. The goanna would often wipe the sides of their jaws on the ground after swallowing their prey. Influence of birds on foraging sites The foraging of these two V. gouldii at Karrakatta Cemetery was influenced by the rainbow bee-eater (Merops ornatus) and the red wattlebird ( Anthochaera carunculata). A bee-eater was seen to swoop on a V. gouldii on 13 different occasions, and chase it away from their nesting holes (into the soft sand areas of the cem¬ etery). The goanna would lower its head during the final moments of such an attack, to avoid contact, and would quickly retreat to a more secure location. Bee- eaters were observed to actually contact a goanna on five occasions. On one occasion, the goanna ran ap¬ proximately 70 m after being harassed to find protection under a grave cover. The red wattlebird was also ob¬ served to swoop on a goanna on 12 occasions, contact with the lizard's head was seen on one occasion. It appeared that goannas had learnt to distinguish the par¬ ticular calls of the red wattlebird and the rainbow bee- eater from other noises, as they would adopt a vigilant posture upon hearing their calls and move away from the direction of the call or under a shrub or tree to avoid being swooped on. Poiani (1991) reports another wattle¬ bird ( Manorina melanophrys ) responding to the presence of V. gouldii with an alarm call. Ill Journal of the Royal Society of Western Australia, 78(4), December 1995 Movement speeds and distance travelled while foraging Little is known of the speed at which varanids forage, although the total distance moved in a given period is known for a number of large varanids. For example, Stebbins & Barwick (1968) report a 4.2 kg V. varius to travel a total distance of 2.9 km over a five day period in November, and Weavers (1993) reports the same species to travel up to 1.63 km day1 in south-eastern New South Wales. Vernet et al. (1988b) report V. griseus in the north¬ western Sahara desert to cover between 0.1 and 2.5 km day1; Pianka (1970, 1971, 1982) reports following V. gouldii for 'over a mile', V tristis for 'nearly a mile' and V. eremins for 'up to a kilometre' in a day, in semi-arid environments of Western Australia. Auffenberg (1981a) reports the much larger V. komodoensis to travel as much as 10 km day1, with a mean of 1.8 km day1, while Alberts (1994) reports V. albigularis in Namibia to travel as much as 4 km day1 during the mating season. V. gouldii , with a body mass of less than 600 g (measured in a previous year during October and November) trav¬ elled an appreciably shorter distance per day (mean 158 m day1; Thompson 1992). In this study of two V. gouldii, the average speed was 2.6 m min1 while foraging; V. gouldii #1 covered an estimated total distance of 7010 m in 42 hours of activity time (a mean speed of movement of 2.78 m min1), while V. gouldii #2 covered an esti¬ mated total distance of 2238 m in 23.5 hours of activity time (a mean speed of movement of 1.59 m min*1). The speed of movement between foraging patches was con¬ siderably higher, at 27.6 m min"1 (se = 1.92, range of 18.6 to 40.2 m min1), which was about ten times the overall mean speed of foraging. Goannas seldom moved in a straight line when moving between foraging sites and when foraging. Therefore, the mean speed of move¬ ment between foraging patches and the mean speed of movement while foraging are probably underestimates because the actual distance covered in each trial was greater than the linear distance recorded. However, the underestimation is probably quiet small, because most of the movement by the two goannas occurred between grave covers where circuitous movements are con¬ strained by the walls of the graves. Trials to establish movement speed between foraging sites were conducted adjacent to grave covers of known dimensions. Body postures V. gouldii frequently stopped during foraging to adopt a 'vigilant' posture. This posture was characterised by the body remaining motionless, the ab¬ domen held in either a prone or erect position, and the head and neck held high. The goanna would slowly rotate its head a couple of times to give a clear view of the surrounding area, while the posterior part of the abdomen and tail rested on the substrate. When a grave cover was nearby, the goanna would sometimes climb onto the side (Fig 2B) or top of the grave cover during foraging, adopt this vigilant posture, then return to the ground to continue to forage. On a few occasions, a goanna was observed to stand erect, by balancing on the hind legs and tail (Fig 2E). This stance appeared to be prompted by the need to see over some obstruction, as it occurred when the V. gouldii were between graves and their vision of the surrounding area was obstructed. A similar erect posture has been reported by Glazebrook (1977) for V. gouldii standing erect on its hind legs to obtain a better view of its surrounding. Auffenberg (1981a) reports adult V. komodoensis standing erect to take food suspended in the air, and young V. komodoensis standing erect to look over tall grass. King & Green (1993a) show a V. panoptes in a similar upright stance. These erect stances are different from the bipedal threat displays of V. mertensi (Murphy & Lamoreaux 1978), V. panoptes (King & Green 1993b, p257; Wilson & Knowles 1992, p318; Bennett 1992), V. gouldii (Grigg et al. 1985, p vii; Horn 1985) and for several other species (Beste & Beste 1977, p89) where the neck is arched, the gular pouch extended and the goanna hisses while rap¬ idly flicking its tongue in and out. The aggressive body postures reported by King & Green (1993a, b) for varanids generally, Auffenberg (1981a) for V. komodoensis, Bennett (1992) for V. panoptes, Stanner (1985) for V. griseus, Johnson (1976), Murphy & Lamoreaux (1978) for V. mertensi, Pianka (1970) and Stirling (1912) for V. gouldii, and Daltry (1991) for V. salvator were not seen for the V. gouldii at Karrakatta Cemetery. Rather, these goannas appeared to adopt an avoidance behaviour when threatened. Tail swipes (4 and 8 in quick succession) were ob¬ served on two occasions. On neither occasion were they associated with an inflated gular pouch and abdomen, or a hissing sound, as reported by King & Green (1993a) to represent an aggressive stance. On the first occasion, V. gouldii #1 approached to within a couple of metres of the observer, raised and turned its head slightly to the side and moved its tail from side-to-side four times (Fig 2G, H and 1). On the second occasion, V. gouldii #2 was digging for prey. I have observed a similar 'tail-swipes' behaviour for a male V. gilleni courting a female of the same species. Additional observations are necessary be¬ fore this behaviour can be accurately interpreted. V. gouldii are primarily quadrupedal and terrestrial. After they had emerged from their overnight burrow, elevated their Tb, and commenced moving around their foraging site, they were observed to 'drag' a substantial proportion of their tail on the ground. As the morning progressed, the tail was often lifted off the ground, or only the distal 20% was dragged on the ground (Fig 2F). A lateral sigmoidal trotting (diagonally opposite feet ad¬ vancing together) while walking was often evident (Fig 2D). The head was normally swung slowly from side- to-side, and at the same level as the body (Fig 2F). The walking posture of V. gouldii was therefore similar to that described by King & Green (1993a) for varanids generally, and by Auffenberg (1981a) for V. komodoensis, except that V. gouldii would sometimes lift their tail off the ground while walking and not leave a tail drag mark. The dorsal surfaces of the rear toes were often dragged along the ground extending the length of their spoor, similar to that reported for V. komodoensis (Padian & Olsen 1984). On twelve occasions, V. gouldii #1 was observed, while stationary, to raise its tail so that the distal end was vertical. The same tail raising posture was not ob¬ served for goanna #2. A similar stance has been reported for V. tristis (Christian 1981); its long tail was curled in a long arc so that the tip would almost touch the back of the neck. The role of this body stance is unknown. 112 Journal of the Royal Society of Western Australia, 78(4), December 1995 Detection by a goanna of a person, dog or other source of potential danger resulted in one of two initial responses. The goanna would either flatten its body on the ground and remain motionless until the source of danger had passed, or it would slowly move behind a nearby object and would remain motionless. Some two to ten minutes later, it would poke its head around the object to determine if the danger had passed. Alterna¬ tively, if the goanna had just emerged from a burrow or a burrow was nearby, it would return to this burrow instead of hiding behind an object. V. gouldii appeared to run to a burrow only if the source of danger came very close. Acknowledgements: Funding for this project was provided by Edith Cowan University. Constructive review of early drafts of the manuscript by P C Withers, M Bull and D King were most appreciated as they im¬ proved the final manuscript, The field assistance of S and W Thompson in undertaking most of the observational study was very much appreciated. Access to the Karrakatta Cemetery was approved by the Metropolitan Cemeteries Board, and the assistance of cemetery staff is acknowledged. All goannas were caught under licence issued by the Department of Con¬ servation and Land Management. Animal experimentation was done with the approval of the Animal Welfare Committee of Edith Cowan Univer¬ sity. References Alberts A 1994 Off to see the lizards: lessons from the wild. Viva¬ rium 5:26-28. Auffenberg W 1981a The Behavioural Ecology of the Komodo Monitor. University of Florida Press, Gainesville. Auffenberg W 1981b Combat behaviour in Varanus bengalensis (Sauria: Varanidae). Journal of the Bombay Natural History Society 78:54-72. Auffenberg W 1984 Notes on the feeding behaviour of Varanus bengalensis (Sauria: Varanidae). Journal of the Bombay Natu¬ ral History Society 80:286-302. Auffenberg W 1988 Gray's Monitor Lizard. University of Florida Press, Gainesville. Auffenberg W 1994 The Bengal Monitor. University of Florida Press, Gainesville Auffenberg W, Arain Q N & Khurshid N 1991 Preferred habitat, home range and moment patterns of Varanus bengalensis in southern Pakistan. Mertensiella 2:7-28. Bennett D 1992 A note on Varanus panoptes rubidus (Storr 1980) in Wanjarri, Western Australia. British Herpetological Society Bulletin 39:28-30. Beste H & Beste J 1977 Goanna lizard. In: Wild, wild world of animals: kangaroos and other creatures from down under. Time-Life Films, USA, p89. Chamov E L 1976 Optimal foraging, the marginal value theorem. Theoretical Population Biology 9:126-136. Christian T 1981 Varanus tristis - a variable monitor. Herpetofauna 12:6-12. Cooper W E 1989 Prey odor discrimination in the varanoid liz¬ ards Heloderma suspectum and Varanus exanthematicus. Ethol¬ ogy 81:250-258. Cowles R B 1930 The life history of Varanus niloticus (Linnaeus) as observed in Natal, South Africa. Journal of Entomology and Zoology 22:1-31. Daltry J 1991 The social hierarchy of the water monitor, Varanus salvator. Hamadryad 16:10-20. Garrett C M & Card W C 1993 Chemical discrimination of prey by naive neonate Gould's monitors Varanus gouldii. Journal of Chemical Ecology 19:2599-2604. Glazebrook R 1977 Old man goanna. North Queensland Natural¬ ist 44:4-6. Green B & King D 1978 Home range and activity patterns of the sand goanna, Varanus gouldii (Reptilia: Varanidae). Australian Journal of Wildlife Research 5:417-424. Grigg G, Shine R & Ehmann H 1985 Biology of Australian Frogs and Lizards. Surrey Beatty, Sydney. Horn H-G 1985 Beitrage zum verhalten von Waranen: Die ritualkampfe von Varanus komodoensis Ouwens, 1912 and V. semiremex Peters, 1869 sowie die imporiephasen der ritiualkampfe von V. timorensis timorensis (Gray, 1831) und V. t. sirnilis Mertens, 1958. Salamandra 21:169-179. Jennrich R I & Turner F B 1969 Measurement of non-circular home range. Journal of Theoretical Biology 22:227-37. Johnson C R 1976 Some behavioural observations on wild and captive sand monitors, Varanus gouldii (Sauria: Varanidae). Zoology Journal of the Linnean Society 59:377-380. King LI 1977 Temperature regulation in the sand goanna Varanus gouldii (Gray). PhD Thesis, University of Adelaide, South Aus¬ tralia. King D 1980 The thermal biology of free-living sand goannas (Varanus gouldii) in southern Australia. Copeia 1980:755-767. King D & Green B 1993a Goanna: The Biology of Varanid Liz¬ ards. NSW University Press, Sydney. King D & Green B 1993b Family Varanidae. In: Fauna of Austra¬ lia. Vol. 2A Amphibia & Reptilia (eds C J Glasby, G J B Ross, & P L Beesley). Australian Government Printing Service, Canberra, 253-260. Murphy J B & Lamoreaux W E 1978 Threatening behaviour in Mertens water monitor Varanus mertensi (Sauria: Varanidae). Herpetologica 34:202-205. Padian K & Olsen P E 1984 Footprints of the Komodo monitor and the trackways of fossil reptiles. Copeia 1984:662-671. Pianka E R 1970 Notes on the biology of Varanus gouldii flavirufus. Western Australian Naturalist 11:141-144. Pianka E R 1971 Notes on the biology of Varanus tristis. Western Australian Naturalist 11:180-183. Pianka E R 1982 Observations on the ecology of Varanus in the Great Victoria Desert. Western Australian Naturalist 15:1-13. Pianka E R 1994 Comparative ecology of Varanus in the Great Victoria Desert. Australian Journal of Ecology 19:395-408. Poiani A 1991 Anti-predator behaviour in the Bell Miner Manorina melanophrys. Emu 91:164-171. Samuel M D & Garton E O 1985 Home range: a weighted normal estimate and tests of underlying assumptions. Journal of Wildlife Management 49:513-519. Sokolov V E, Sukhov V P & Chernyshov Y M 1975 A radiotelemetic study of diurnal fluctuations of body tempera¬ ture in the desert monitor (Varanus griseus). Zoological Zhurnai 54:1347-1356. Stanner M 1985 Monitoring the desert monitor. Israel - Land and Nature 16:142-145. Stebbins R C & Barwick R E 1968 Radiotelemetric study of ther¬ moregulation in a lace monitor. Copeia 1968:541-547. Stirling E C 1912 Observations on the habits of the large central Australian monitor (Varanus giganteus), with a note on the "fat bodies" of this species. Transactions and Proceedings of the Royal Society of South Australia 36:26-34. Thompson D F & Homer W 1963 A preliminary' account of the herpetology of the Great Sandy Desert of central Western Australia. Proceedings of the Royal Society of Victoria 77:217- 237. Thompson C» G 1992 Daily distance travelled and foraging areas of Varanus gouldii (Reptilia: Varanidae) in an urban environ¬ ment. Wildlife Research 19:743-753. Thompson G 1994 Activity area during the breeding season of Varanus gouldii (Reptilia: Varanidae) in an urban environment. Wildlife Research 21:633-641. Vemet R, Lemire M, Grenot C & Francaz J-M 1988a Ecophysi- ological comparisons between two large Saharan lizards. 113 Journal of the Royal Society of Western Australia, 78(4), December 1995 Uromastix acanthinurus (Agamidae) and Varanus griseus (Varanidae). Journal of Arid Environments 14:187-200. Vemet R, Lemire M, & Grenot C 1988b Field studies and water balance of a desert monitor Varanus griseus (Reptilia, Varanidae). Journal of Arid Environments 15:81-90. Weavers B W 1983 Thermal ecology of Varanus varius (Shaw), the Lace Monitor. PhD Thesis, Australian National University, Canberra. Weavers B W 1993 Home range of male lace monitors, Varanus varius (Reptilia: Varanidae), in South-eastern Australia. Wild¬ life Research 20:303-313. Wilson S K & Knowles D G 1992 Australia's Reptiles: A Photo¬ graphic Reference to the Terrestrial Reptiles of Australia. Cornstalk, Sydney. 114 Journal of the Royal Society of Western Australia, 78:115-120, 1995 THE ROYAL SOCIETY OF WESTERN AUSTRALIA INCORPORATED Constitution and Rules and Regulations (Revised March 1996) Constitution 1. The scientific society known as 'The Royal Society of Western Australia Incorporated" (hereinafter re¬ ferred to as "The Society") has the objects, purposes, and powers hereinafter mentioned. Objects, purposes, and powers 2. To promote and assist in the advancement of science. 3. To print and publish, or join in printing and pub¬ lishing, one or more journals, periodicals, books, newsletters or other documents relating to science. 4. To promote, produce and to hold, or join in promot¬ ing, producing and holding meetings, lectures, con¬ ferences, symposia, excursions and exhibitions in furtherance of any objects of The Society. 5. To found, subsidise or contribute to, for the benefit of any institutions or persons, scholarships, grants, prizes, or monetary or other rewards in connection with all or any of the objects or purposes of The Society. 6. To acquire by purchase, lease, exchange, hire, or by way of loan or otherwise, any real or personal prop¬ erty in furtherance of any objects of The Society. 7. To invest, sell, lease, hire, mortgage, charge, lend, surrender or otherwise dispose of, or deal with all or any part of the assets or property, real or per¬ sonal, of The Society, or to borrow money and grant security therefore and to liquidate, redeem or dis¬ charge any obligations undertaken in furtherance of any objects of The Society. 8. To improve, develop or extend all or any of the property or rights of The Society, also to build, erect or alter any buildings or erections, and to furnish, fit up, and maintain the same and provide fittings, equipment and appliances. 9. To amalgamate, co-operate or affiliate with any other society, association or body having objects wholly or in part similar to those of this Society. 10. To protect or assist in protecting the interests or rights of any member or members of The Society. Rules and Regulations Membership 11. The Society shall consist of members who are all individuals involved, assisting or interested in the promotion and/or advancement of science. Mem¬ bers shall be divided into the classes hereinafter mentioned, and the members who are, have been, or may be duly elected as members of such respec¬ tive classes shall have the rights and privileges here¬ inafter specified as appertaining to such respective classes, subject always to the due and punctual pay¬ ment of the annual subscription or other sum here¬ inafter provided to be paid by the respective mem¬ bers of each class, namely: (a) Ordinary Members: An Ordinary Member if financial shall have the following rights and privileges: (i) To be present and to speak and vote at Ordinary, Special or Annual General Meet¬ ings of The Society, to vote in postal ballots conducted by The Society, and to attend excursions, meetings, lectures or other © Royal Society of Western Australia 1995 activities arranged from time to time by The Society. (ii) To be eligible for election as a member of the Council hereinafter referred to and also to any office or position in The Society. (iii) To submit to the Council for consideration for publication papers prepared on any sci¬ entific subject. (iv) To receive the Journal of The Society. (v) To receive other publications or documents issued by The Society, upon such condi¬ tions as the Council may from time to time determine. (vi) To borrow books, periodicals, papers, or other documents belonging to The Society, subject to the approval of the Council. (vii) To propose or second candidates for ad¬ mission as Ordinary, Associate or Student Members of The Society. (viii)At any meeting held by The Society to in¬ troduce visitors. Such visitors shall not be 115 Journal of the Royal Society of Western Australia, 78(4), December 1995 entitled to vote at any such meeting or ex¬ cursion, but may express opinions on any matter under discussion at the invitation of the chairman or leader. (ix) To propose or second any Honorary or financial Ordinary Member for election to the Council. (b) Associate Members: Persons elected as Associ¬ ate Members shall be entitled to attend and speak at general meetings, excursions and other activities conducted by The Society, but shall not be entitled to vote at any meeting, and if any such member should vote, then that vote shall not be counted. Associate Members shall be entitled to submit papers for publication, but shall not receive the Journal of the Society ex¬ cept at a price fixed by the Council from time to time and subject to the approval of the Council, may borrow books, periodicals or other docu¬ ments belonging to The Society, but shall not be entitled to be proposed for, or elected to, any office in The Society, nor to propose or second others for membership or office in The Society. (c) Student Members: The Council may, on appli¬ cation to be made each year in such form it may require, admit as Student Members, persons who are undertaking full-time studies in West¬ ern Australia, or as agreed by Council. Student Members shall have the same rights and privi¬ leges, and shall be subject to the same restric¬ tions, as Associate Members, except that they shall receive the Journal of the Society. (d) Honorary Members: The Society at its Annual General Meeting in any year may, on the rec¬ ommendation of the Council, admit as Honor¬ ary Members persons distinguished in Science or as patrons thereof, but only in so far as the number of such members shall not henceforth at any time exceed five percent of the total Ordi¬ nary, Associate and Student membership of the Society. Honorary Members shall have the same rights and privileges as Ordinary Members, but without liability for any subscription. (e) Honorary Associate Members: The Society at its Annual General Meeting in any year may, on the recommendation of the Council, admit as Honorary Associate Members persons inter¬ ested in science, but only in so far as the num¬ ber of such members shall not exceed five per¬ cent of the total Ordinary, Associate and Stu¬ dent membership of the Society. Honorary As¬ sociate Members shall have the same rights and privileges, and be subject to the same restric¬ tions, as Associate Members, but without liabil¬ ity for any subscription. (f ) Corporate Members: Corporate membership may be made available by the Council to organisations under such terms and with such rights and privileges as the Council may deter¬ mine. 12. Every person desirous of becoming an Ordinary or Associate Member of The Society shall make application in the form prescribed from time to time by the Council and shall be proposed by an Ordi¬ nary or Honorary Member of The Society and be seconded by at least two other Ordinary or Honor¬ ary Members of The Society, to each of whom the applicant is known personally. Each application shall be accompanied by the subscription applicable to the class of membership sought and shall be lodged with one of the Secretaries. Each Ordinary and Honorary Member of The Society shall be in¬ formed of the application forthwith by such means as the Council may from time to time determine. The Council shall accept or reject each candidate at a meeting of the Council held not less than one month and not more than six months after issue of advice of the candidate's application to the Ordi¬ nary and Honorary Members. Admission shall be made by a two thirds majority of those Council members present and voting. 13. The Secretary shall inform each candidate of his ad¬ mission or non-admission to the class of member¬ ship sought. Upon admission. Ordinary or Associ¬ ate Members shall receive a copy of The Society's Constitution and Rules and Regulations. Their membership shall be deemed to apply to the whole of the financial year in which they were admitted so that, subject to their availability, they shall be en¬ titled to receive all publications and documents is¬ sued during that year to all members of their class. 14. Members of The Society, of whatever class, shall be bound to observe and perform, and not commit any breach of, the Rules and Regulations of The Society from time to time in force. 15. Members shall notify the Secretary in writing of their address for receiving notices or communica¬ tions to be forwarded to them, and shall from time to time notify in writing any change in such ad¬ dress. If any member shall fail to give such notifica¬ tion they shall have no claim against The Society for publications or notices of the meetings or other ac¬ tivities of The Society; notification is deemed to have occurred at the time of posting. 16. Members, on paying to The Society all subscriptions or moneys owed by them and returning all books, papers, manuscripts, or property of The Society which they may have borrowed or received, may resign their membership by giving notice in writing to the Secretary of The Society; and any members ceasing by resignation, death, or otherwise to be a member of The Society shall not, nor shall their rep¬ resentatives have any claim upon or interest in the funds or property of The Society; but nothing herein contained shall prejudice the right of The Society to recover any moneys owing or property of The Soci¬ ety borrowed, held, or received by members at any time. 17. Members whose subscription may be in arrears for a period of more than two years shall, on a resolu¬ tion of the Council being at any time thereafter passed, be declared to be no longer members and thereupon shall cease to have the rights and privi¬ leges to which they may have been entitled, pro¬ vided always that nothing herein contained shall 116 Journal of the Royal Society of Western Australia, 78(4), December 1995 prejudice the right of The Society to recover from such members all moneys or subscriptions due, ow¬ ing, or payable by them up to the date of such ter¬ mination of their membership, and also to recover all books, papers, manuscripts or property belong¬ ing to The Society which may be held or have been received by such members at any time. 18. Members in the respective classes of membership in The Society shall use their best efforts to promote the objects of The Society and shall not do or com¬ mit any act, deed, or thing which may be deemed by the Council to be prejudicial to the interests of The Society. 19. The Council may, by an affirmative vote of two thirds of its total membership, remove or suspend from membership or expel any member of The Soci¬ ety. Notice of such removal, suspension, or expul¬ sion, shall be sent by registered post to the last known address of the member concerned within seven days after the decision of the Council. Mem¬ bers against whom such decision of removal, sus¬ pension, or expulsion shall be made, shall be en¬ titled to appeal to a Special General Meeting of The Society by notice to be forwarded by them in writ¬ ing, addressed to the Secretary within two months after the date of such removal, suspension, or expul¬ sion, stating in such notice the grounds of appeal. It shall be the duty of the Council to summon a Spe¬ cial General Meeting of The Society for the purpose of considering any such appeal, and of hearing statements by any Member of the Council or by the member who may have been removed, suspended or expelled, and if a majority of the members present at such meeting uphold the decision of the Council, then the decision of the Council shall be confirmed, but if such majority shall uphold the ap¬ peal, then the decision to remove, suspend, or expel such member shall be set aside. No members shall be entitled to exercise such right of appeal after the expiration of the said period of two months. In the event of any such removal, suspension, or expulsion taking effect, any members concerned shall remain liable for all moneys or subscriptions due or pay¬ able by them as at the date of such removal, suspen¬ sion, or expulsion, and for the return of all property belonging to The Society. Subscriptions 20. The subscription for each ordinary membership shall be set from time to time on recommendation of the Council and approval by a resolution passed by a two-thirds majority of Ordinary Members voting at any Ordinary or Annual General Meeting of The Society, of which at least twenty eight days notice has been given, and in which notice the proposed alterations have been specified. Ordinary Members whose subscriptions are not in arrears shall be granted Ordinary Membership for Life upon payment of a fee of twenty years sub¬ scription at the rate of the current year. 21. Associate Membership shall not be more than 60 percent of the Ordinary Membership rate. 22. Student Membership shall not be more than 60 per¬ cent of the Ordinary Membership rate. 23. The financial year of The Society shall be from the first day of July in each year to the thirtieth day of June in the following year. 24. All subscriptions shall be payable in advance and shall become due on the first day of July in each year. Any member whose subscription is unpaid six months after the due date in any year shall be deemed to be not financial. After this date a late fee shall be payable at a rate determined by Council from time to time. Management 25. The management of the business and affairs of The Society shall be vested in a Council consisting of a President, two Vice-Presidents, Treasurer, three Sec¬ retaries, Librarian, Editor, Journal Manager, Imme¬ diate Past President and seven Ordinary or Honor¬ ary Members of The Society. 26. All members of the Council shall be elected at the Annual General Meeting of The Society to be held on a date normally in July at a place to be decided by the Council and serve for a term of one or two years. For the purpose of such election the Council then in office shall submit, at an Ordinary or Special General Meeting held not less than one calendar month before the Annual General Meeting, a list of names of Honorary or financial Ordinary Members proposed by the Council for election indicating the length of term of office and shall appoint a Return¬ ing Officer. Any Ordinary or Honorary Member present at such Ordinary General Meeting shall be entitled to propose another financial Ordinary or any Honorary Member to hold any position on the Council, and if such nomination be duly seconded and if the candidate nominated has signified will¬ ingness to accept office if elected, then the name of such nominee shall be added as a candidate. Fur¬ ther, any Ordinary or Honorary Member of The So¬ ciety may lodge with the Returning Officer within seven days after that Ordinary or Special General Meeting, preceding the Annual General Meeting, a nomination in writing in favour of any financial Or¬ dinary or any Honorary Member for any position on the Council. Such written nomination shall be seconded by another Ordinary or Honorary Mem¬ ber, and shall carry a statement by the candidate nominated as to their willingness to accept the office if elected. Upon receipt of such nomination, the Returning Officer shall add the name of the nomi¬ nee as a candidate for the Council. 27. If the number of members duly proposed for elec¬ tion does not exceed the number of vacancies for the various offices, the Chairman of the Annual General Meeting shall declare the persons nominated as duly elected. If the number of nominations for any office exceeds the number of vacancies, an election shall be held and the results ascertained by preferential ballot. For the purpose of such election, a ballot paper containing the names of all persons duly pro¬ posed for all contested positions on the Council 117 Journal of the Royal Society of Western Australia, 78(4), December 1995 shall be prepared by the Returning Officer and sent to all Ordinary and Honorary Members. Such members desirous of voting shall cause to be deliv¬ ered to the Returning Officer such ballot papers, duly completed in accordance with any instructions on voting procedure before the formal opening of the Annual General Meeting. The Annual General Meeting shall appoint at least two scrutineers, who shall examine the procedures of the Returning Offi¬ cer in the counting of votes and who shall report the results of the count to the Chairman of the Annual General Meeting. Such Chairman shall announce the result of the ballot at the Annual General Meet¬ ing or at the next Ordinary or Special General Meet¬ ing of The Society. 28. All members of the Council shall be eligible for re- election. 29. Any casual vacancy in the Council may be filled by resolution of the Council, and any member so ap¬ pointed shall hold office until the next annual elec¬ tion of Council. Any vacancy not filled at the an¬ nual election shall be deemed a casual vacancy. 30. The Council may define the duties of the three Sec¬ retaries and may add any distinguishing word to the title of any or all Secretaries in accordance with the nature of the duties to be performed. 31. The Council shall meet at least once in each month from February to November inclusive in each year (unless otherwise decided by Council), at such times and places as may be appointed by the President, or in the President's absence, by one of the Vice-Presi¬ dents, and due and sufficient notice shall be previ¬ ously sent to each member of the Council. 32. A quorum for a meeting of the Council shall be the President or one of the Vice-Presidents, a Secretary who may be specially appointed for that meeting, and four other members of the Council, and no busi¬ ness shall be transacted at any Council meeting un¬ less such quorum is present. 33. If any members of the Council (including holders of offices) shall fail to attend three consecutive meet¬ ings of the Council without satisfactory explanation or reason, or without leave of absence having been first granted to them, then the position or office of such member may by resolution of the Council be declared vacant, and on passing of such resolution they shall cease to be a member of Council and holder of any office to which they may have been elected. 34. Council shall inform members of activities of the Society including deliberations of the Council by publishing and distributing to the members a Pro¬ ceedings following each Ordinary or Annual Gen¬ eral Meeting. 35. The Council shall present at each Annual General Meeting a report giving a review of the work of The Society during the preceding year and some details and information with regard to its progress and affairs, and shall publish this report. 36. The Council may appoint a Committee from among the financial members of the Society, and by two thirds majority resolution may include non-mem¬ bers in such a Committee. A Committee may be of such number as the Council may decide and may consider and make recommendations on any subject on which Council may require advice, provided that such a Committee includes at least one Council Member. Each Committee so appointed shall be reviewed at regular intervals to be determined by the Council. 37. The Council may from time to time make, alter, and repeal by-laws to enable it more effectually to carry out the management of the affairs of The Society, and to regulate the conduct of members and assist in the protection of its property, and for such pur¬ poses as may be calculated to advance the welfare of the Society, provided that such changes are not inconsistent wTith these Rules and Regulations. 38. The Council, without limiting its general powers of management and carrying on the business and af¬ fairs of The Society, may exercise and do all things necessary except such as may be required or di¬ rected to be exercised by General Meetings, includ¬ ing power to appoint and remove all or any officers, assistants, employees, or others deemed by the Council to be necessary in connection with the work of The Society., and that with or without remunera¬ tion and upon such terms and conditions as the Council may think fit. The Council may also del¬ egate all or any of its powders or authorities to any committee or sub-committee from time to time, and may pay all or any expenses or liabilities incurred from time to time and take any steps or proceedings which may be deemed desirable for the purpose of carrying out or securing the fulfilment of any of the objects or purposes of The Society. President and Vice-Presidents 39. The duties of the President shall be to preside at all meetings of The Society and Council, and regulate all the proceedings therein and generally to execute or see to the execution of the Rules and Regulations and by-laws of The Society. In the case of an equal¬ ity of votes at any meeting, the President or Vice- President or member presiding shall have a casting vote in addition to a deliberative vote. 40. In the absence of the President from any meeting of the Council, the President's place shall be filled by one of the Vice-Presidents. In the absence of the President from any other meeting or excursion of The Society, the President's place shall be filled by one of the Vice-Presidents or by an Ordinary or Honorary Member of The Society elected as Chair¬ man or leader by the Ordinary and Honorary Mem¬ bers there present. Duties of Officers 41. The Secretaries shall conduct correspondence, ar¬ range meetings, cause notices of meetings and Pro¬ ceedings to be sent out, keep adequate records of all meetings, membership and other activities of The Society, and generally perform such duties as are usually assigned to persons holding such office and 118 Journal of the Royal Society of Western Australia, 78(4), December 1995 comply with the directions, requests, or instructions issued from time to time by the Council. 42. The Treasurer shall keep a correct record of all re¬ ceipts and disbursements, including subscriptions and all moneys received for the benefit of The Soci¬ ety, and shall pay those moneys forthwith to the credit of The Society at its bankers. It shall be the duty of the Treasurer to pay all accounts passed by the Council. No moneys shall be withdrawn from the bank account except by cheque signed by any two of a group of three or more members of the Council which shall include the Treasurer and two or more other members of Council so authorized by Council. The Treasurer shall cause the books of The Society to be posted regularly, and shall bring his books to balance as on the thirtieth day of June in each year, or on such other date as the Council may from time to time decide. 43. The Treasurer shall annually submit to an Auditor or Auditors all books and accounts kept by the Trea¬ surer in connection with the affairs of The Society, made up to the date last mentioned. The Society shall, not later than its Ordinary General Meeting immediately prior to the Annual General Meeting, appoint an Auditor or Auditors, but if such Ordi¬ nary General Meeting shall fail to make such ap¬ pointment, then the Council may appoint an Audi¬ tor or Auditors. 44. It shall be the duty of the Auditor or Auditors of The Society to submit a written report each year on the financial affairs of The Society through the Council to an Ordinary, Special or Annual General Meeting of The Society. 45. The Librarian shall maintain records and generally manage the books, periodicals and documents of The Society according to the directions of the Coun¬ cil. 46. The Editor shall supervise the production of The Society's Journal and such other publications as the Council may direct. 47. The Journal Manager shall supervise the distribu¬ tion of The Society's Journal and other publications as authorized by Council. 48. Other members of the Council shall assist in the general management of The Society according to the needs which may arise from time to time. Ordinary General Meetings 49. Ordinary General Meetings of The Society shall be held at 8 p.m. on the third Monday of the months March to June and August to December inclusive in each year, unless the Council determines otherwise, but at least three Ordinary General Meetings shall be held in each financial year. Notice of each Ordi¬ nary General Meeting shall be sent to all members of each class. A quorum for an Ordinary General Meeting shall be seven Ordinary or Honorary Mem¬ bers personally present. Conduct of an Ordinary General Meeting shall be at the discretion of the President or Chairman elected by such meeting. Annual General Meetings 50. The Annual General Meeting of The Society shall, unless the Council determines otherwise, be held on the third Monday of July in each year at a time and place determined by the Council. Notice of each Annual General Meeting shall be sent to all mem¬ bers of each class at least one calendar month before such meeting. 51. The proceedings of the Annual General Meeting, un¬ less otherwise determined by the Council, shall be as follows: (a) Presentation of the minutes of the previous An¬ nual General Meeting. (b) Reading of nominations of candidates for Coun¬ cil, and, if necessary, appointment of scrutineers, and counting of votes. (c) Presentation of the Annual Report of the Coun¬ cil. (d) Presentation of the Balance Sheet, Statement of Accounts and (if available) Auditor's Report. (e) Report (if necessary) of the scrutineers on the ballot and declaration of the results by the retir¬ ing President. (f) Address by the retiring President. (g) Installation of the new President. (h) Any other business of which notice may have been given prior to or agreed during the meet¬ ing to be considered. 52. A quorum for transaction of business at an Annual General Meeting shall be seven Ordinary or Honor¬ ary Members of The Society personally present. Special General Meet mgs 53. Special General Meetings of The Society may be called by the Council whenever it may deem such meeting expedient, or on the requisition of ten Ordi¬ nary or Honorary Members made in writing to one of the Secretaries and specifying the purpose for which the meeting is required. Upon receipt of such requisition, that Secretary shall call the meeting within not less than seven days nor more than twenty eight days. Notice of such meeting shall be sent to all members of each class. The meeting shall be chaired by the President or a Vice-President or, in their absence, an Ordinary or Honorary Member elected by the meeting. 54. A quorum for a Special General Meeting of The So¬ ciety shall be seven Ordinary or Honorary Members personally present. The Journal of the Society 55. The Society shall publish a Journal at least once a year, in which papers communicated to The Society during or before that year may be printed. The Journal shall be printed in such form as decided by the Council. 56. Every paper intended to be published in the Society's Journal must be sent to the Editor for con- 119 Journal of the Royal Society of Western Australia, 78(4), December 1995 sideration by Council. The Editor shall Chair a Pub¬ lications Committee, if such is appointed by Coun¬ cil, to provide assistance and advice to the Editor. 57. The Editor shall seek an expert opinion from any person or persons the Editor may select as referees to judge the suitability of any paper for publication. The Editor shall recommend to the Council whether or not a paper submitted for publication should be accepted for the Society's Journal. A paper of which the Editor is author or co-author shall be edited by a member of Council, so authorised by Council, who is not a co-author of the paper. 58. It shall be the duty of the Council to decide, on the recommendation of the Editor or authorised Council member, whether or not a contribution shall be ac¬ cepted for publication. 59. Publication in the Society's Journal shall be avail¬ able to all categories of members. 60. The original copy of every paper accepted for publi¬ cation by The Society, with its illustrations, shall be¬ come the property of The Society, unless the Coun¬ cil decides otherwise. Authors shall not be at liberty to publish elsewhere papers submitted to The Soci¬ ety for publication in its Journal, unless permission for doing so is given by the Council, or unless the Society fails to publish the paper in the Journal of the year in which it is accepted or of the succeeding year, or does not accept the paper for publication. 61. The published price of the Journal shall be fixed by the Council from time to time. 62. Offprints of papers shall be available to authors on such terms as shall be decided from time to time by the Council. The Medal of the Royal Society of Wester n Australia 63. A Medal, for which Council will call for nomina¬ tions through its Proceedings, shall be awarded by the Council every fourth year or at such other times or periods as the Council may from time to time decide for distinguished work in science connected with Western Australia. The Council shall appoint a Medal Committee consisting of five members of the Council to assess the nominations and recom¬ mend a recipient of the Medal. The recipient shall, on the awarding of the Medal, be requested to de¬ liver a public address to be known as the Royal So¬ ciety of Western Australia Medal Lecture. Formation of Sections 64. Sections may be formed for the purpose of any par¬ ticular branch of science. Any member of The Soci¬ ety may be enrolled as a member of one or more sections. Each section shall appoint a Chairman and Secretary, with the approval of Council. Sections shall not incur expenditure without first obtaining the approval of the Council. Any communication to a section may be presented subsequently at a gen¬ eral meeting of The Society. Common Seal 65. The Common Seal of The Society shall be in the custody of the President or one of the Vice-Presi¬ dents, and the President or any one Vice-President shall respectively be authorized to use the same, and when required to be affixed to any deed, document, or writing, shall be so affixed and signed by either the President or one of the Vice-Presidents and countersigned by a Secretary of The Society. Interpretation of the Constitution and Rules and Regulations 66. The Council of The Society shall be the sole author¬ ity for the interpretation of the Constitution and of the Rules and Regulations of The Society, and the decisions of the Council on questions of interpreta¬ tion shall be final and binding on all members. Alteration of the Constitution and Rules and Regulations 67. The Constitution or the Rules and Regulations or any of them may be amended, altered, enlarged or repealed from time to time by a resolution passed by 75 percent of Ordinary or Honorary Members voting in a postal ballot conducted by the Council. Notice of intention to conduct such a ballot shall be given with notice of an impending Ordinary or An¬ nual or Special General Meeting of The Society and the proposed amendment or amendments shall be presented at that meeting at least one calendar month before the ballot is held. Non-Profit Organisation 68. The Society is operated as a non-profit organisation. The assets and income of The Society shall be ap¬ plied solely in furtherance of its above mentioned objects and purposes and no portion shall be dis¬ tributed directly or indirectly to the members of The Society except as bone fide compensation for ser¬ vices rendered or expenses incurred on behalf of The Society. Winding up of The Society 69. The Society may be wound up by a resolution to be passed by a four fifths majority of the Ordinary and Honorary Members of The Society present and vot¬ ing at a Special General Meeting summoned for such purpose, whereof at least twenty eight days notice shall be given. If a resolution to wind up The Society be passed, and there remains, after satisfac¬ tion of all The Society's debts and liabilities includ¬ ing the costs, charges and expenses of that winding up, any property whatsoever, the same shall not be paid to or distributed among the members but shall be given or transferred to some other association or associations which are themselves incorporated un¬ der the Associations Incorporation Act 1987 and are therefore non-profit organisations, or for charitable purposes, or both, as determined by resolution of the members. 120 Journal of the Royal Society l*clsr of Western Australia cuRtin University of Technology Perth Western Australia Museum of Victoria 401 Geoscience de Laeter Symposium on Isotope Science Curtin University November 1995 Standard: Volume 79 Part 1 March 1996 ISSN 0035-922X 3 < C .2 £ w M) C/5 o d o S- c £ o S H ~ ^ < § o 0 g£ H '£ £ < C c *- IP ^ c QJ U QJ ^ rs U -j V 73 a*N a> X> a> > o GO sD £ o 'cT 0- £ 3 Jz — j tn r> "c Q T3 ca c 0) g £ - .«> .2 c -* '5 CA ra 2 3 ^ C Q ^ 2 2 (Sg m 2 ^ D eg Si ^ 55 2 A O H 2 ^ u-1 Q U £ UJ £ V~) f- O $ £ £ ca ^ C i i * | ££ gl £3 c< ?5 -jg os — - ^ r/! I g n 2 h -c y & ^ -o = -3 D 2 X 2 0 | h w u- > Proceedings of The de Laeter Symposium on Isotope Science1 edited by Philip C Withers and Kevin J R Rosman Contents Foreword K J R Rosman de Laeter Symposium Sir M Oliphant Science to de Laeter: What now comes later. H S Peiser High-accuracy mass spectrophotometry for isotopic abundances. E L Garner Ultra-high acccuracy isotopic measurements: Avogadro's constant is up! P De Bievre Atomic weights: From a constant of nature to natural variations. N E Holden Ion optical design for mass spectrometers. S W ] Clement Clean laboratories: Past, present and future. J R Moody Meteorites recovered from Australia. A W R Bevan Isotopic anomalies in extraterrestrial grains. T R Ireland Solar and solar system abundances of the elements. M Ebihara Origin of the terrestrial planets and the moon. S R Taylor Cosmogenic noble gases in silicate inclusions of iron meteorites: Effects of bulk composition on elemental production rates. O Jentsch & L Schultz Aspects of low energy nuclear fission. ] W Boldeman Isotopic studies of the Oklo fossil reactors and the feasibility of geological nuclear waste disposal. R D Loss Isotopic signatures: An important tool in today's world. D J Rokop, D W Efurd, T M Benjamin, J H Cappis, J W Chamberlin, H Poths & F Roensch Stable heavy isotopes in human health. B L Gulson Lead isotopes and pollution history. K J R Rosman & W Chisholm Surface ionization sources and applications. J E Delmore, A D Appelhans, J E Olson, T Huett, G S Groenewold, ] C Ingram & D A Dahl SHRIMP: Origins, impact and continuing evolution. W Compston Zircons: What we need to know. R T Pidgeon A review of Pb-isotope constraints on the genesis of lode-gold deposits in the Yilgarn Craton, Western Australia. N J McNaughton & D I Groves Isotopic constraints on the age and early differentiation of the Earth. M T McCulloch A tale of two era tons: Speculations on the origin of continents. A F Trendall Neutrinos: Ghosts of creation. J R de Laeter iii iv 1 5 11 21 27 29 33 43 51 59 67 73 81 85 91 97 103 109 119 123 131 141 143 Held at the Curtin University of Technology, Perth, 6-8 November 1995. Journal of the Royal Society of Western Australia OF VICTq£ CONTENTS Page Recent Advances in Science in WA 149 Installation of Professor Michael Jones as the new President of the Royal Society of Western Australia. His Excellency Major General Michael Jeffery, AC MC, Governor of Western Australia 153 The impact of vegetated buffer zones on water and nutrient flow into Lake Clifton, Western Australia. P M Davies & J A K Lane 155 A new species of Lerista (Lacertilia: Scincidae) from Western Australia: Lerista eupoda. L A Smith 161 Biogeography of the herpetofauna of the Archipelago of the Recherche, Western Australia. L A Smith & R E Johnstone 165 Population and plant growth studies of six species of Eremophila (Myoporaceae) from central Western Australia. G S Richmond & E L Ghisalberti 175 Volume 79 Part 2 June 1996 Museum of Victoria 40336 ISSN 0035-922X Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Senior Vice-President Junior Vice-President Hon Secretaries Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1996-1997 M G K Tones S Hopper H Recher A George P Gardner V Hobbs P Lavery R Froend P C Withers J E O'Shea M A Triffitt W A Cowling J Dodd D Gordon K Rosman V Semeniuk L N Thomas G G Thompson MA PhD BSc (Hons) PhD BSc PhD BA BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BAgricSci (Hons) PhD BA Msc PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc MSc PGDipEIA MEd The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March-December) at 8 pm. Kings Park Board offices. Kings Park, West Perth, W A 6005. Individual membership subscriptions for the 1996/1997 financial year are $40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countnes. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design : Mangle's kangaroo paw ( Anigozanthos manglesii) and the numbat (Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia K; CONTENTS Plant breeding for stable agriculture: Presidential Address 1994. W A Cowling Egg laying by thorny devils ( Moloch horridus) under natural conditions in the Great Victoria Desert. G Pianka, E R Pianka & G G Thompson Sea breeze activity and its effect on coastal processes near Perth, Western Australia. G Masselink A genetic perspective on the specific status of the Western Rock Lobster along the coast of Western Australia - Panulirus cygnus George 1962 or P. longipes A Milne-Edwardes 1868? A P Thompson Roost selection by the lesser long-eared bat, Nyctophilus geoffroyi, and the greater long-eared bat, N. major (Chiroptera: Vespertilionidae) in Banksw woodlands. D J Hosken Waterbirds and aquatic invertebrates of swamps on the Victoria-Bonaparte mudflat, northern Western Australia. S A Halse, R J Shiel & G B Pearson Page 183 195 199 207 211 217 Volume 79 Part 3 September 1996 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia PATRON Her Majesty the Queen VICE-PATRON His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Senior Vice-President Junior Vice-President Hon Secretaries Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1996-1997 M G K Jones S Hopper H Recher A George P Gardner V Hobbs P Lavery R Froend P C Withers J E O'Shea M A Trifitt W A Cowling J Dodd D Gordon K Rosman V Semeniuk L N Thomas G G Thompson MA PhD BSc (Hons) PhD BSc PhD BA BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BAgricSci (Hons) PhD BA Msc PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc MSc PGDipEIA MEd The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March-December) at 8 pm. Kings Park Board offices, Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1996/1997 financial year are $40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the Membership Secretary, 7c W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design: Mangle's kangaroo paw ( Anigozanthos manglesii) and the numbat (Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia CONTENTS Symposium on the Design of Reserves for Nature Conservation in South-western Australia 00^ Page Preface P Horwitz in History of conservation reserves in the south-west of Western Australia G E Rundle 225 Assessing the conservation reserve system in the Jarrah Forest Bioregion N L McKenzie, S D Hopper, G Wardell-Johnson & N Gibson 241 A floristic survey of the Tingle Mosaic, south-western Australia: applications in land use planning and management G Wardell-Johnson & M Williams 249 Terrestrial invertebrates in south-west Australian forests: the role of relict species and habitats in reserve design Barbara York Main 277 Aquatic fauna of the Warren Bioregion, south-west Western Australia: does reservation guarantee preservation? K M Trayler, J A Davis, P Horwitz & D Morgan 281 Ecosystem dynamics and management in relation to conservation in forest systems R J Hobbs 293 Forests reservations: an overview A R Main 301 Volume 79 Part 4 December 1996 Museum of Victoria 43322 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia Patron Her Majesty the Queen Vice-Patron His Excellency Major General Michael Jeffery AD MC Governor of Western Australia President Immediate Past President Senior Vice-President Junior Vice-President Hon Secretaries Hon Treasurer Hon Editor Hon Journal Manager Hon Librarian Members COUNCIL 1996-1997 M G K Jones S Hopper H Recher A George P Gardner V Hobbs P Lavery R Froend P C Withers J E O'Shea M A Triffitt W A Cowling J Dodd D Gordon K Rosman V Semeniuk L N Thomas G G Thompson MA PhD BSc (Hons) PhD BSc PhD BA BEng (Hons) GDipCSci DipEd BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BA ALIA BAgricSci (Hons) PhD BA Msc PhD BSc (Hons) PhD BSc (Hons) PhD BSc (Hons) PhD BSc MSc PGDipEIA MEd The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March-December) at 8 pm, Kings Park Board offices, Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1996/1997 financial year are S40 for Ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the Membership Secretary, % W. A. Museum, Francis Street, Perth WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The journal is indexed and abstracted internationally. Cover Design: Mangle's kangaroo paw ( Anigozanthos tnanglesii ) and the numbat ( Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences embraced by the Royal Society of Western Australia. (Artwork by Dr Jan Taylor). Journal of the Royal Society of Western Australia, 79:iii, 1996 FOREWORD The John R de Laeter Special Issue In 1995, when Professor John de Laeter retired from the post of Deputy Vice-Chancellor (Research and Development) at Curtin University, a unique opportunity arose to combine formal recognition of his broad spectrum of research achievements with an international conference on Isotope Science, the field in which the bulk of his research has taken place. Although initial plans for the conference included Science Education and Science Policy, it was eventually decided that in the two days available we could only do justice to one field. The de Laeter Symposium on Isotope Science was held on 6-8 November 1995 and attracted 72 del¬ egates from the European Community, the United States of America, Canada, Japan and Australia. Twenty one colleagues and associates journeyed to Perth to contribute papers covering the many areas of Isotope Science to which John has contributed. The papers in this special issue of the Royal Society's journal are based on these presentations. They span a very wide range of the traditional disciplines and serve to highlight both the interdisciplinary nature of isotope science and the tremendous range of John's interests. The style of the papers is equally diverse: a proportion are current reviews, w'hile others explore the origin and development of important aspects of isotope science. New data and new developments are also presented while other original, but less scientific, contributions also appear. With only minor rearrangement, this volume closely follows the four themes of the Symposium; Mea¬ surement Science; Cosmochemistry, Meteorites & Nucleosynthesis; Nuclear & Environmental Science; and Geochronology & Isotope Geology. The papers are prefaced with an inspiring opening address by Sir Marc Oliphant, who recounts some of his experiences at the Cavendish Laboratory. Sir Marc's participation in the symposium was particu¬ larly significant, since a lecture he delivered at the University of Western Australia in 1950 prompted the late Peter M Jeffery to initiate isotope ratio mass spectrometry and geochronology in Australia. Peter Jeffery's infectious enthusiasm for mass spectrometerv caught many of us, including John de Laeter. Overwhelming support for the Conference was received from all parts of Curtin University. Sponsors included The Vice Chancellor's Office, The Division of Engineering and Science, The School of Physical Sciences and The Department of Applied Physics. Kevin J R Rosman Convenor of the Organising Committee for the de Laeter Symposium on Isotope Science Curtin University of Technology Perth, WA 6102 Australia alictc the 'Pi&fawo* i 'Prow 70 Ho Sontd All tHc Atottui Sy 0T* \ S&cH 72 Sample Dissolution & Preparation r Initial Mass Spectrometry - Procedure V Separated Isotope Purification & Purity Verification Identification and Control of Error Sources Calibration Standards ■\ V Comparison of Standards and RM r Precise Mass Spectrometric - Procedure Accurate Isotope Ratio Accurate \ Atomic Weight - Certified as a Standard Reference Material Figure 1. Analytical methodology for high accuracy isotopic abundance. surements calibrated for the effects of bias in the mass spectrometry. Consequently, there was a critical need for new data and, if at all possible, calibrated or "abso¬ lute" measurements to resolve the differences in mass spectrometric data reported for the same element, as well as discrepancies between chemical and mass spec¬ trometric determinations of atomic weights. Transition to atomic weights determined by mass spectrometry Having played a significant role in the adoption of the unified atomic weight scale, and having exercised a similar role in the review process leading to the 1961 International Table of Relative Atomic Weights, NBS had great interest in redetermining the atomic weights of key elements. In particular, the focus was on those elements which could be used to assess the accuracy of many of the chemically determined atomic weights. In fact, NBS had a long and distinguished history of research on the atomic weights of the elements. Noyes (1907) reported the first of these measurements, a value of 1.00783 for the atomic weight of hydrogen based on the classical method of the electrolysis of water. Very shortly there¬ after, Noyes & Weber (1907) reported the atomic weight of chlorine. This was of particular importance since con¬ temporary work at Harvard on the ratio of silver to chlo¬ rine had led to the determination of the atomic weights of a number of elements based on the atomic weight of chlorine* The work on chlorine was followed a few years later by Weber's (1913) publication reporting the atomic weight of bromine. Between 1913 and 1959, work on the atomic weight of a variety of elements was conducted using classical chemical methods. In 1959, NBS was on the verge of starting a program which would within 5 years result in the atomic weights of silver (Shields et al. 1960), chlorine (Shields et al. 1962), and bromine (Catanzaro et al. 1964). The analytical methodology was based upon the use of isotopic stan¬ dards of known composition which were prepared by gravimetrically blending chemically pure and very nearly isotopically pure separated isotopes of the ele¬ ments. The significant event that moved NBS away from a narrow focus on the isotopic abundances of uranium and plutonium to other elements of the periodic table was a broad based program in science to improve our knowledge of the fundamental constants. High accu¬ racy determinations of the Faraday constant were of par¬ ticular interest because of its relationship to Avogadro's constant. Historically, the Faraday was determined by reacting a pure substance electrolytically with an as¬ sumed efficiency of 100%. While the electrochemical data of the time were evaluated as self-consistent, they did not agree as well as needed with the Faraday calcu¬ lated by combining other fundamental constants. In an effort to bring some clarity to this situation, Craig et al. (1960) conducted experiments leading to a determina¬ tion of the electro-chemical equivalent of high purity silver and a value for the Faraday using a silver perchlo¬ ric acid coulometer. The work of Craig and his cowork¬ ers had critical implications for Avogadro's constant, the Faraday constant and the atomic weights of a number of elements as determined chemically by their combining 6 Journal of the Royal Society of Western Australia, 79(1), March 1996 weight ratios. The initial task for mass spectrometry at NBS was to give assurance that the process used to pu¬ rify the silver for Craig's measurements had not pro¬ duced either an isotopically enriched or depleted high- purity silver sample. Because mass spectrometry gave credible evidence that some enrichment of the light iso¬ tope did occur, it was deemed necessary to do the fol¬ lowing; (1) establish the absolute isotopic abundance of silver used in the electrochemical determination of the Faraday; (2) establish limits for natural isotopic variations of silver; and (3) shed some light on the nearly 8% range in mass spectrometric data reported between 1948 and 1959. At the conclusion of these measurements, the extension of the program to the atomic weights of chlorine, bromine and other elements was an obvious out¬ come. The methodology for high accuracy By 1960, all of the essential and critical technologies for high accuracy were in place and tested. The nuclear community had served as a great incubator for analyti¬ cal chemistry, mass spectrometry and instrumentation advances. The proof of principle in blending highly en¬ riched separated isotopes to produce standards of known isotopic abundance had been conclusively and routinely demonstrated at uranium diffusion plants where mass spectrometers equipped with electron im¬ pact ion sources had made measurements with ex¬ panded uncertainties between 5 in 104 and 2 in 104. The most critical technology for the development of high ac¬ curacy uranium isotopic abundance measurements was the availability of relatively large amounts (kilograms) of highly enriched separated isotopes of uranium. The electromagnetic separation technology which produced these nuclear materials also resulted in the availability of economically affordable, but in some cases still very expensive, highly enriched and chemically pure sepa¬ rated isotopes of many other elements. These separated isotopes were the key ingredient for preparing standards of known isotopic abundance to calibrate mass spec¬ trometers for the effect of bias for a wide range of ele¬ ments other than those of critical and immediate impor¬ tance to the nuclear industry. The methodology to determine the isotopic abun¬ dance of an element with high accuracy (Fig 1) is in fact a combination of analytical chemistry and mass spec¬ trometry. The least recognized, and perhaps least ap¬ preciated component, is the analytical chemistry. The ultimate products from the analytical chemistry are; (1) a high purity Reference Material to be calibrated, and (2) a series of "calibration standards" with known isotopic ratios covering the range of isotopic ratios of the Ref¬ erence Material to be characterized. The critical components and potential sources of un¬ certainty in the analytical chemistry are sample dissolution, stoichiometry, chemical purification, interfering ions, gravimetry and high precision assay of the separated isotope solutions for the analyte element. Using gravi¬ metric procedures, weighed aliquots of the separated isotope solutions are blended to produce the calibration standards. Knowing the assay or concentration of the analyte element in the separated isotope solutions, the weight of aliquots blended to produce the standards, and the composition of the separated isotopes from mass spectrometric analysis, the isotopic abundance of the calibration standards can be calculated. Ideally, it is desirable to know the calibration standards with an ex¬ panded uncertainty of 1 in 105. More realistically, the uncertainty contribution from the calibration standards was 1 to 2 in 104. Mass spectrometry instrumentation and procedures For most of the high-accuracy work at NBS, solid sample, thermal or surface ionization mass spectrom¬ eters were used. A notable exception was by Barnes et al (1975) for the isotopic abundance of a reference sample of high purity silicon by electron impact ioniza¬ tion of a gaseous sample. The other notable exception was the use by Gramlich et ah (1973) of a double mag¬ netic stage instrument arranged in an "S" configuration and equipped for pulse counting detection. For the atomic weight determination of nickel by Gramlich et al (1989), a commercially designed multicollector instru¬ ment was used. Otherwise, all mass spectrometers were single focussing, magnetic sector instruments, either 15 cm or 30 cm radius of curvature, and/or 60°, 68°, or 90° sectors. These instruments (Shields et al 1967) were equipped with interchangeable and nearly identical elec¬ tronic components, the same basic design of a multi¬ element, deep-bucket Faraday cup collector, and either a standard Nier type ion source (prior to 1964) or a thin lens "z" focussing ion source (after 1964). Mass spectrometric procedures for all elements were based upon: (1) a knowledge of the sources of uncertainty in the measurement process; (2) control of the parameters and influence factors which could contribute to the uncertainty or bias the measurement; (3) identifying and selecting the critical analytical parameters; and (4) establishing an acceptable tolerance specification for each parameter which would be the same for all analyses. Because of isotopic fractionation effects in the ion source of a thermal ionization mass spectrometer, it was critical to develop a methodology which yielded pre¬ cisely the same isotopic fractionation pattern for each analysis. The goal was to quantify the correction for this effect as an experimentally determined constant which was unique for the analytical conditions of the measure¬ ment. In general, three types of isotopic fractionation trends or patterns were observed for the ratio of the lighter to heavier isotope of an element: (1) constant and non-changing with time; (2) decreasing with time; and (3) increasing with time. 7 Journal of the Royal Society of Western Australia, 79(1), March 1996 For some elements and matrices, all three patterns could be observed in the same analysis. By controlling the time base for each analysis and the other critical parameters of the measurement process, it could be demonstrated that the same correction for isotopic frac¬ tionation could be applied to each analysis. In addition to isotopic fractionation, it was necessary and essential to characterize the effect of secondary electrons in the collector as well as non-linear effects in the entire mea¬ suring system. Complete characterization of the mea¬ surement uncertainty from the mass spectrometry in¬ cluded an evaluation of possible sources of uncertainty, influence factors and suspected influence factors in the sample handling system, the ion source, analyzer, ion detector, and the data handling system. A graphic rep¬ resentation of the general scheme used to completely characterize the uncertainty is shown in (Fig 2), includ¬ ing an allowance for the use of judgment-based limits. Knowledge of influence factors and suspected influence factors was integrated into the measurement process as a modification or refinement of control procedures or specifications for control of a parameter. Figure 2. Characterization of measurement uncertainty. Results Atomic weight determinations, certification of the iso¬ topic abundances of uranium and plutonium, and trace element analysis by isotope dilution mass spectrometry were the broad areas where high accuracy isotopic abun¬ dance measurements were made at NBS from 1960 to 1985. The scope of this paper does not permit full attention to either of these areas, so atomic weights will be used as the prime example. During the period in question, the atomic weights of 14 different elements were determined. These measurements included; silver (Shields et al. 1960); chlorine (Shields et al. 1962); copper (Shields et al. 1964); bromine (Catanzaro et al. 1964); chromium (Shields et al. 1966); magnesium (Catanzaro et al. 1966); lead (Catanzaro et al. 1968); rubidium (Catanzaro et al. 1969); boron (Catanzaro et al. 1970); rhenium (Gramlich el al. 1973); silicon (Barnes et al. 1975); potas¬ sium (Garner et al. 1975); Thallium (Dunstan et al 1980); silver (Powell et al. 1981) and strontium (Moore et. al. 1982). After the period in question, the absolute isotopic abundance ratios and the atomic weight of gallium (Machlan et al. 1986) and nickel (Gramlich et al. 1989) were reported. The absolute isotopic abundance ratios for a representative number of these elements are shown in Table 1. Table 1. Absolute isotopic abundance ratios by thermal ionization mass spectrometry. Element Year Ratio Value Ag 1960 107Ag/109Ag 1.0755 ± 0.0013 Ag 1961 107Ag/109Ag 1.07597 ± 0.00135 Br 1964 79Br/81Br 1.02784 ± 0.00190 Cu 1964 63Cu/65Cu 2.2440 ± 0.0021 Rb 1969 85Rb/87Rb 2.59265 ± 0.00170 Re 1973 185Re/187Re 0.59738 ± 0.00039 T1 1979 205T1/203T1 2.38714 ± 0.00101 Ag 1981 107Ag/109Ag 1.07638 ± 0.00022 The uncertainty of an isotopic ratio was reported as an overall limit of error and was determined by sum¬ ming the 95% confidence limits of the uncertainty com¬ ponents. The overall or expanded uncertainty of the measurement included terms to cover random effects as well as possible systematic error. Each overall or ex¬ panded uncertainty included a random component for the "mass spectrometric analytical error," a component for "possible systematic error in composition of sepa¬ rated isotopes," and a component for "possible system¬ atic error in chemical analysis." The only notable excep¬ tion was the atomic weight of silver by Shields et al. (1960) which included a component for nuclidic masses, since at that time it was not clear as to whether the unified atomic weight scale would be adopted. Because rounding-off could have impact on the atomic weight calculation, most ratios were reported to one figure be¬ yond significance. While there is some uniqueness about each of the measurements in the atomic weight series, silver is con¬ sidered in a separate niche because; (1) it had the distinction of being the "pathfinder element" (Shields et al. 1960) of the series; (2) the only element for which additional measure¬ ments were made and data published beyond the original measurements (Shields et al. 1961); and (3) the only element for which a completely indepen¬ dent redetermination was made by a different team of scientists (Powell et al. 1981). In addition, silver was the only element for which the expanded uncertainty was perceived to be very near the limits achievable with the instrumentation, analytical chemistry and measurement process then developed. 8 Journal of the Royal Society of Western Australia, 79(1), March 1996 Discussion The single unifying concept which brought together the critical elements of analytical chemistry and mass spectrometry to yield high accuracy isotopic abundance measurements was the concept of measurements as a process. Pontius & Cameron (1967) first documented this concept in describing mass measurements as a pro¬ duction process. A fundamental assumption of this con¬ cept is that measurement is analogous to a production or manufacturing process. When a production process is in a state of control, the product is uniform and re¬ producible, reflecting the degree of control of the pro¬ cess. Analogous to the industrial production process, the product of an isotopic abundance measurement pro¬ cess is an isotopic ratio. When the entire process is in a state of statistical control, the associated uncertainty of the process is valid and predictable. When the measure¬ ment process is out of statistical control, predictability is lost and the uncertainty of the measurement under con¬ trol conditions is not applicable to the out-of-control condition. In conjunction with measurement as a process, the concept of measurement assurance was the other gen¬ eral principle which underpinned the pathway to high accuracy isotopic abundance measurements. This pro¬ cess was perceived and implemented as a continuous improvement process. The technique for mass spec¬ trometry included but was not limited to; (1) the use of sound experimental design principles so that the entire measurement process, its components and relevant influencing factors could be well char¬ acterized, monitored and controlled; (2) complete experimental characterization of the un¬ certainty for the measurement process to include statistical variations, contributions from all known or suspected influence factors, imported uncertain¬ ties, judgment based components, and the propaga¬ tion of uncertainties throughout the measurement process; and (3) continuously monitoring the performance and state of control of the complete measurement process, both analytical chemistry and mass spectrometry, with locally tested and proven quality control or process control techniques to include the measure¬ ment of stable isotopic abundance check standards along with the normal workload. The state of statistical control observed for an element did not preclude the measurement assurance require¬ ment to continuously monitor, control, and evaluate the process. After completion of a unique or distinct set of measurements such as an atomic weight determination, then it was appropriate to study, evaluate, test and ex¬ periment with the measurement process. The results and findings would then become the basis for changing, improving, redesigning or even developing a substan¬ tially different process. After the next set of unique or distinct measurements the cycle was repeated. Where applicable, knowledge gained about one element was utilized in developing and evaluating the measurement process for other elements. Perhaps the best illustration of the value of a sound measurement assurance program was the first attempt to determine the atomic weight of bromine in 1962. An independent, third-party statistical analysis of the data, a practice which was followed in all of the high accu¬ racy isotopic measurements, revealed an out-of-control statistical condition during the preparation of the cali¬ bration mixes. Re-evaluation of the analytical chemistry revealed that the separated isotopes were not blended over a short time span of hours but a period of several days. As a result the isotopic abundances of the calibra¬ tion standards were dependent on the time of mixing and it was necessary to void the entire set of isotope ratio measurements. This hard-earned lesson became a cornerstone of the methodology for all future measure¬ ments in which solutions were blended to prepare cali¬ bration mixes or to perform isotope dilution. Summary The products of isotopic abundance mass spectrom¬ etry are the value for the abundance ratio(s) and the uncertainty associated with the ratio. Initial isotopic ra¬ tio measurements by thermal ionization mass spectrom¬ etry at NBS between 1958 and 1960 were made with an expanded uncertainty of 2% over the ratio range 1:20 to 20:1. With major developments and improvements in instrumentation and measurement process, the ex¬ panded uncertainty was reduced to 0.5% for the ratio range 1:100 to 100:1. The breakthrough to an expanded uncertainty of 1 or 2 in 104 as a realistic limit for high accuracy isotopic abundance measurements by thermal ionization mass spectrometry occurred when it was re¬ alized that the best strip chart recorders available were a limiting factor. When performing to manufacturer's specification, strip chart recorders measured a full scale output signal to ±0.5 %. By modifying the strip chart recorder to operate in an expanded scale measuring mode, the contribution to the uncertainty of the mea¬ surement was reduced by a factor of 10. With improve¬ ments in the measurement process and the advent of digital measurement systems, the ultimate accuracy for thermal ionization mass spectrometry was realized for the determination of the atomic weight of silver (Powell et al. 1981). In looking at the future from a 1995 perspective, the potential for ultra high accuracy measurements of 1 in 105 is still over the horizon. Assuming ideal conditions of equal atom ratio measurements, ultra high enriched separated isotopes of 99.9999%, highly reproducible iso¬ topic fractionation, and ideal ion current intensities, ul¬ tra high accuracy is not a practical goal without major breakthroughs or different approaches in the analytical chemistry and mass spectrometry. References Barnes I L, Moore L J, Machlam L A, Murphy T J & Shields W R 1975 Absolute isotopic abundance ratios and atomic weight of a reference sample of silicon. Journal of Research of the National Bureau of Standards (US) 79A:727-735. Catanzaro E J, Murphy T J, Gamer E L & Shields W R 1964 Abso¬ lute isotopic abundance ratio and the atomic weight of bro¬ mine. Journal of Research of the National Bureau of Stan¬ dards (US) 68A:593-599. Catanzaro E J, Murphy T J, Garner E L & Shields W R 1966 Abso¬ lute isotopic abundance ratios and atomic weight of a reference 9 Journal of the Royal Society of Western Australia, 79(1), March 1996 sample of magnesium. Journal of Research of the National Bureau of Standards 70A:453-458. Catanzaro E J, Murphy T J, Shields W R & Garner E L 1968 Abso¬ lute isotopic abundance ratios of common, equal atom and radiogenic lead isotope standards. Journal of Research of the National Bureau of Standards (US) 72A:261-267. Catanzaro E J, Murphy T J, Gamer E L & Shields W R 1969 Abso¬ lute isotopic abundance ratio and atomic weight of terrestrial rubidium. Journal of Research of the National Bureau of Stan¬ dards (US) 73A:51 1-516. Catanzaro E J, Champion C E, Garner E L, Marienenko G, Sappenfield K M & Shields W R 1970 Standard reference ma¬ terials: Boric acid; isotopic and assay standard reference ma¬ terials. National Bureau of Standards (US) Special Publica¬ tion: 260-17. Craig D N, Hoffman J 1, Law C A & Hamer W J 1960 Determina¬ tion of a value of the Faraday with a silver perchloric acid coulometer. Journal of Research of the National Bureau of Standards (US) 64A:381-402. Dunstan L P Gramlich J W Barnes I L Purdy W C 1980 Absolute Isotopic Abundance and the Atomic Weight of a Reference Sample of the Thallium. Journal of Research of the National Bureau of Standards (US) 85:1-10. Garner E L, Murphy T J, Gramlich J W, Paulsen P J & Barnes I L 1975 Absolute isotopic abundance ratios and the atomic weight of a reference sample of potassium. Journal of Re¬ search of the National Bureau of Standards (US) 79A:713-725. Gramlich J W, Murphy T J, Garner E L & Shields W R 1973 Abso¬ lute isotopic abundance ratio and atomic weight of a refer¬ ence sample of rhenium. Journal of Research of the National Bureau of Standars (US) 77A:691-698. Gramlich J W, Machlan L A, Bames I L & Paulsen P J 1989 Abso¬ lute isotopic abundance and the atomic weight of a reference sample of nickel. Journal of Research of the National Bureau of Standards (US) 94:357-362. Machlan L A, Gramlich J W, Powell L J & Lambert G M 1986 Absolute isotopic abundance ratios and atomic weight of a reference sample of gallium. Journal of Research of the Na¬ tional Bureau of Standards (US) 91:323-331. Moore L J, Murphy T J & Bames I L 1982 Absolute isotopic abun¬ dance ratios and atomic weight of a reference sample of stron¬ tium. Journal of Research of the National Bureau of Stan¬ dards (US) 87:1-8. Noyes W A 1907 The atomic weight of hydrogen. Bulletin of the National Bureau of Standards 4:179-204. Noyes W A & Weber H C P 1907 The atomic weight of chlorine. Bulletin of the National Bureau of Standards 4:345-364. Pontius P E & Cameron J M 1967 Realistic uncertainties and the mass measurement process. National Bureau of Standards, Monogram 103. Powell L J, Murphy T J & Gramlich J W 1982 Absolute isotopic abundance and atomic weight of reference sample of silver. Journal of Research of the National Bureau of Standards (US) 87:9-19. Shields W R 1967 Analytical mass spectrometry section: Sum¬ mary' of activities July 1966 to June 1967. National Bureau of Standards (US) Technical Note 426. Shields WT R, Craig D N & Dibeler V H 1960 The absolute isoto¬ pic abundance ratio and atomic weight of silver. Journal of American Chemical Society 82:5033-5036. Shields W R, Garner E L & Dibeler V H 1961 Absolute isotopic abundance of terretrial silver. Journal of Research of the Na¬ tional Bureau of Standards 66A:l-3. Shields W R, Murphy T J & Gamer E L Dibeler V H 1962 Abso¬ lute isotopic abundance ratio and the atomic weight of chlo¬ rine. Journal of the American Chemical Society 84:1519-1522. Shields W R, Murphy T J & Gamer E L 1964 Absolute isotopic abundance ratio and the atomic weight of copper. Journal of Research of the National Bureau of Standards (US) 68A:589-592. Shields W R, Murphy T J, Catanzaro E J & Gamer E L 1966 Abso¬ lute isotopic abundance ratio and the atomic weight of a ref¬ erence sample of chromium. Journal of Research of the Na¬ tional Bureau of Standards (US) 70 A. T 93-197. Weber H C P 1913 The atomic weight of bromine. Bulletin of the National Bureau of Standards 9:131-150. 10 Journal of the Royal Society of Western Australia, 79:11-19, 1996 Ultra-high accuracy isotopic measurements: Avogadro's constant is up! P De Bievre Institute for Reference Materials arid Measurements, European Commission-JRC, B-2440 GEEL, and Department of Chemistry, University of Antwerpen, Belgium Abstract Ultra-high accuracy isotopic measurements have been achieved on abundance ratios of Si isotopes by achieving 10~5 reproducibilities on the ratio measurements and calibrating the results by synthetic isotope mixtures prepared to 2 10'5 combined relative uncertainty. This enabled us to attain a relative combined uncertainty of 3 10 5 on the abundance ratios in natural Si samples and, consequently, 10'7 on the Si molar mass. The route to these results is described, followed by a description of the improvement of our knowledge of the Avogadro Constant N A through these measurements. Combined with measurements of the lattice constant and density in a near-perfect Si single crystal. Si molar mass measurements lead to a totally independent value of NA. This value is compared to the authoritative CODATA evaluation of the interrelationships of our fundamental constants. After the year 2000 an improved value of NA will almost certainly play a key role in a redefinition of the kilogram, our primary standard of mass. The acquired expertise in measurement instrumentation and measurement procedures for mea¬ surements of Si isotope amount ratios can now be extended to a more general use in measure¬ ments of isotope amount ratios (i.e. in other elements). It can also be combined with isotope dilution. I describe how the latter combination may open the possibility of realising direct traceability of an amount-of-substance measurement to the measurement procedure and instru¬ mentation leading to the Avogadro Constant. Perhaps a traceability to the mole , i.c. to SI, is being developed. Introduction Some elements have their abundance ratios known to 10'3 combined relative uncertainty, some abundance ra¬ tios have been measured to 1CH, but only for one ele¬ ment, silicon, have the abundance ratios been measured to a confirmed relative uncertainty in the 10"5 range, all uncertainty components included. It required 1CD re¬ producibility in abundance ratio measurements, cali¬ brated by 2 10'5 accurate synthetic isotope mixtures, to achieve a 3 105 combined relative uncertainty. Since it takes an enormous effort to achieve this, the incentive to do it must be important. Indeed it is. It is the molar mass M (numerically equal to the atomic weight) which constituted for many years the limiting factor in the knowledge of the Avogadro con¬ stant Nx (in mol*1) as determined from X-ray density, lattice constant and molar mass measurements on a near-perfect Si single crystal (an " Avogadro crystal ) through the relation Na =M (Si) / p (eq. 1) V a in which NA is calculated as the ratio of molar volume M (Si) / p (in m3 mol'1) to atomic volume V (in m3); M © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 (Si) is the molar mass in kg mol'1 and p the density in kg nr3. The value of Va is determined through an X-ray inter¬ ferometric measurement of the lattice constant d22Q. With the lattice parameter ao = d22Q vU the atomic volume V =a \ Molar masses are derived from isotope abun¬ dances f (lft = 1) and atomic masses M('Si). In fact, abundance ratios R. are measured relative to the abun¬ dance f of one isotope j as. IfMCSi) M(Si) = If M ('Si) = - rrr - A/, l\fM('Si)]/f lRtM('Si) (lf)/f ' XR, (eq. 2) The isotope abundances (Table 1) and resulting molar mass (Table 2) of natural Si are not constant enough in nature to be used in this approach. Individual calibrated measurements of the Si in the "Avogadro crystal" must be carried out. Combined relative uncertainties which have been achieved on M, Va and p by 1994 in the cur¬ rently running three Avogadro projects in the world, are summarized in Table 3. A combined relative uncertainty of 10'7 on M(Si) requires a combined relative uncertainty of 3 10'5 on the abundance ratios of the Si isotopes in a Si single crystal. Before 1990 this uncertainty was simply the largest uncertainty contributor to the uncertainty of Na (Table 4). 11 Journal of the Royal Society of Western Australia, 79(1), March 1996 Table 1 International Union of Pure and Applied Chemistry (IUPAC) iso¬ tope abundances of Si. Isotopes IUPAC IUPAC IUPAC IRMM 1993 representative selected evaluated = IUPAC isotopic range of selected composition of "Best natural "Best terrestrial Measure- occurrences Measure- material ment" ment" in 1995 /(“Si) 0.9223 0.9223104 0.9241-0.9214 0.9223104 1 46 46 /PSi) 0.0467 0.0467536 0.0473-0.0457 0.0467536 1 33 33 /P°Si) 0.0310 0.0309360 0.0314-0.0301 0.0309360 1 36 36 Table 2 International Union of Pure and Applied Chemistry (IUPAC) molar masses of Si. Molar mass (atomic weight) MfSiVg-mol1 IUPAC representative 28.0855 atomic weight of 3 terrestrial material IRMM 1993 28.0854349 76 IUPAC selected "Best Measurement" up to 1995 28.0855 2 after 1995 28.0854349 76 Table 3 Achieved combined relative uncertainties on measurements of lattice parameter ao, density p and molar mass M, as measured on near-perfect Si single crystals. IRMM = Institute for Refer¬ ence Materials and Measurements, Geel; PTB = Physikalisch Technische Bundesanstalt, Braunschweig; NRLM = National Research Laboratory for Metrology, Tsukuba; IMGC = Istituto di Metrologia Gustavo Colonnetti, Torino. Na Project 1 2 3 a 0 V a PTB 1 6 Kb8 1.8 l(b7 IMGC2 3 10* 9 10-8 P PTB1 7.7 Kb7 IMGC3 1.5 lO'7 NRLM 4 1.1 l(b7 Af(Si) IRMM1 3.2 10-7 IRMM1 3.2 lO’7 IRMM1 3.2 IO*7 *De Bievre et al. (1995); 2Basile et al (1995);3Saccomi et al. (1995); 4Fujii et al. (1995) Table 4 The isotopic composition of natural silicon with combined relative uncertainties compared with uncertainties of measurements at IRMM on "Avogadro Si crystal" samples. IUPAC values for natural abundances IRMM 1989/90 IRMM 1992“ IRMM 1995b /(28Si) 0.9223 1 ±1 10-* ±1 io-5 ±1 io-6 ±1 io-6 /(29Si) 0.0467 1 ±2 10*3 ±2 10r* ±2 10-5 ±2 IO'5 /r si) 0.0310 1 ±3 Kb3 ±3 10-1 ±3 10-5 ±3 IO*5 M(Si) 28.0855 ±1 10-5 ±1 io-6 ±1 IO7 ±1 10-7 n{29S\)/n{2fiSi) 0.050 6 ±2 10 -3 ±2 Kb4 ±2 l(b5 ±2 IO'5 Si) 0.033 6 ±3 10 -3 ±3 10-4 ±3 IO'5 ±3 IO*5 aSeyfried et al. (1992); bDe Bievre et al. (1995) Measurement of M(Si) and its relative uncertainty The Si crystal samples were converted to SiF4 gas as described by De Bievre et al. (1995). The Avogadro I Mass Spectrometer (an IRMM upgraded MAT-CH5) had yielded a 1 Kb6 M(Si) uncertainty which had led to 1.1 Kb6 relative uncertainty on the 1992 NA value (Seyfried et al. 1992). Using the acquired experience, an Avogadro II Mass Spectrometer was assembled, essentially based on a standard MAT 271 instrument, into what is now known as the IRMM /MAT 271 Avogadro II mass spec¬ trometer. The development of appropriate measurement procedures in which correction for every single signifi¬ cant error as well as a full orthodox calculation of every 12 Journal of the Royal Society of Western Australia, 79(1), March 1996 uncertainty contribution was properly incorporated, proved to be of overriding importance. Creating near- to-ideal vacuum conditions, realizing close-to-ideal cir¬ cumstances for the SiF4 gas, also proved to be crucial for the ultimate quality of the results and has been described elsewhere (De Bievre et al. 1994). Then, of course, syn¬ thetic isotope mixtures had to be made to calibrate the isotopic measurements on the Si of the "Avogadro" crystal. They were made from enriched isotopes (see Table 5) which were incorporated in higly stable, non-hygro- scopic, pure Cs?SiFh molecules. This enabled the prepa¬ ration of isotope mixtures of about "natural" isotopic composition (Fig 1). Typical values of such a mixture with their uncertainties are given in Table 6, which also shows a full analytical uncertainty budget that will ulti¬ mately contribute to the uncertainty of the Avogadro constant. The complete correct formula to calculate the value of such a mixture is (De Bievre et al. 1995); Table 5 Isotopic composition of the enriched starting materials "'Si" (af¬ ter chemical purification and assay) as measured on their SiF3+ ions. Combined standard uncertainties (uc) are given below the digits to which they apply. Enriched //28Si" Enriched "29Si" Enriched "30Si" /(28Si) 0.9953463 10 0.0343310 200 0.0364759 44 fi29 Si) 0.0028138 5 0.9639600 210 0.0028005 43 /(30 Si) 0.0018399 8 0.0018090 11 0.9607236 63 n(29Si) n(28Si) m('Cs28SiF")f( 29.,,28”) m(“Cs29SiF") •/( 29."29") mC'Cs^SiF ")-f( 29."30") MCCs^SiFJ MCCs^SiFJ M("Cs230SiF6") mC'Cs^SiFJfi 28."28”) M("Cs28SiF ") m(^'Cs29SiF ")f (28. "29") M("Cs29SiF") mC'Cs^SiFJfi 28."30") M("Cs230SiF6") (eq. 3) Table 6 Calculated values with combined relative uncertainties for one of the Si synthetic isotope mixtures as propagated from the contributing uncertainty sources. Also listed are the chemical uncertainties for Ar(Si):18. abundance 107combined relative standard uncertainty (amount of substance as contributed to the total by the measurements of the: fraction) in the synthetic mixture total abundances of enriched isotope materials3 chemical impurities stoichiometry weighings atomic masses of the isotopes f(28Si) 0.9233923 18 18 12 6.5 6.5 8.4 0.2 fCSi) 0.0463008 14 14 10 4.6 4.6 6.1 0.14 A30 Si) 0.0303069 11 11 7.9 3.0 3.0 6.1 0.09 Ar(Si)b = 28.0837260 27 1.0000000 lc 25 18 9 9 12 <1 afrom measurements of the enriched isotopes; b this "sample atomic weight" applies only to this synthetic mixture; crounding off uncer¬ tainty only. Table 7 Isotopic composition and silicon atomic weight of the "test material" which is now a certified isotopic reference material (IRMM 018). Combined standard uncertainties (u) are given below the digits to which they apply. Abundance ratio Abundance (amount of substance fraction) Atomic weight (relative atomic mass) w(29Si)/n(28Si) ni^SD/n^Si) /(28Si) /(29Si) /(* Si) A(Si) 0.0508442 0.0335851 0.9221440 0.0468857 0.0309703 28.085635 24 33 35 21 29 6 13 YEAR Journal of the Royal Society of Western Australia, 79(1), March 1996 GIVEN MATERIALS INTERMEDIATE PRODUCTS TARGET IONS FOR MS MEASUREMENT Figure 1. General layout of preparation and measurements of Si isotope mixtures. 6.022 080 6.022 090 6.022 1 00 6.022 1 1 0 6.022 1 20 6.022 1 30 6.022 1 40 /VA/1023-mol'1 Figure 2. Values and uncertainties (as stated by authors) for N A in the period 1974-1986. Values are from Deslattes et al. (1974 1976) Cohen & Taylor (1987), Deslattes (1988), Seyfried et al. (1992), Basile et al. (1995), De Bievre et al. (1995). 14 Journal of the Royal Society of Western Australia, 79(1), March 1996 where " indicates the isotopically enriched materials, not pure isotopes. In the course of the project, a 10 kg Si02 batch was measured and calibrated and made into a certified reference material, IRMM-018 (De Bievre et al. 1994), the values of which are given in Table 7. What now is Avogadro's Constant? The various determinations of N . over the last 20 A years are given in Table 8 and Figure 2. They show the gradual 'Tuning in" of the value as well as the reduction of the uncertainty. As is apparent from the CODATA evaluation of the interrelationships of the fundamental constants (Cohen & Taylor 1987; Taylor & Cohen 1990), the Avogadro constant obtained through the molar mass/ lattice constant/ density route is consistent at the lO^-N^ level with other fundamental constants such as the Planck constant, the Boltzmann constant, the Uni¬ versal gas constant etc. It is interesting to note (De Bievre & Peiser 1994) how this uncertainty has been de¬ creasing at a steady rate of a factor of 10 every 15 years since about 1860 (Fig 3). Table 8 Values of NA with their combined relative uncertainties as detemined in the period 1974-1994, compared to the 1986 CODATA evaluation. Year with uncertainty 1974d 6.0220943 1023 63 1976b 6.0220941 1023 53 1988c 6.0221318 1 023 73 1992d 6.0221363 1 023 66 1994e 6.0221365 1023 51 1994' 6.0221379 1023 25 1994' 6.0221430 1023 8 CODATA* 6.0221367 1 023 1986 36 ‘Deslattes et al. (1974), bDeslattes et al. (1976), cDeslattes (1988), dSeyfried et al. (1992), t-De Bievre et al. (1994), fBasile et al. (1994), *Cohen & Taylor (1987). b < A io+2 io+1- 10 - 10' 10" 10 - 10 10 ' 1 - T Jl A aA || Tu T < < < (k - 1 |1 i a i A A AA I A = /Va 6.022 136 7 • 1023mol‘1 A A A *iV A Positive A values shown by tip of arrow pointing down @A Negative A values shown by tip of arrow with bar pointing down A |r* U, the absolute (single-deviation) uncertainty, shown by A pointing up, is estimated for the time of initial publication 1986 CODATA T The ringed numbers, to clarify overlapping entries, REFERENCE VALUE i refer to the order number in Figure 1b . . . . . . . 1 f 1 M 1 < 1 M ' i 1 1 M ' 1 '' '' 1 ' ' ' ' 1 ' ' ' 1 1 11 1 1 " 1 1 1 " " 1 " "'1 ''T- 1 1 " " 1 1 1 rrl " " I ' 1 rr r'"r"T"Tr,m ",,l,"TMT'rrr o o o CD o o o in o> - 10+1 -10” -10" -10" 10" -10“ -10" 10 10" o o YEAR Figure 3. Uncertainties of published values for Avogadro's constant. 15 Journal of the Royal Society of Western Australia, 79(1), March 1996 Possible consequences of the achieved uncertainty on NA It is now well accepted (Kind & Quinn 1995) that there are variations of the order of 5 10-8 kg in the mass of the prototype of the kilogram, our primary standard of mass (see Fig 4). Hence, a new definition of the kilogram is needed. It will probably evolve around "V12 of the mass of (NJ atoms of 12C (x 1000)". For a smooth transition of definitions, an uncertainty of at least 5 10'8 Na must be attained on NA. Work is in progress to achieve this. Inversely, the observation that NA is consistent with other fundamental constants to better than 10* NA, one could say that the closely-knit network of fundamental constants supports a value of NA to at least 10* NA com¬ bined relative uncertainty and therefore also supports the claimed combined relative uncertainty of the molar mass, lattice constant and density measurements as a group - not necessarily singly. Hence, this network also supports the measurements of the isotope amount ratios R. of Si at the 3 10'5 R level. Apart from the fact that the molar mass measurements are traceable to prepared iso¬ tope amount ratios (synthetic isotope mixtures), the en¬ tire measurement procedure to achieve 3 10*5 uncertainty on an isotope amount ratio can be said to be, so to say, under the supreme quality control of the Avogadro constant and its interrelationships with other fundamental con¬ stants. There is indeed a strong "quality assurance"; the measurement procedure and instrumentation appear to be monitored by the "network" of inter-related funda¬ mental constants. Are the measurement methods leading to the Avogadro constant useable for other metro¬ logical purposes? Given the supreme "quality assurance" exerted by the network of fundamental constants on the measurement methods used, the results obtained with these methods are indeed confirmed by this network within the stated uncertainties. Are results of measurements, by these methods, not reliable and credible in other fields of application to within the stated uncertainties? Does the X-ray interferometric method used not yield a result of the same uncertainty in other lattice constant measure¬ ments? Are density measurements using the same method on other materials not credible to the stated un¬ certainty? Are molar mass determinations not reliable to the stated uncertainty when performed on other (ideal) gases? The underlying reasoning for these ques¬ tions is that measurement methods which are monitored by a quantified connection to the network of interrelated 1890 1910 1930 1950 1970 1990 YEAR Figure 4. Observed variations in mass of realisations of the kilogram (indicated by their number identification) against the mass of the prototype kg of the kilogram. 16 Journal of the Royal Society of Western Australia, 79(1), March 1996 fundamental constants are subject to the highest "qual¬ ity assurance" programme imaginable, one which does not make use of a human convention (written procedure or other) as the ultimate judge but of nature itself by referring to its unalterable fundamental constants. What is more, all quantities measured in equation 1 are properly traceable to SI units, and, in addition, equa¬ tion 1 only contains ratios of values expressed in SI units, kg molA /kg m'3 = - - ~~ = mol- {eq. 4) This ultimately delivers a number per mole, only based on ratios of values which are expressed in SI units. Isotope mass spectrometrists have always made cali¬ brated measurements of a molar mass (numerically equal to atomic weight or mean relative atomic mass) which were in fact traceable to ratios of amounts of pure isotopes in gravimetrical preparations from enriched iso¬ tope materials. Through various corrections e, such as for the lack e of 100% isotopic purity of the enriched isotopes, for impurities ejm and for deviations of stoi¬ chiometries e t of the compounds containing the en¬ riched isotopes, the weight or mass ratio of isotopically enriched materials (’E) leading to synthetic isotope mix¬ tures, is converted into an amount ratio of (pure) iso¬ topes; n('E) m ('£) M (>E) 1-eJj'E) l-e„('E) 1-e nQE) - m (E) ' M ('£) ' 1-eb('E) ’ 1-sJE) ' 1 -eimp('E) (eq ’ Also, this equation only contains ratios (the corre¬ sponding uncertainty budget is given in Table 6). The isotope amount ratio then serves to calibrate the mea¬ sured abundance ratios. The Na measurement as a "primary method" for amount ratios As described above, a completely understood and fully described measurement procedure does result from the isotopic measurements of Si for the determination of Na. The equations relating "measurement signals" to what is intended to be measured are all in expressed (in ratios of quantities expressed) in SI units and do not contain any empirical term. We think therefore that they qualify as a "primary method of measurement". If so, other amount ratio measurements, using this procedure can also be considered as having been measured by a "primary measurement method" which is subject to the supreme "quality assurance" exerted by the fundamen¬ tal constants network. Ensuing results thus borrow their credibility and stated uncertainty from that network of fundamental constants. Consequences for isotope dilution as measurement method An isotope dilution measurement procedure com¬ pares an unknown amount of a representative isotope of an element in a sample X to a known amount ("spike") of a representative but other isotope of the same element in a sample Y (De Bievre 1990,1993) through the measurement of an amount ratio R in their isotopic blend B, R = ” (,£) x n (>E) Y (eq. 6) All other isotopes in unknown sample and spike are only correction terms to the above equation (De Bievre 1990,1993). The full general equation is. n x = Ry ~ Rb ^ R,x (eq. 7) n (E) y Rb - Rx SR, y The connection of the measurement of a (very) small amount or a (very) small concentration can therefore be visibly and traceably made to a pure substance (the iso¬ topically enriched spike material). This in itself can be expressed in mol kg'1 through corrections of mass due to impurities, deviation of 100% isotopic purity, a known (deviation of) stoichiometry and known atomic weights. Alternatively, the spike can be measured by comparing it to a pure substance of natural isotopic composition (reverse isotope dilution). Thus traceability to the value in mol kg'1 in a pure substance can be established using a measurement method which in itself is controlled by the Avogadro constant and the network of fundamental constants. When the kilogram itself has been redefined as the mass of so many atoms as there are contained in 0.012 kg of 12C, exclusive traceability to the mole will have been established since the kg will have disappeared from the traceability chain. Other evidence for a fundamental role of isotope amount ratio measurements When measuring the abundance ratios of the Si isotopes on SiF4 gas, it is obvious that the inlet of the gas into the ion source of the mass spectrometer must be done through pinholes in a (gold) leaf (Fig 5). This process continuously changes the isotopic composition of the gas (actually enriching the sample in the expansion vessel prior to the inlet, in the heavier isotopic molecules). The size of this effect is derived directly from kinetic gas theory; it is the very well known square-root-of-mass- ratio. Performing measurements over many hours, of exponentially changing values of the isotope amount ra¬ tio over time or, better, over the continuously measured ion current of one isotope, enables back-extrapolation of the isotope amount ratio to find its value at time zero, the opening of the valve of the sample container. The gas flow rate into the ion source is, -dNt(t) /dt = R ■ « M- * -N, (f) (eq. 8) where N.( t) is the number of molecules in the container at t- 0, and ft is a constant for a given effusion barrier. Integrating from t to to yields; N (0 = N, (to) • exp (-6 ■ 'M-4’ -t) (eq. 9a) and N (t) = N. (to) ■ exp (-ft • 'M~4, -t) (eq. 9b) Thus, for isotopic gas molecules, one establishes that, IJt) In f— = -ft • ('M_4> - >M~*) -t (eq. 10) 17 Journal of the Royal Society of Western Australia, 79(1), March 1996 shut-off valve X sample inlet from ampoule (viscous flow) expansion vessel shut-off valve gold foil leak y< to ion source t gaseous diffusion (molecular flow) Figure 5. Inlet path of gas into ion source of gas isotope mass spectrometer. Table 9 Theoretical and experimental values of M(14N2)/M(14N15N); uncertainties (Is) are given below the digits to which they apply. predicted by kinetic gas theory experimentally observed values difference relative difference [(28SiF4)/ (28SiF4)]05 1.00000 1.00026 0.00016 [(29SiF4)/ (28SiF4)]° 5 1.00480 1.00469 20 0.00011 20 -1.1 io-1 [(30SiFJ)/ (28SiF4)]05 1.00957 1.00911 31 0.00046 31 -4.6 10-4 and also that lft) In -fjj y = -fi • CM-* -0 (eq. 11) Consequently, r 'M-* ln y o = \_iM-* ~ 1 P'O' and values for (iM/iM) ‘t> can be estimated from the slope of ln/.^t) against lnJ.(t). From a best fit of these experimental values for the changing ratio, experimental values for 4> can be ob¬ tained. Optimizing instrument design, vacuum condi¬ tions and measurement procedures, the experimental values from Table 9 have been observed. The gratifying agreement with gas kinetic theory values, gives direct support that the measurement process is in a better than 10'3 agreement with this other law of basic physics, ki¬ netic gas theory. The measurement process is moni¬ tored for "accuracy" by the kinetic gas theory, which gives it an additional feature for being qualified as "pri¬ mary method of measurement". Conclusions Measurements of ratios of isotope amounts (isotopic measurements) are revealed to be of great fundamental importance not only because of their inherent potential for very small uncertainties but also for their measure¬ ment potential directly in SI units. They are fully under¬ stood and are corroborated by the fundamental con¬ stants network at the 3 10’5 combined relative uncer¬ tainty level. The measurement process itself seems to conform with kinetic gas theory at the 10'3 combined relative uncertainty level or better. When combined with isotope dilution, they offer the potential of opening a new route for establishing traceability of amount mea¬ surements to (a realization of) the mole, perhaps one could say, to the Avogadro constant. Acknowledgements: A number of persons contributed substantially to the work presented and their contribution is warmly acknowledged; S Valkiers, T Murphy, S Peiser, G Lenaers, H Ku, P Hansen and D Vendelbo. I would also like to warmly acknowledge the painstaking re-editing work of H Kerslake for typing text and tables, and H Koekenberg for preparing figures. References Basile G, Bergamin A, Cavagnero G, Mana G, Vittore E & Zosi G 1995 The (220) lattice spacing of Silicon. IEEE Transactions on Instrumentation and Measurement 44:526-529. Cohen E R & Taylor B N 1987 The 1986 adjustment of the funda¬ mental physical constants. Review of Modern Physics 59:1121-1148. De Bievre P 1990 Isotope dilution mass spectrometry: What can it contribute to accuracy in trace analysis? Fresenius Journal of Analytical Chemistry 337:766-771. De Bievre P 1993 Isotope dilution mass spectrometry as a pri¬ mary method of analysis. Analytical Proceedings 30:328-333. De Bievre P & Peiser H S 1994 Calling for research in the science of measurement, a process exemplified by redeterminations 7(0 ln /TFT + Vo) (eq,12) 18 Journal of the Royal Society of Western Australia, 79(1), March 1996 of the Avogadro constant. Proceedings of the National Con¬ ference of Standards Laboratories, Workshop and Sympo¬ sium, Chicago, 565-578. De Bievre P, Lenaers G, Murphy T J, Peiser H S & Valkiers S 1995 The chemical preparation and characterization of specimens for "absolute" measurements of the molar mass of an ele¬ ment, exemplified by silicon, for redeterminations of the Avogadro constant. Metrologia 32:103-110. De Bievre P, Valkiers S & Peiser H S 1994 New Values for Silicon Reference Materials, certified for isotope abundance ratios. Journal of Research of the National Institute of Standards and Technology 99:201-202. De Bievre P, Valkiers S, Peiser H S, Becker P, Ludicke F, Spieweck F & Stiimpel J 1995 A more accurate value for the Avogadro Constant. Conference on Precision Electromagnetic Measure¬ ments. IEEE Transactions on Instrumentation and Measure¬ ment 44:530-532. De Bievre P, Valkiers S, Schaefer F, Peiser FI S & Seyfried P 1994 High-accuracy isotope abundance measurements for Metrol¬ ogy. PTB Mitteilungen 104:225-236. Deslattes R 1988 X-ray interferometry and y-ray wavelengths. In: The Art of Measurement (ed. B Kramer). VCH, Weinheim, 193-208. Deslattes R D, Hennis A, Bowman H A, Schoonover R M & Carroll C L 1974 Determination of the Avogadro constant. Physical Review Letters 33:463-466. Deslattes R D, Hennis A, Schoonover R M, Carroll C L & Bow¬ man H A 1976 Avogadro constant - Corrections to an earlier report. Physical Review Letters 36:898-899. Fujii K, Tanaka M, Nezu Y, Sakuma A, Leistner A & Giardini W 1995 Absolute measurement of the density of silicon crystals in vacuo for a determination of the Avogadro Constant. IEEE Transactions on Instrumentation and Measurement 44:542- 545. Kind K & Quinn T 1995 Metrology: Quo Vadis? IEEE Transac¬ tions on Instrumentation and Measurement 44:85-89. Sacconi A, Peuto A M, Mosca M, Panciera R, Pasin W & Pettorruso S 1995 The IMGC volume-density standards for the Avogadro Constant. IEEE Transactions on Instrumenta¬ tion and Measurement 44:533-537. Seyfried P, Becker P, Kozdon A, Ludicke F, Spieweck F, Stiimpel J, Wagenbreth H, Windisch D, De Bievre P, Ku H H, Lenaers G, Murphy T J, Peiser H S & Valkiers S 1992 A determination of the Avogadro constant. Zeitschrift fur Physik B. Condensed Matter 87:289-298. Taylor B N & Cohen E R 1990 Recommended values of the fun¬ damental physical constants: a status report. Journal Re¬ search National Institute of Standards and Technology 95:497- 523. 19 Journal of the Royal Society of Western Australia, 79:21-25, 1996 Atomic weights: From a constant of nature to natural variations N E Holden High Flux Beam Reactor, Brookhaven National Laboratory, Upton NY USA 11973 Abstract The concept of Atomic Weight of an element as a constant of nature has played a key role in science for almost two hundred years. Two centuries ago, it helped provide credence to the atomic theory of matter. One hundred fifty years ago, the periodicity in the properties of various elements as a function of their atomic weight helped lead to the discovery of the Periodic Table and the classification of the chemical elements. Early in this present century this constant of nature concept was shaken when elements were found, which had different atomic weight values as well as different radioactive properties but they had the same chemical properties and there¬ fore were located in the same position in the Periodic Table. To solve this problem, the concept of isotopes was born. Finally fifty years ago, the variation in nature of the composition of the stable isotopes in carbon and oxygen was found, which led to a variation in their respective atomic weights. Now mass dependent chemical reactions, nuclear reactions both natural or man-made and radioactive decay processes force us to accept the idea that atomic weights are more likely to vary in nature than they are to be constants of nature. What should we expect from atomic weights in the future? Introduction The world is made up of an apparently endless vari¬ ety of substances; if each one is an entity in itself, the nature of matter must be forever incomprehensible. The Greeks first introduced the idea of atoms as elementary constituents of matter, but their atom was a vague gen¬ eral idea unattached to any specific facts or processes. John Dalton introduced his atomic theory and his table of atomic weights at the start of the 19th century. The Periodic Table was constructed using the periodicity of the chemical properties of elements in the ascending or¬ der of the atomic weights of these elements. As a result, the world was now made up of a moderate number of different real substances related in a single system and the nature of matter became comprehensible, but why did it take almost seventy years from the inception of the atomic weight concept to the publication of the Peri¬ odic Table? In the first decade of the twentieth century, new sub¬ stances were being discovered, almost daily, which had similar chemical properties to existing elements but with different atomic weight values. Bewildered scientists could not decide where to place these new discoveries in the Periodic Table. Did this invalidate the concept of atomic weights as a useful, chemical tool? Treating the variation in the lead atomic weight as a special case, atomic weights were still considered to be constants of nature into the latter half of this century. Has this view changed? Is the atomic weight concept still useful today? We will investigate these questions in detail below. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 Prehistory The ancient Greeks first developed the idea 2500 years ago that matter was composed of atoms (Greek: indivis¬ ible). They taught that all matter was composed of four elements; fire, water, air and earth (Holden 1984). The 16th century alchemist Paracelsus added to those ele¬ ments sulphur, salt and mercury- In the seventeenth century, the Irishman Robert Boyle denied both the Greek notion that the basic elements were fire, water, air and earth and Paracelsus' salt, sulphur and mercury. He developed chemical analysis - the tech¬ nique for breaking down substances into their most elemental parts. He defined an element as a material that could be identified by scientific experiment and could not be broken down into still simpler substances. This is the definition that is still in use today. The French scientist Antoine Lavoisier revolutionized chemistry by introducing accurate weighing. He deter¬ mined that a given amount of matter has a total mass as measured by a weight, which remains the same when it changes in chemical combination, whether in the solid, liquid or gaseous state. The French chemist Joseph Proust's analyses showed that a particular chemical compound always contained the same elements united in the same definite proportion by weight. Dalton's atomic theory The English school-teacher John Dalton tested Proust's law and noted that the same elements com¬ bined in different proportions to produce different sub¬ stances. In his atomic theory, all matter was made up of particles called atoms, which were alike in everything except their weight. In chemical reactions, atoms pre¬ served their identity and are not destroyed. When Dalton published his atomic theory, he included tables of atomic weight values (Dalton 1805). 21 Journal of the Royal Society of Western Australia, 79(1), March 1996 Dalton assigned weights to atoms and expressed the relations between atoms of elements in precise numeri¬ cal terms. It is possible to assign relative weights by determining the ratio in which elements reacted with each other. Having assigned hydrogen as his reference atom with atomic weight one, he calculated atomic weights by comparing weights of other atoms with that of hydrogen. When two elements combine in a com¬ pound, it is insufficient to merely determine the percent¬ age of each element in the compound. One must also determine the valence of each element in the compound. Valence is a measure of how many atoms of one ele¬ ment combine with an atom of the other element, e.g. is water HO, or H20, or perhaps H202? Dalton assumed that if only one compound of two elements is known, it contains one atom of each element. This led to many difficulties in the application of his atomic theory. Equivalent weights (atomic weight/valence) were quoted rather than atomic weights. Although his calculations were wrong, the principle was correct. However, the listing of atomic weights for some elements and fractions of atomic weights for other elements was very confusing and persisted for half a century. The English physician, William Prout (Prout 1815) noted that Dalton's atomic weight values of elementary gases were nearly exact multiples of that of hydrogen and suggested that hydrogen was the primordial matter from which all elements are formed. For a while, it ap¬ peared that a number of atomic weight values agreed with this "Law". Testing the "Law" led to a major mea¬ surement effort of atomic weight values over the re¬ mainder of the century. The Italian physicist Amedeo Avogadro suggested (Avogadro 1811) that all gases under the same condi¬ tions of temperature and pressure contain the same number of molecules and a molecule (Greek: a small mass) may contain more than one atom. He made a distinction between the chemical atom (smallest part of matter that can enter into combination) and physical molecule (smallest particle that can exist in a free state). This could have helped to solve the equivalent weight problem but unfortunately he used the term molecule throughout his discussion with a series of qualifying adjectives; integral, constituent and elementary. In those days, the terms atom and molecule were often used in¬ terchangeably. Some scientists understood Avogadro to imply that there could be half-atoms. This confusion caused Avogadro to be ignored for half a century. The French physicist, Joseph Gay-Lussac determined (Gay-Lussac 1809) that gases form compounds with each other in simple (numerical) volume ratios proving that Dalton's idea of combining gases by weight alone was insufficient. Atomic weights and the periodic table At the Karlsruhe Congress in September 3-5, 1860, about 140 of the leading European chemists met to for¬ mulate an area of agreement among chemists regarding the nature of atoms and molecules and to reach a con¬ sensus with respect to a mutually satisfactory atomic weight scale. The Italian chemist, Stanislao Cannizzaro presented his "Sketch of a Course in Theoretical Chem¬ istry" (Cannizzaro 1858), where he called attention to the value of Avogadro's distinction between atoms and molecules as an organizing device for the interpretation of chemical phenomena. Lother Meyer and Dimitri Mendeleev both attended this congress and subse¬ quently developed periodic tables of the chemical elements based on revised atomic weight values. Mendeleev left open spaces, when no known element filled that space (Mendeleev 1869). He also predicted the properties of these unknown elements. When scandium, gallium and germanium were discovered over the next sixteen years and agreed with Mendeleev's predicted chemical properties and atomic weight, the periodic table was established and the usefulness of atomic weights was further enhanced. As mentioned above, Prout's law spurred chemists to prodigious effort to measure atomic weights during the nineteenth century. Compare the Table from 100 years ago (Clarke 1896) with that of the International Com¬ mission for 1959 (the last one prepared on the oxygen = 16 scale). The elements not included in the 1895 Table were the noble gases and some rare earths, which had yet to be separated. Two thirds of the 1895 values listed agree to better than 1% and almost 40% agree to better than 0.1% with the 1959 values. Radioactivity and atomic weights At the end of the 1800s, many scientists felt that future progress was to be looked for in the measurement of variations in the sixth decimal place of fundamental con¬ stants such as the atomic weights. Roentgen's discovery (Roentgen 1895) of X-rays followed by Becquerel's ra¬ dioactivity discovery (Becquerel 1896) quickly changed that viewpoint. As radioactive materials were studied, many sub¬ stances were being found with various atomic weight values. The English chemist, Frederick Soddy, showed (Soddy 1911) the chemical identity of mesothorium (228Ra) and radium. In 1913, he concluded that there were chemical elements with different radioactive prop¬ erties and different atomic weights but with the same chemical properties and therefore occupying the same position in the Periodic Table. He coined the word "iso¬ tope" (Greek: in the same place) to account for these radioactive species. The study of the natural radioactive decay chains for thorium and uranium led to speculation that these par¬ ent isotopes, 232Th and 23MU would decay into different daughter isotopes of lead, 208Pb and 206Pb, respectively. The lead from radioactive minerals should differ in atomic weight according to the proportion of uranium and thorium in the mineral. The atomic weight value for "common" lead (from a non-radioactive source ma¬ terial) was measured to be 207.2 (Baxter & Wilson 1908). Soddy & Hyman (1914) measured lead in a thorium sili¬ cate mineral to have an atomic weight value of 208.4. Richards & Lembert (1914) measured the atomic weight of lead in uranium minerals as low as 206.4. Could stable lead be made up of a mixture of iso¬ topes, each of a different whole number atomic weight? Was the overall atomic weight a fraction only because it 22 Journal of the Royal Society of Western Australia, 79(1), March 1996 was an average? Radioactivity contributed to this prob¬ lem in the case of lead, but what about the case of non¬ radioactive elements? J J Thomson discovered the electron, which was found to be over a thousand times less massive than even the lightest atom. He then studied the rare gas neon in 1912 by sending a stream of cathode ray electrons through the neon gas. These cathode ray electrons knocked some electrons off of neon atoms, which left these neon atoms with a positive electric charge, so- called neon ions. In the combined presence of a magnet and an electric field, the neon ions move in a curved path. If all neon ions had the same mass, all would fol¬ low the same curve. If some were more massive than others, the more massive ones would curve less. Thomson detected the neon ions at the end of their path on a photographic plate. He measured the darkening of the plate and found two locations which from the amount of curvature had to be 20Ne and 22Ne. The inten¬ sity of darkening indicated amounts of 90% and 10%, respectively- The overall atomic weight of neon, 20.2, was the average atomic weight of these two isotopes. Thomson's instrument, the fore-runner of the "mass spectrometer", was the first one capable of separating isotopes. The Englishman, Francis W Aston used a mass spec¬ trograph (Aston 1929) to analyze a sample of lead show¬ ing lines on the photographic plate at masses 206, 207 and 208 with intensities of 100, 10.4 and 4.5, respec¬ tively. Aston concluded that mass 207 must be the end product of the actinium radioactive decay series and was probably derived from an isotope of uranium and it would have a mass of 235. 2^U was found six years later. Lead has four isotopes, of which only 204Pb is not produced from radioactive decay. The American physi¬ cist, Alfred Nier used this peak as a reference in a mass spectrometer. He showed (Nier 1938) that the relative abundances of the lead isotopes varied widely even in common lead, which had a nearly constant atomic weight value. Nier's work on lead's isotopic composi¬ tion was also useful for dating purposes and the mea¬ surement of geological time (Nier 1939). There is no longer a case of an element like lead hav¬ ing varying isotopic compositions but a constant atomic weight because atomic weights are now determined by isotope mass spectrometry almost exclusively (De Bievre 1973). The atomic weight scale The atomic weight scale H = 1 was originally con¬ ceived used by Dalton and was used for 100 years. The Commission on Atomic Weights changed to the O = 16 scale with it's 1906 report (Holden 1984). Both hydrogen and oxygen were thought to not have isotopes. The dis¬ covery of oxygen isotopes in infrared spectra (Giauque & Johnson 1929a,b) led to a situation where the chemists scale of O = 16 differed from the physicists scale of lhO = 16. When a variation was found in oxygen's atomic weight in water versus air (Dole 1935), this implied a variation in the isotopic composition of oxygen and the two scales took on a small but variable difference. In April 1957 at a hotel bar in Amsterdam, Nier suggested (Holden 1984) that the 12C = 12 scale be adopted because of carbon's use as a secondary standard in mass spec¬ trometry. Physicists' approval was obtained, and in 1961 the atomic weights were officially given on the 12C = 12 scale for the first time (Cameron & Wichers 1962). Variations Although the atomic weight scale difficulty had been solved, another problem began to plague the Atomic Weights Commission. Nier & Gulbransen (1939) had made measurements on carbon which showed 5% varia¬ tion in the isotopic composition. The atomic weight would vary depending on the source of the material studied. Although the lead atomic weight variation could be ignored, the variation in carbon and in oxygen, mentioned earlier, made it apparent that atomic weights were not constants of nature. Variations in many light elements have since been found as well as variations due to radioactive decay in a parent affecting the isoto¬ pic composition and atomic weight of the daughter. For 30 years, restrictions on quoted atomic weight values have acknowledged these variations. Speculations and conclusions After the problem with lead, Richards had speculated on whether the supposed constant atomic weight mag¬ nitudes in chemistry were really variable? If so, how much effort should be expended in determining atomic weight values? One must determine the detailed varia¬ tion to understand causes of the variation. Evaluations of isotopic compositions have been added to the respon¬ sibility of the Atomic Weights Commission. Examples will illustrate some of the interesting problems that are now addressed. Reference has already been made to the use of the isotopic variations in uranium dating of geological times. There are a host of other dating methods which involve selecting a decay system with the same magni¬ tude of half-life as the age of the material to be studied. Boron is an element with a large probability for react¬ ing with neutrons, so boron was used as a standard for measuring other elements. However, these measure¬ ments at different laboratories gave dissimilar results, which was traced to the use of boron samples with dif¬ ferent atomic weights and isotopic composition at these labs. This variation now restricts the accuracy with which the atomic weight of boron can be quoted. In 1972, uranium ore from the Oklo mine in Gabon, West Africa was shown to contain too low an amount of 235U compared to normal uranium. Additional analyses indicated that the 235U had been burned up in a natural fission chain reaction under the ground about 2 billion years ago. The isotopic composition of various chemical elements was not consistent with normal samples of these elements but was consistent with the yield of the various isotopes as produced in the fission process (Ruffenach et al. 1980). These variations are now made note of when reporting the standard atomic weight val¬ ues. 23 Journal of the Royal Society of Western Australia, 79(1), March 1996 Table I. Comparison of 1895 & 1959 atomic weight values based on oxygen = 16.000 scale Element 1895 1959 Element 1895 1959 Element 1895 1959 Actinium Unknown (227) Glucinum 9.08 (Beryllium] Praseodymium 143.5 140.92 Aluminum 27.11 26.98 Gold 197.24 197.0 Promethium Artificial (145) Americium Artificial (243) Hafnium Unknown 178.50 Protactinium Unknown (231) Antimony 120.43 121.76 Helium Uncertain 4.003 Radium Unknown (226) Argon Uncertain 39.944 Holmium Unlisted 164.94 Radon Unknown (222) Arsenic 75.09 74.92 Hydrogen 1.008 1.0080 Rhenium Unknown 186.22 Astatine Unknown (210) Indium 113.7 114.82 Rhodium 103.01 102.91 Barium 137.43 137.36 Iodine 126.85 126.91 Rubidium 85.43 85.48 Berkelium Artificial (249) Iridium 193.12 192.2 Ruthenium 101.68 101.1 Beryllium (Glucinium) 9.013 Iron 56.02 55.85 Samarium 150.0 150.35 Bismuth 208.11 208.99 Krypton Unknown 83.80 Scandium 44.0 44.96 Boron 10.95 10.82 Lanthanum 138.6 138.92 Selenium 79.0 78.96 Bromine 79.95 79.916 Lead 206.92 207.21 Silicon 28.40 28.09 Cadmium 111.93 112.41 Lithium 7.03 6.940 Silver 107.92 107.873 Calcium 40.08 40.08 Lutetium Unknown 174.99 Sodium 23.05 22.991 Californium Artificial (251) Magnesium 24.29 24.32 Strontium 87.61 87.63 Carbon 12.01 12.011 Manganese 54.99 54.94 Sulfur 32.07 32.066 Cerium 140.2 140.13 Mendelevium Artificial (256) Tantalum 182.6 180.95 Cesium 132.89 132.91 Mercury 200.0 200.61 Technetium Artificial (99) Chlorine 35.45 35.457 Molybdenum 95.98 95.95 Tellurium 127.07 127.61 Chromium 52.14 52.01 Neodymium 140.5 144.27 Terbium 160.0 158.93 Cobalt 58.93 58.94 Neon Unknown 20.183 Thallium 204.15 204.39 Columbium 94.0 (Niobium) Neptunium Artificial (237) Thorium 232.63 (232) Copper 63.60 63.54 Nickel 58.69 58.71 Thulium 170.7 168.94 Curium Artificial (247) Niobium (Columbium) 92.91 Tin 119.05 118.70 Dysprosium Unlisted 162.51 Nitrogen 14.04 14.008 Titanium 48.15 47.90 Einsteinium Artificial (254) Nobelium Artificial (254) Tungsten 184.84 183.86 Erbium 166.3 167.27 Osmium 190.99 190.2 Uranium 239.59 238.07 Europium Unknown 152.0 Oxygen 16.000 16.000 Vanadium 51.38 50.95 Fermium Artificial (253) Palladium 106.36 106.4 Xenon Unknown 131.30 Fluorine 19.03 19.00 Phosphorus 31.02 30.975 Ytterbium 173.0 173.04 Francium Unknown (223) Platinum 194.89 195.09 Yttrium 88.95 88.91 Gadolinium 156.1 157.26 Plutonium Artificial (242) Zinc 65.41 65.38 Gallium 69.0 69.72 Polonium Unknown (210) Zirconium 90.6 91.22 Germanium 72.3 72.60 Potassium 39.11 39.100 24 Journal of the Royal Society of Western Australia, 78:43-54, 1995 Many elements are produced which are enriched in less abundant isotopes and are used as tracers in medi¬ cal diagnoses of processes in humans when use of radio¬ active tracers is not appropriate, e.g. in children and pregnant women. Note should be made that standard atomic weight values may not apply to these "doctored" elements. Carbon has two stable isotopes, 12C and 13C. The study of diet uses the ,3C abundance variation in the two ma¬ jor photosynthetic pathways; plants - wheat, rice, beans and nuts, are depleted in 13C relative to atmo¬ spheric CO, and as compared to C4 plants, such as com and sugar cane, which come from warm environments. Similarly, nitrogen has two stable isotopes, 14N and l5N. The abundance of l?N is enhanced in marine plants relative to land plants. This can be used to study changes in diet, when our ancestors moved from a hunting society to one dependent on marine life and on to the cultivation of plants. Earth and planetary science studies the isotopic anomalies (Shima 1989; Shima & Ebihara 1989) in meteor¬ ites and moon rocks to understand differences in pro¬ cesses of origin of the solar system 1-1.5 1010 years ago compared to the earth some 4 or 5 10y years ago. We have seen how the concept of atomic weights has evolved over the past two centuries. There was much interest when atomic weights were considered constants of nature and even more interest now that they are known to be variable. The demonstrated uses (de Laeter 1988, 1990; de Laeter et al 1992) of the underlying fun¬ damental isotopic compositions exceed the few ex¬ amples cited and I anticipate even more extensive uses of isotopes will be found in the future. Acknozvledgements: This paper is dedicated to Professor J de Laeter on his retirement as Deputy Vice Chancellor, Curtin University. This work was performed under the auspices of the US Department of Energy (Contract DE-AC02-76CH00016). References Aston F W 1929 Mass spectrum of uranium lead and the atomic weight of protactinium. Nature 123:313. Avogadro A 1811 Essai d'une maniere de determiner les masses relatives des molecules elementaires des corps, et les propor¬ tions selon lesquelles elles entrent dans ces combinaisons. Journal de Physique 73:58-76. Baxter G P & Wilson J H 1908 A revision of the atomic weight of lead. Journal of the American Chemical Society 30:187-195. Becquerel 1 1 1896 Sur les radiations emises par phosphorescence. Comptes Rendus des Seances de l'Academie des Sciences Paris 122:420-421. Cameron A E & Wichers E 1962 Report of the International Com¬ mission on Atomic Weight (1961). Journal of the American Chemical Society 84:4175-4197. Cannizzaro S 1858 Sunto di un corso di filosofia Chimica. Nuovo Cimento 7:321-366. Clarke F W 1896 Third annual report of committee on atomic weights. Results published during 1895. Journal of the Ameri¬ can Chemical Society 18:197-214. Dalton J 1805 On the absorption of gases by water and other liquids. Memoirs of the Literary and Philosophical Society of Manchester, second series 1:271-287. De Bievre P 1973 Atomic weights and isotopic abundances of the elements: a future concern for the analytical chemist. Zeitschrift fuer Analitisch Chemie 264:365-371. de Laeter J R 1988 Mass spectrometry in nuclear science. Mass Spectrometry Reviews 7:71-1 11. de Laeter J R 1990 Mass spectrometry in cosmochemistry. Mass Spectrometry Reviews 9:453-497. de Laeter J R, De Bievre P & Peiser H S 1992 Isotope mass spectrometry in metrology. Mass Spectrometry Reviews 11:193-245. Dole M 1935 The relative atomic weight of oxygen in water and in air. Journal of the American Chemical Society 57:2731. Gay-Lussac J L 1809 Memoire sur la combinaison des substances gazeuses, les unes avec les autres. Memoires de physique et de chemie de la Societe d'Arcueil 2:207-234. Giauque W F & Johnson H L 1929a An isotope of oxygen, mass 18. Nature 123:318. Giauque W F & Johnson H L 1929b An isotope of oxygen of mass 17 in the earth's atmosphere. Nature 123:831. Holden N E 1984 The International Commission on Atomic Weights - an early historical review. Chemistry International 6:5-12. Mendeleev D I 1869 The relationships between the properties of elements and their atomic weights. Journal of Russian Chemi¬ cal Society 1:60-77. Nier A O 1938 Variations in the relative abundances of the iso¬ topes of common lead from various sources. Journal of the American Chemical Society 60:1571-1576. Nier A O 1939 The isotopic constitution of radiogenic leads and the measurement of geological time II. Physical Review 55:153-163. Nier A O & Gulbransen E A 1939 Variations in the relative abun¬ dance of the carbon isotopes. Journal of the American Chemi¬ cal Society 61:697-698. Prout W 1815 On the relation between specific gravities of bodies in their gaseous state and the weights of their atoms. T Thomp¬ son Annals of Philosophy 6:321-330. Prout W 1816 Corrections of a mistake in the essay on the rela¬ tionship between specific gravities of bodies in their gaseous state and the weights of their atoms. T Thompson Annals of Philosophy 7:111-113. Richards TW & Lembert M E 1914 The atomic weight of lead of radioactive origin. Journal of the American Chemical Society 36:1329-1344. Roentgen W C 1895 Ueber eine neue Art von Strahlen. Sitzungsberichte der Physikalisch-Medicinischen Gesellschaft zu Wurzburg. 132-141. Ruffenach J C Hagemann R & Roth E 1980 Isotopic abundance measurements a key to understanding the Oklo phenomenon. Zeitschrift fuer Naturforschung 35a:171-179. Shima M 1989 Isotopic composition of elements in extra-terres¬ trial materials II. Shitsuryo Bunseki 37:195-227. Shima M & Ebihara M 1989 Isotopic composition of elements in extra-terrestrial materials. I. Shitsuryo Bunseki 37:1-31. Soddy F 1911 The chemistry of mesothorium. Journal of the Chemical Society 99:72-83. Soddy F & Hyman H 1914 The atomic weight of lead from Ceylon Thorite. Journal of the Chemical Society 105:1402- 1408. 25 Journal of the Royal Society of Western Australia, 79:27, 1996 Ion optical design for mass spectrometers S W J Clement Research School of Earth Sciences, Australian National University, Canberra ACT 6001 Abstract The efficient operation of mass spectrometers, since the development of the earliest instru¬ ments, has depended on the analogy between the optics of light and the 'focussing' effect of various electromagnetic fields acting on charged particles. Without the first order refocussing of the simple magnetic sector field, even modest mass resolution would be achieved only by ex¬ treme cohimation of the ion beam and at a heavy penalty in sensitivity. Over the years, the increasing level of sophistication in the theoretical analysis of mass spectrometers as ion optical systems has led to real improvements in the operational performance of the instruments as analytical tools. It is now common to find mass spectrometers with non-normal incidence to the magnet, curved field boundaries, contoured magnet pole-faces, and spherical or toroidal electro¬ static analysers. These and other devices are now being used to provide first and higher order focussing in both the dispersive and non-dispersive planes. Further insight into the operation of mass spectrometers has been gained by the application of the phase space concepts originally introduced in the beam transport field of nuclear physics. The ideas of beam emittance and instrument acceptance are now widely understood and their correct matching is seen as fundamental to the ion optical design of a new instrument. Liouville's Theorem has also emphasised the relationship between mass resolution, ion transmission effi¬ ciency and instrument parameters such as accelerating potential and magnet radius. Sensitive High Resolution Ion Micro-Probe (SHRIMP) at Curtin University in the Isotope Science Research Centre. See Compston (1996, in this issue) for the development of SHRIMP. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 27 / Journal of the Royal Society of Western Australia, 7 9:29-32, 1996 Clean laboratories: Past, present and future J R Moody Chemical Science and Technology Laboratory, National Institute for Standards and Technology, Gaithersberg MD USA 20899. Abstract Clean chemistry laboratories have become rather ubiquitous for the ultra-trace analysis of elements, especially among those laboratories that work in the area of isotope geochemistry. The need for the control of analytical blanks has almost always been appreciated, but became the primary concern of analysts as new measurement technology opened the door to elemental measurements on amounts of substance equal to a nanogram or less. Since common laboratory environments expose samples to a contaminant deposition of up to one microgram in a 24 hour period, cleaner environments were recognized as necessary. In the 1960s, a number of laborato¬ ries applied clean air technology from the space and electronics industries to the chemical labora¬ tory. The results from these pioneering efforts and the need to accurately measure elemental and isotopic compositions on the Apollo lunar samples led the National Bureau of Standards to build several complete clean air laboratories dedicated to sample preparation wet chemistry. These laboratories were simple but effective in design and have been extensively copied by other labora¬ tories. By the early 1980s, the issue of air quality (the Class of the clean air) was beginning to be recognized by many as less important than the other issue in clean laboratory design. That is, contamination reduction may depend as much or more on keeping gross quantities of analyte elements out of the laboratory and preventing the formation of undesirable corrosion products. This was not news for the geochemistry community, but by the 1980s it had become a main¬ stream problem for trace element analytical chemistry, primarily because of the new focus of analytical chemistry on environmental measurements. The question now is where does clean laboratory design go to exact further improvements in laboratory performance. Part of the answer may lie in examining other trends in analytical chemistry and projecting how a clean laboratory might accommodate them. Introduction Clean chemistry laboratories for the ultra-trace analysis of elements have become common, especially among those laboratories that work in the area of isotope geo¬ chemistry. However, back in the 1960s most analysts were still working in open laboratories. Most wet analy¬ sis was performed in glass apparatus using commer¬ cially available reagents equivalent to the American Chemical Society's (ACS) Reagent Grade. Elemental blanks from the reagents and apparatus alone amounted to more than a few micrograms. Some geologists were fortunate enough to be working with isotopes of ele¬ ments of relatively low abundance. Few analytical chem¬ ists considered the environmental blank contribution to be a limiting factor in their analyses. Fortunately, iso¬ tope geologists were among the first to look for ways to lower the analytical blank since they were often limited in sample size or they were analyzing isotopic composi¬ tions at very low elemental concentrations. At the US National Bureau of Standards (NBS; now National Insti¬ tute for Standards and Technology, NIST) in the late 1960s, scientists were measuring the elemental deposi¬ tion from the air in the laboratories and they could eas¬ ily find significant contaminations of 0.1-0. 5 pg per day for elements like lead. As the emphasis in inorganic analytical chemistry turned more toward trace analysis, analysts in many laboratories were discovering contamination problems in their measurements. The isotope geology laboratories and semiconductor electronics laboratories were the first to take positive steps to minimize contamination. Zief & Mitchell's (1976) book on contamination control was a real turning point in trace analysis for the general ana¬ lytical community; it detailed procedures that had been in use at select laboratories for a number of years. A paper by Patterson & Settle (1976) represents another milestone in approaches to contamination control. Al¬ though both may be difficult to obtain now, they are superlative references since relatively little has changed in clean lab procedures from the principles presented by these authors. The NBS laboratories have published a guide to clean laboratory construction (Moody 1982) that has been widely used by many laboratories around the world. Although the basic design is still valid, we will try to project what the analytical clean laboratory might look like in ten years. Historical Approaches to Contamination Control The simplest approach to the control of contamina¬ tion from laboratory air was to enclose the experiment inside a plastic box. Contamination was thus reduced to the amount of contamination contained within the box. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 29 Journal of the Royal Society of Western Australia, 79(1), March 1996 A more sophisticated approach was to use a filtered sup¬ ply of nitrogen or another inert gas to flush the sample containment box and to keep it under a slight positive pressure. This prevented the influx of contaminants from outside of the box. Gloved ports were used for access to the sample. The major problem with this ap¬ proach was the inconvenience involved. Dissolutions in¬ side the box could be tricky and some sort of fume eradicator (a canopy to siphon off fumes) was neces¬ sary. In the late 1960s and early 1970s, a few laboratories started to explore the use of clean air technology for application to the laboratory. The object was to over¬ come some of the difficulties of working in a closed system and still preserve the contamination control as¬ pects in a new configuration. At this time, clean air was already extensively used in electronics manufacturing, medicine, and aerospace. The HEPA (High Efficiency Particulate) filters used were typically 99.97% efficient at 0.5 micrometer particle sizes. In an ordinary laboratory, a few filters could provide a huge improvement in the overall laboratory air quality and relatively high air quality in the designated clean bench area. Several of these designs were published (Patterson & Settle 1976; Zeif & Nesher 1974; Mitchell 1973). All required rela¬ tively stringent laboratory access procedures to keep the laboratory environment clean. None would meet strict air quality definitions but all were certainly useful. An example of such a laboratory was the clean labo¬ ratory at the University of Ghent, Belgium. Designed by Versieck and Cornelis, the laboratory was used for their pioneering work on the analysis of trace elements in biological and clinical samples. Mitchell (1973) at the Bell Telephone Laboratories in the USA was another early pioneer in the development of a working clean laboratory for chemical operations. Typical clean air flow amounted to the equivalent of one room air change in a period of one to three minutes. NBS purchased a commercial clean room for chemical analysis in 1968 with much higher air flows, although nothing else in the laboratory was optimised for chemistry. This laboratory did not perform well for chemical applications. The greatest change in the approach to chemistry clean labo¬ ratory design probably occurred in 1971 at NBS when the inorganic mass spectrometry group built a new clean laboratory specifically for the analysis of the Apollo lu¬ nar rock samples. The adoption by NBS of a new ap¬ proach to clean laboratory design would not have hap¬ pened without the prior four years experience with a commercial clean room and the results observed in the efforts of other trace analysts. The resulting NBS clean laboratory and its successor were described in detail by Moody (1982). These labora¬ tories were simple but effective in design and have been extensively copied by other laboratories. By the early 1980s, the issue of air quality (the Class or cleanliness of the clean air) was beginning to be recognized by many as less important than other issues in clean laboratory design. That is, contamination reduction in a clean labo¬ ratory may depend as much or more on keeping gross quantities of analyte elements out of the laboratory and preventing the formation of undesirable corrosion prod¬ ucts. This was not news for the geochemistry commu¬ nity, but, by the 1980s contamination and its prevention had become a mainstream problem for trace element analytical chemistry, primarily because of the new focus of analytical chemistry on environmental measurements. Present Day Clean Laboratory Design At the present time, the International Standards Organisation (ISO) is preparing a new standard for the measurement and classification of clean air and other standards are also undergoing some revisions. At this time it is premature to quote new draft technical stan¬ dards. These standards will harmonise nomenclature and clarify classifications of laboratories but the stan¬ dards do not address real laboratory performance for chemistry applications. Technical standards aside, the actual quality of the HEPA filter has been improved greatly over the last twenty years allowing the air qual¬ ity in the laboratory to be improved by a factor of one hundred or more over what was achievable in the early 1970s. The design of the filter has also improved, allow¬ ing the trace analyst to specify both a more efficient filter and metal-free filter body. It would be a mistake, however, to assume that improvements in the clean laboratory air quality have led to a corresponding im¬ provement in the analytical blank. Most of the earliest clean rooms for chemistry em¬ ployed horizontal air flow. They also had a relatively small HEPA filter area relative to the volume or area of the laboratory. The NBS (NIST) design adopted a modi¬ fied vertical flow industrial design. This approach re¬ quired a total filter area, volume of air flow, and short¬ ness of residence time for air in the clean laboratory that was significantly different compared to contemporary chemistry laboratories employing clean air. While the problems in chemistry applications are different from industry, the NBS (NIST) design shared more than a few common elements with industrial design. A deliberate attempt was made to keep the entire laboratory involved in the air circulation pattern to both reduce the residence time of the air in the laboratory and to effectively flush particles from the room that were produced by chemical processes. Fume hoods were designed to handle perchloric acid in a laminar flow clean air environment Whereas most industrial clean rooms operate at almost 100% recycling of the clean air, the NBS laboratory exhausted 35% or more of the total clean air through the fume hoods, thus reducing the recycled air to less than 65%. This meant that a high proportion of dirty make up air had to be introduced to the room to maintain the pressurisation of the laboratory. This created another break with traditional clean room design. With near 100% recycling, environ¬ mental controls (temperature and humidity) have to be built into the clean air handling system. It was not un¬ usual to see facilities that devoted as much space to the installation of technical equipment as was devoted to the clean air space. This was expensive and wasted valu¬ able laboratory space. Because the NBS laboratories were retrofitted into ex¬ isting spaces, all available space was at a premium. Con¬ sequently NBS turned to a system of modular clean air units that could be incorporated into the ceilings of ex¬ isting laboratories. Since fume exhaust was to be about 35% of the total room air, the make up air normally 30 Journal of the Royal Society of Western Australia, 79(1), March 1996 provided to the laboratory could handle all of the air conditioning and heating needed without having to modify any of the existing building air handling sys¬ tems. The design was significantly cheaper than other approaches and over the years it has performed as well as or better than more conventional clean room designs. The entire sample handling areas have very high quality vertical laminar flow clean air which effectively isolates the samples from the analyst(s). All aspects of sample handling, dissolution, and separation were accommo¬ dated within the clean laboratory to eliminate the neces¬ sity of going in or out of the laboratory to perform an¬ other sample preparation step. The result of these and other factors such as ultra-purified reagents led to a sub¬ stantial improvement (factor of 10 to 1000 or better) in the analytical blank in our laboratory and in others with similar operating conditions. Literally hundreds of cop¬ ies of the laboratory have been constructed over the years with little change in the basic design. The NBS (NIST) designers tried to provide the maxi¬ mum performance for the least cost. However, despite the enormous improvements in measured air quality in the newest NIST clean laboratory, there has been little if any improvement in the analytical blanks attributable to the clean laboratory. Where blank reductions have been achieved, the results were through improvements in the analytical method, separations, or choice of reagents em¬ ployed. The clean laboratory blanks seem to have been constant. One reason for this is that our laboratories are not totally metal free, leading to contamination by sec¬ ondary causes such as touching or handling of artifacts in the laboratory that ultimately transfer to the samples when handled. Another reason is that to keep the labo¬ ratory as fully utilized as possible and to make its use attractive to staff, personnel hygiene restrictions have been relatively relaxed. Thus, improvements in proce¬ dures could be used to improve the laboratory perfor¬ mance even further. Nevertheless, the clean laboratory itself does not seem to be a major contributor to analyti¬ cal blanks. Other factors such as ion exchange resins, reagents, and labware are larger contributors to the blank problem. Recently, a few laboratories have attacked this one weakness of the NBS/NIST design clean laboratory. Two laboratories that have done an outstanding job of reducing secondary contamination sources (contamina¬ tion not transmitted by the HEPA filter) are Boutron's laboratory in France (Boutron 1990) and De Bievre's laboratory at the Institute for Reference Materials and Measurements (IRMM) in Belgium. Both laboratories have made extensive use of plastics to replace other materials of construction in the laboratory. The IRMM laboratory in Belgium is probably the ultimate achievement for the present clean air technology and laboratory de¬ sign. Since the NBS or NIST designs are widely known and published, it may be more useful to examine how IRMM improved on the last NIST design. As with all discussions of clean laboratory design, the reader must judge the best design or feature for their application. Planning for the IRMM facility began with NBS assis¬ tance in 1985 and continued for several years before a final design was accepted. All commercial clean rooms employ the principle of shelling or compartmentalizing clean operations inside of progressively cleaner environ¬ ments. Patterson and Settle’s (1976) chemical and sam¬ pling operations were based on the same principle, and these have been proven to be effective and necessary in instances of extremely low analyte concentration. Where the NBS laboratory employed zero or minimal progres¬ sion into the cleanest environment, the IRMM facility has four distinct air quality zones starting with the com¬ plete isolation of the clean laboratory building from the adjacent mass spectrometry facilities. The laboratory fa¬ cilities are classified as Class 1000, Class 100, and Class 10. Actual performance is better than specified. The entire clean laboratory building at IRMM is physically isolated and supplied with HEPA filtered clean air. Although our experience at NIST has indi¬ cated that the HEPA filters may last for four to five years, there is little doubt that the relatively poorly fil¬ tered air supply to the NIST laboratories shortens the useful lifetime of the HEPA filters. The major justifica¬ tion for HEPA filtering for the entire building air supply at IRMM was that it would extend the clean laboratory HEPA filter life and also provide a high quality air sup¬ ply for the building to further reduce the chance of con¬ tamination in the laboratories. Initially, this approach is expensive, but the redundant HEPA filter chambers used for the building air conditioning mean that the fil¬ ters may be serviced or replaced with no interruption of service to the laboratories. In addition to the highest level of air quality, all of the laboratory components are made of plastic or of ma¬ terials that are fully protected from corrosion caused by wet chemical processes. The restrictions on use by labo¬ ratory personnel together with the modular design and metal free constructions go as far as seems possible to¬ wards reducing secondary contamination from the labo¬ ratory and from the chemists. The IRMM laboratory de¬ sign probably represents the final stage of clean room evolution using conventional laboratory approaches. The question now is where does clean laboratory design go to extract further improvements in laboratory perfor¬ mance. Part of the answer may lie in examining other trends in analytical chemistry and projecting how a clean laboratory might accommodate them. The Clean Laboratory of the Future The largest consumer of clean room apparatus is the electronics industry. All of the clean laboratories in the world constitute an almost imperceptible fraction of the clean rooms constructed for the electronics industry each year. One may start with the premise that the only equipment that one can buy is the equipment that is manufactured for this industry. The medical industry is another large consumer of clean room equipment, but their equipment is still derived from the semiconductor market. One disturbing trend for analysts is a current trend in the semiconductor industry. With the ever de¬ creasing size of microelectronic circuitry on chips, the demand for better air quality (ULPA filters, for instance) is leading to some manufacturing changes. In the 1970s we were careful to move from steel filter frames to alu¬ minium frames and then to wood frames (metal-free) for the HEPA filters. With the demand for low particle counts, the wood frame is being phased out to accom¬ modate the use of metal filter frames. Metals do not 31 Journal of the Royal Society of Western Australia, 79(1), March 1996 contract or expand with temperature or humidity as much as wood. The reality is that we will have to adapt to the use of metal frames again over the next five years or so. This will create some problems in making these filters suitable for use in a trace metals laboratory. Par¬ ticle counts will be lower, but much more care will be needed to protect the HEPA filter from corrosion. The other trend that one may see in the electronics industry is a rapid adoption of automation to remove the human operator from the process. This in turn is leading to changes in equipment design and philosophy. In analytical chemistry, you may also see this trend with the rapid adoption of laboratory automation. The moti¬ vation for the laboratory manager may be to reduce laboratory costs more than to improve performance. Nevertheless, it is probably safe to assume that automa¬ tion will become a majority factor in laboratories of the near future. Wet chemists with the knowledge to per¬ form many laboratory operations are also disappearing. Based upon these trends, this author speculates that the laboratory of the future will be built around auto¬ mated apparatus. Further, it is likely that the most effi¬ cient way to do this is to adopt the semiconductor in¬ dustry developed "mini-environment/' The mini-envi¬ ronment is a fully self contained clean process with clean air, robotics, wet benches, etc. built into a closed cham¬ ber. These may be grouped and ganged together to per¬ form sequential operations. They provide much better air quality with total isolation from the clean room in which the mini-environment is located. Ventilation and heating requirements may be greatly reduced since the environment provided is what the process needs, not what humans require. Chemists would have relatively little direct contact with a sample in this kind of clean room. Finally, there is one other advantage to the mini¬ environment approach. The cost is much less than would be needed to achieve similar process performance by conventional means. If building a clean laboratory of about 100 m2 today, one might spend well in excess of $(US)2, 000,000. The cost for a clean room of similar performance utilizing mini-environments would be much less than half of that. Operating costs could be similarly cheaper for the minienvironment. Such a laboratory could not be built today since the automated laboratory equipment does not yet exist for many analytical operations. Finally, the analytical chemical process itself may be significantly improved with a more consistently clean operation with¬ out human intervention. No other large improvements in clean laboratory performance seem likely with the present design approaches. References Boutron, C F 1990 A Clean laboratory for ultralow concentration heavy metal analysis. Fresenius Zeitschrifter Analtische Chemfuerie 337:482-491. Mitchell J W 1973 Ultra purity in trace analysis. Analytical Chemistry 45:492 A-500 A. Moody J R 1982 NBS Clean laboratories for trace element analy¬ sis. Analytical Chemistry 54:1358A-1376A. Patterson C C and Settle D M 1976 The reduction of orders of magnitude errors in lead analysis of biological materials and natural waters by controlling the extent and sources of indus¬ trial lead contamination introduced during sample collecting, handling and analysis. Proceedings of the Seventh Materials Research Symposium, US Government Printing Office, Wash¬ ington DC. Zief M & Mitchell J L 1976 Contamination Control in Analytical Chemistry. Wiley; New York. Zief M & Nesher A G 1974 Clean environment for ultratrace analysis. Environmental Science & Technology 8:677-678 32 Journal of the Royal Society of Western Australia, 79:33-42, 1996 Meteorites recovered from Australia A W R Bevan Department of Earth and Planetary Sciences, Western Australian Museum, Francis Street, Perth 6000 Abstract Meteorites are our only tangible source of information on the earliest history of the Solar System. Over the last ten years the number of meteorites described from Australia has doubled, and this has stimulated many new lines of enquiry. To date, fragments from a total of 474 distinct and authenticated meteorites have been recovered in Australia. The material, including 13 mete¬ orites observed to fall, comprises 389 stones (21 achondrites, 366 chondrites and 2 unclassified stones), 71 irons, 13 stony-irons and one meteorite of unknown class. Two hundred and fifty- seven distinct meteorites are currently known from Western Australia, 123 from South Australia, 48 from New' South Wales, 20 from Queensland, 12 from the Northern Territory, 10 from Victoria and 4 from Tasmania. Discoveries include the first lunar meteorite (a fragment of the Moon), Calcalong Creek, found outside of Antarctica. Five meteorites (Veevers [iron], Wolf Creek [iron], Henbury [iron], Boxhole [iron], and Dalgaranga [mesosiderite stony-iron]) are associated with craters. Another eighteen impact structures (lacking meteorites) are known in Australia and the country has one of the world's best preserved impact cratering records stretching back more than 500 million years. Most meteorites in Australia have been found in the Nullarbor Region, which for climatic and geological reasons is one of the most prolific areas of the world for meteorite recoveries outside of Antarctica. Since 1971, several thousand specimens of an as yet unknown total number of distinct meteorites have been recovered from the Nullarbor, including many rare types. ,4C terrestrial ages of Nullarbor meteorites combined with population statistics are provid¬ ing important information about the number of meteorites falling with time. Moreover, weather¬ ing studies of ancient stony meteorite finds, and the stable isotopic composition of carbonate contamination derived from the Nullarbor limestones is yielding palaeoclimatic information for that region of Australia over the last 30,000 years. Introduction Meteorites are an unique source of information about the earliest history of the Solar System. Mostly fragments broken from small planetary bodies, called asteroids, in solar orbits between Mars and Jupiter, many meteorites have remained virtually unaltered since their formation 4.55 Ga ago. Some rare types of carbonaceous meteorite contain water and complex carbon compounds, includ¬ ing amino acids. These rocks may be similar to the origi¬ nal materials from which the Earth gained the water for its oceans, the gases for the atmosphere we breathe, and the building blocks of life. However, aside from the fun¬ damental question of the origin of life, basic research on meteorites is helping us to understand many other as¬ pects of our natural environment and history. Research on Australian meteorites essentially started with the publication by Haidinger (1861) of a description of two masses of the Cranbourne iron meteorite found in Victoria in 1854. On numerous occasions in the past, meteorites found in Australia have been reviewed, or listed {eg. Cooksey 1897; Anderson 1913; Prior 1923; Hodge-Smith 1939; Prior & Hey 1953; Hey 1966; Mason 1974; Gibbons 1977; Graham et al 1985; Bevan 1992a). Most recently, Bevan (1992a) provided a comprehensive review of meteorite recovery in Australia. However, since the early 1990's there has been a surge in the recovery of meteorites in Australia that has opened up many new lines of re¬ search. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 Figure 1 shows the numbers of meteorites known at various times from Australia during the period 1897- 1996. Mason (1974) documented a total of 184 distinct meteorites from Australia. During the period 1974-1992, data were published for 93 new Australian meteorites (Bevan 1992a). However, since the last review by Bevan (1992a), data for an additional 197 meteorites from Aus¬ tralia have appeared in the literature. This remarkable recovery rate is largely due to discoveries in the Western Australian and South Australian Nullarbor Region. For climatic and geological reasons, the Nullarbor Region is one of the most prolific desert areas of the world for meteorite recoveries outside of Antarctica (Bevan & Reviews Figure 1. Cumulative histogram of meteorites known from Aus¬ tralia at various times during the period 1897-1996. 33 Journal of the Royal Society of Western Australia, 79(1), March 1996 Binns 1989a, b,c; Bevan 1992a; Bevan & Pring 1993). Re¬ coveries from the Nullarbor alone account for more than 50% of all meteorites currently known from Australia, and more than a thousand recently recovered, and po¬ tentially new, Nullarbor meteorite fragments remain to be described (Bevan 1992b; Koeberl et al. 1992). The purpose of this paper is to review Australian me¬ teorites, particularly those recovered since the last re¬ view by Bevan (1992a), and with special reference to the unique environment of the Nullarbor Region, to exam¬ ine the climatic and physiographic factors that contrib¬ ute to the recovery of meteorites in Australia. Meteoritic materials and origin Only a brief resume of meteorite classification and genesis is presented here. For more detailed accounts, the reader is referred to the bibliography and references therein {e.g,. Dodd 1981; McSween 1987). Traditionally, meteorites have been divided into three major categories depending on the relative amounts of silicate and metallic minerals they contain. Iron meteor¬ ites are composed predominantly of iron-nickel metal; stony meteorites (often called 'stones') consist mainly of silicates, but also contain some metal and other acces¬ sory and minor mineral phases; and stony-irons com¬ prise metal and silicates in roughly equal amounts. About 95% of modern observed meteorite falls are stones, around 4% are irons and only 1% are stony-irons. This simple classification scheme, however, conceals the diversity of materials that have fallen to Earth, and mod¬ ern research has delineated numerous distinct groups and sub-types of meteorites. Stones Of the two main groups of stony meteorites recognised, the chondrites are the most numerous, ac¬ counting for 87% of all meteorites observed to fall. Chondrites contain millimetre-sized beads of stony min¬ erals, called chondrttles , from which the name of the group originates (Fig 2). Chondrules are unknown in any rocks from Earth and for more than a century argu¬ ments have continued over how these enigmatic objects might have formed. While there is still no consensus on Figure 2. Photomicrograph of the Forrest Lakes LL5 ordinary chondrite showing numerous rounded chondrules. the mechanism by which chondrules were generated, most researchers agree that chondrules were among the earliest materials to have formed in the Solar System. An understanding of the origin of chondrules, and sub¬ sequent chondrites, is fundamental to our understand¬ ing of how materials (and ultimately planets) formed in the infant Solar System. Chemical variations between chondritic meteorites define a number of distinct groups. The largest group, collectively known as the ordinary chondrites , accounts for more than half of all known me¬ teorites (observed falls + chance finds) {e.g. see Fig 3). Although 75% of their bulk is made up of silicate miner¬ als, ordinary chondrites contain substantial amounts of iron both in the form of silicates and as metal and iron- sulphide. Two other rarer groups of chondrite, enstatite and carbonaceous chondrites, represent extremes in com¬ position. The enstatite chondrites are rich in metal and sulphide, but the main silicate mineral they contain (enstatite) is a pure magnesium-silicate containing no iron, in contrast, carbonaceous chondrites contain little or no metallic iron, but their silicate minerals are mostly iron-rich. Figure 3. Mass of the ordinary chondritic (H5) meteorite that fell on Binningup beach, Western Australia, at 10:10 am on 30 Sep¬ tember, 1984. Carbonaceous chondrites are among the most important and intriguing classes of meteorites. Sometimes rich in complex carbon compounds such as amino and fatty acids, carbonaceous chondrites can also contain appre¬ ciable amounts of water (up to 20%), and water-bearing minerals that formed by the hydrothermal alteration of other minerals. A rare type of carbonaceous chondrite [Cl], named for the type meteorite lvuna (that fell in Tanzania), of which only a few are known, is composed almost entirely of minerals that formed at low tempera¬ tures. Significantly, some of these meteorites have chem¬ istries that closely match the Sun's, and are our best samples of 'average' Solar System material. The majority of meteorites are believed to be fragmental debris from the collision of asteroids, but there is astronomical evi¬ dence suggesting that some carbonaceous chondrites may have had a cometary origin. The achondrites make up around 8% of modern mete¬ orite falls. So-called because they lack chondrules, achondrites generally have textures showing that they formed in similar ways to some of the igneous and volcanic rocks on Earth. Although most achondrites are 34 Journal of the Royal Society of Western Australia, 79(1), March 1996 asteroidal in origin, twelve are fragments of the Moon. These "lunar" meteorites were probably ejected from the Moon by large, low-angle impacts. We are able to recognise lunar meteorites by comparison with the ref¬ erence collection of lunar rocks returned by the Apollo space missions and knowledge of the Moon's chemistry. Another twelve achondritic meteorites are planetary in origin and may be fragments of Mars (c.g. see Gladman et al 1996 and references therein). Unlike meteorites of asteroidal origin, these meteorites crystallized only 200- 1300 million years ago and must have come from a large, planetary-sized body that remained hot for much longer than the asteroids. Irons The chemical make-up of irons suggests that most solidified from molten accumulations of metal that could only have formed deep in the interiors of a num¬ ber of small asteroids (Fig 4). Irons often consist of two iron-nickel minerals arranged in a regular trellis-work structure of interlocking crystal plates. When cut and polished surfaces of some irons are treated with acid, this structure, called a Widmanstatten pattern is revealed (Fig 5). Widmanstatten structures formed as the result of extremely slow cooling of hot metal in the solid state. Cooling rate calculations indicate that many irons are fragments of the cores of small asteroidal bodies ranging up to a few hundred kilometres in diameter. Like achon- drites, most iron meteorites tell us that some bodies smaller than planets in the early Solar System melted and differentiated, separating metal from silicate to form 'cores' and 'crusts' like the Earth's. Other iron meteor¬ ites were never completely melted and some contain in¬ clusions of silicate. The modern classification of irons is based on chemistry, notably the abundance of Ni, Ga, Ge and Ir that they contain. Figure 4. Main mass, weighing 480 kg, of the Haig (group IIIAB) iron meteorite found on the Nullarbor Plain by Mr A J Carlisle in 1951 (photograph by D. Elford). Figure 5. Cut, polished and acid treated slice of the Haig iron meteorite showing the Widmanstatten pattern characteristic of this group of irons (sawn edge measures 8 cm). Stony-irons Stony-irons are by far the rarest of the main catego¬ ries of meteorites. Two main groups of stony-irons are recognised. Meteorites of the largest group, the paUasites , are composed of crystals of olivine set in metallic iron- nickel. Members of the other major group of stony-irons, called mesosiderites, are made up of mixtures of frag¬ ments of rock similar to some achondrites in composi¬ tion, and nuggets and veins of iron-nickel metal. There are also several anomalous meteorites (e.g. Bencubbin) that fit structurally into the stony-iron category, al¬ though these have no relationship to the two major groups. Bencubbin is a complex mixture of metal and silicate components, including a variety of chondritic xe- noliths an example of which is seen as a dark area on the cut face of the meteorite (Fig 6). Bencubbin is not related to any of the known major groups of meteorites although some components are chemically and isotopically Figure 6. Mass (originally 54 kg) of the Bencubbin meteorite found in July 1930. Scale bar is 10 cm. (photograph by K Brimmell). 35 Journal of the Royal Society of Western Australia, 79(1), March 1996 similar to the CR group of carbonaceous chondrites and an unique Antarctic chondrite Allan Hills 85085 (Barber & Hutchison 1991; Weisberg et al. 1995). The stony-irons may have formed by the mixing of both solid and liquid metal and silicates at various depths within their parent asteroids. A close relation¬ ship between some pallasites and one of the groups of irons suggests that they may have formed in the semi- molten regions between the metallic cores and rocky outer skins of small planet-like asteroids, whereas mesosiderites originated as mixtures of solid and liquid metal and achondritic rocks of diverse origins. Recovery of meteorites in Australia Currently, fragments from a total of 474 distinct and authenticated meteorites have been described from Aus¬ tralia. The material comprises 389 stones (21 achondrites, 366 chondrites and 2 unclassified stones), 71 irons, 13 stony-irons and one meteorite of unknown class. One meteorite, Murchison Downs, previously thought to be distinct has been shown by Bevan & Griffin (1994) to be a transported fragment of the Dalgaranga mesosiderite and is not included in the total. Two hundred and fifty- seven distinct meteorites are currently known from Western Australia, 123 from South Australia, 48 from New South Wales, 20 from Queensland, 12 from the Northern Territory, 10 from Victoria and 4 from Tasma¬ nia. Discoveries include the first lunar meteorite, Calcalong Creek, found outside of Antarctica (Hill et al. 1991). Only thirteen well-documented observed meteorite falls have been recorded from Australia (Bevan 1992a). The most recently recovered fall (Fig 3) is a single stone of an ordinary chondrite weighing 488.1 grams, that fell on Binningup beach in Western Australia on 30 Septem¬ ber, 1984 (Bevan et al. 1988). Several large fireballs, some associated with sonic phenomena, from which meteor¬ ites may have been deposited, have been recorded in Australia over the last few years (e.g. see McNaught 1993). However, no known material that can be linked to these events has been recovered. A list of the authen¬ ticated observed meteorite falls from Australia is given in Table 1. One of the most recent discoveries is a 34 kg mass of ordinary chondrite found near Broken Hill in December 1994 (Grossman 1996) In Australia, most chance meteorite recoveries, or finds, have resulted from the clearing of land for agricul¬ ture and pastoralism, and also mining and prospecting activity. The distribution of meteorite falls and finds in Australia is shown in Fig 7. The general lack of discov¬ eries in tropical Australia (north of latitude 23° S) prob¬ ably reflects a climate and physiography that are not conducive to the preservation and recognition of mete¬ orites, respectively. As noted by Mason (1974) and Bevan (1992a), there remain surprisingly few docu¬ mented discoveries from central Australia and Queensland. To date, the largest single mass of meteorite found in Australia is an 11.5 tonne fragment of the Mundrabilla iron found in 1966 on the Nullarbor Plain in Western Australia (Wilson & Cooney 1967). Since the discover}7 of this mass, more than twelve additional masses of the same meteorite totalling more than 22 tonnes have been recovered from a large area of the central Nullarbor in Western Australia ( e.g . see De Laeter 1972; De Laeter & Cleverly 1983; De Laeter & Bevan 1992 and references therein). Five meteorites (4 irons and 1 stony-iron) are associ¬ ated with meteorite impact craters (Bevan 1992a, 1996). Figure 7 shows the locations of the Dalgaranga, Veevers, Wolfe Creek, Boxhole and Henbury impact craters. Mount Darwin crater in Tasmania (Fig 7) is undoubt¬ edly of impact origin but no meteorites have been col¬ lected (Fudali & Ford 1979). An additional small crater, the Snelling crater, has recently been discovered in West¬ ern Australia (E S Shoemaker, pers. comm.). However, no meteorites are reported to have been collected from the locality. Throughout Australia another eighteen larger structures are known to varying degrees of certainty to be the deeply eroded remains of giant meteorite or aster- oidal impact craters (Shoemaker & Shoemaker 1988, 1996). No meteoritic material is known from these sites although there is, in many cases, abundant evidence of meteorite impact that may include characteristic macro¬ scopic shatter-cones, microscopic shock-metamorphic ef¬ fects in the target rocks, and noble metal geochemical anomalies. Table 1. Australian observed meteorite falls (in chronological order) Name Date of fall class State co-ordinates Tenham (spring) 1879 L6 Qld 25° 44'S 142° 57'E Rockhampton* (spring) 1895 stone Qld 23° 23’S 150° 31'E Emmaville 1900 Eucrite NSW 29° 28'S 151° 37'E Mount Browne 17.7.1902 H6 NSW 29° 48'S 141° 42'E Narellan 8.4. 1928 L6 NSW 34° 3'S 150° 41' 20"E Moorleah Oct. 1930 L6 Tas 40° 58.5'S 145° 36'E Karoonda 25.11.1930 CK4 SA 35° 5'S 139° 55'E Forest Vale 7.8.1942 H4 NSW 33° 21 ’S 146° 51' 30"E Millbillillie Oct. 1960 Eucrite WA 26° 27'S 120° 22'E Woolgorong 20.12.1960 L6 WA 27° 45 'S 115° 50'E Wiluna 2.9.1967 H5 WA 26° 35' 34"S 120° 19' 42"E Murchison 28.9.1969 CM2 Vic 36° 37'S 145° 12’E Binningup 30.9.1984 H5 WA 33° 09' 23"S 115° 40’ 35"E ^specimen lost 36 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 7. Geographical distribution of meteorite finds and observed falls in Australia and sites of five meteorite impact craters associated with meteorites, Wolfe Creek (W), Dalgaranga (D), Veevers (V), Henbury (H), Boxhole (B), and one crater, Mount Darwin (M), at which meteorites are lacking. The Nullarbor Region The anomalously large number of meteorites found in the Nullarbor Region (Fig 7) does not mean that more meteorites have fallen there than anywhere else in Aus¬ tralia, but reflects an unique physiographic environment and a sustained research effort to recover meteorites from the region. The Nullarbor Region is coincident with a geological structure, the Eucla Basin, that straddles the border between South Australia and Western Australia. The sedimentary basin comprises essentially flat-lying limestones of Lower-Middle Miocene Age ( ca . 15 Ma) outcropping over an area of ca. 240,000 km2. (Lowry 1970). The arid to semi-arid climate of the Nullarbor that has persisted for tens of thousands of years or more, combined with a lack of vegetation and pale country rock, has made the Nullarbor ideal for the prolonged preservation and easy recognition of meteorites. Essen¬ tially, meteorites have been accumulating in the Nullarbor since climatic conditions allowed for their preservation. Moreover, in the Nullarbor there is good evidence to suggest that meteorites are lying on, or near, the surfaces on which they fell and that, physiographically, the region has remained essentially undisturbed for at least the last 30,000 years (Benbow & Hayball 1992). Many of the early meteorite recoveries from the Nullarbor resulted from a programme of search and re¬ covery by personnel from the Kalgoorlie School of Mines (see Cleverly 1993 and references therein), al¬ though numerous recoveries have been made by rabbit trappers, notably the Carlisle family from Kalgoorlie (Bevan 1992a). Until recently, there were few meteorites known from the area of the Nullarbor in South Austra¬ lia. However, collecting by rabbiters and prospectors has resulted in a great number of new recoveries from the area that account for most of the large increase from the 50 meteorites reported by Bevan (1992a) to the current 123 known from South Australia. Unfortunately, most of this South Australian material now resides in collec¬ tions outside of Australia. Nomenclature The described meteorites from the Nullarbor Region now account for more than 50 % of all meteorites known from Australia. As meteorites are named after the geo¬ graphical localities where they are found, the general lack of geographical names in the Nullarbor, and the great number of new recoveries has caused difficulties for meteorite nomenclature. The problem has been over¬ come by the introduction of a system of meteorite no¬ menclature based on geographically named areas. Sev¬ enty-four named areas have been delineated (47 in West¬ ern Australia; 27 in South Australia) in the Nullarbor Region and new and distinct meteorites take the name of the area in which they are found and a three digit number (e.g. Cook 005), usually in chronological order of discovery (Bevan & Binns 1989a; Bevan & Bring 1993). Some nomenclatural anomalies have occurred and these include the Haig, Rawlinna (stone). Cook 003 and Maralinga meteorites, the localities of which lie outside the newly designated areas with the same names. 37 Journal of the Royal Society of Western Australia, 79(1), March 1996 Palaeoclimatic information from meteorites in the Nullarbor Region As soon as meteorites enter the Earth's atmosphere they are subject to contamination from, and alteration by, the terrestrial environment. Prolonged weathering transforms many of the minerals in meteorites, masks their original textures, redistributes elements, and even¬ tually destroys them. However, the processes of weath¬ ering leave a terrestrial 'fingerprint' in meteorite finds that may be used in climatic research. Meteorites that survive prolonged weathering are potential recorders of environmental conditions during their period of terres¬ trial residence. Although in its infancy, the use of an¬ cient Nullarbor meteorite finds as indicators of palaeoclimate is yielding promising results. Recent research on meteorites from the Nullarbor and other hot desert regions of the world for which terres¬ trial age data are available (Jull et al 1990; Jull et al. 1995) has suggested that the weathering characteristics of ancient meteorite finds may reflect the climatic condi¬ tions within a millennium or so of their fall (Bland et al. 1995a,b). This discovery has stimulated a completely new area of palaeoclimatic research, and the Nullarbor region is proving to be one of the most significant areas of the world for the use of meteorites as palaeoclimatic indicators. Within the limits of the data currently available (Jull et al. 1995), the distribution of ,4C terrestrial ages of ordi¬ nary chondritic meteorites from the Nullarbor Region show an apparently uninterrupted exponential decrease from the present day to around 30 ka BP (Fig 8). The oldest terrestrial age of a stony meteorite yet published from the Nullarbor (27±1.4 ka) is in good agreement with the estimated age (<30 ka) of the present calcareous clay cover of the Nullarbor (Benbow & Hayball, 1992). Figure 8. Distribution of 14C terrestrial ages of chondritic meteor¬ ite finds from the Nullarbor Region of Australia (after Jull et al. 1995). The preservation and accumulation of meteorites as the result of prolonged aridity in the Nullarbor from ca. 30 ka BP is consistent with palaeoclimatic evidence from a wide variety of geomorphological, palaeontological and biological studies of the region. For example, a number of workers, notably Thorne (1971), have sug¬ gested on the basis of faunal remains from caves that the climate of the Nullarbor has not changed significantly during the last 20 ka (e.g. see Wyrwoll 1979; Davey et al. 1992 and references therein). However, pollen from three caves in the Nullarbor examined by Martin (1973) indicated that the period 20 ka to ca. 10-8 ka BP was slightly more arid than that of today, with annual rain¬ fall averaging ca. 180 mm. The palynological evidence suggests that from around 10-8 ka BP to 5-4 ka BP the rainfall increased on the Nullarbor, and has since main¬ tained an annual average of about 250 mm. Further evi¬ dence from pollen from the dessicated guts of some dated mummified mammalian carcasses (Ingram 1969), supports the conclusion that the annual rainfall in the Nullarbor Plain area has remained roughly constant since about the middle Holocene (ca. 5 ka BP). Currently, the Nullarbor has no active surface drain¬ age. However, higher lake levels and relict stream courses traversing the region testify to a period of greater effective precipitation (Jennings 1967a,b, 1983; Lowry 1970; Lowry & Jennings 1974; Graaff et al 1977; Street-Perrott & Harrison 1984). When these channels were last active is unknown, although U-series dating of calcite speleothems from Nullarbor caves (Goede et al. 1990) suggests that no significant calcium carbonate deposition has taken place during the last 300-400 ka. Presently, active deposition of speleothems in Nullarbor caves is almost exclusively gypsum and halite (Goede et al. 1990; Goede et al. 1992). As a mechanism for halite deposition, Goede et al. (1992) suggest that periodic changes to slightly higher effective precipitation in the Nullarbor re-initiated percolation and, provided that the seepage was subject to strong evaporation in the cave atmosphere, led to the deposition of halite speleothems. U-series ages of halite speleothems from the Nullarbor have been reported by Goede et al. (1990, 1992). One large (2.78 metres long) broken salt stalag¬ mite from Webbs Cave gave a 'bulk' age indicating pro¬ longed deposition during the Late Pleistocene between ca. 37 ka and 20 ka BP (Goede et al, 1992). Previous dating (Goede et al, 1990) of a small (0.16 m) halite stalagmite from the same cave yielded an age of 2.5±1.2 ka indicating that there have been at least two phases of halite speliothem formation in Webbs Cave within the last 37 ka. The work of Goede et al. (1990, 1992) shows that during the Late Pleistocene (ca. 30-20 ka BP) and again in the Holocene (ca. 2.5 ka BP) there were minor changes to more humid conditions in the Nullarbor following periods of prolonged aridity, and support the conclu¬ sions of Martin (1973) and Lowry & Jennings (1974). Significantly, the age range of the oldest salt stalagmite (37-20 ka) from Webbs Cave overlaps with the apparent onset of accumulation of stony meteorites from around 30 ka BP on the Nullarbor surface. If any stony meteor¬ ites significantly older than 30 ka exist in the Nullarbor, it is possible that they are buried in the calcareous clay cover. However, it should be noted that 2hAl/*-'Mn dat¬ ing of the Mundrabilla iron meteorite by Aylmer et al. (1988) gave a terrestrial age >1 Ma, indicating that it is the oldest meteorite fall yet recovered from the Nullarbor. 38 Journal of the Royal Society of Western Australia, 79(1), March 1996 Bland et al (1995a, b) have attempted to quantify the state of weathering of a number of ordinary chondrite finds from the Nullarbor using 37Fe Mossbauer spectros¬ copy. By comparing the abundance of ferric iron oxide/ oxyhydroxide species in individual meteorites against terrestrial age, Bland et al (1995b) suggest that meteorite weathering is sensitive to climate at the time of fall. Moreover, meteorites appear to obtain their weathering characteristics within ca. 1000 years of fall. Gradual weathering rates, like those that have persisted in the Nullarbor, allow the formation of stable surface oxide layers, and a reduction in the porosity of the meteorites that provides protection against further significant weathering during periods of more effective precipita¬ tion. Moreover, carbonates derived by the meteorite from the Nullarbor limestone also fill pore space in some stones (Bevan & Binns 1989b). It appears that once a stony meteorite reaches a state of temporary equilibrium after initial weathering, the energy of the surrounding environment needs to be raised significantly to alter the remains further, or destroy them. Figure 9 shows a plot of total ferric oxidation (%) in Nullarbor H-group ordinary chondrites against terres¬ trial age (after Bland et al 1995b). Periods of climatic change derived from biological and geomorphological studies outlined above are marked for comparison. Even within the limits of the small data set currently available there is a remarkable co-incidence between the 'rustiness' of chondritic meteorites as measured by Mossbauer (Bland et al 1995b), and periods of alternately higher and lower effective precipitation in the Nullarbor dur¬ ing the Late Pleistocene and Holocene. Figure 9. Plot of total ferric oxidation (%) of weathered H-group ordinary chondrites from the Nullarbor Region of Australia as determined by Mossbauer against their I4C terrestrial ages. Marked above are significant palaeoclimatic events in SW Aus¬ tralia for the same period; see text for references (after Bland et al 1995b) Jull et al (1995) have made a preliminary study of the carbonates from some weathered ordinary chondrites from the Nullarbor. The results show that there are some variations in 8I3C, and there is a weak correlation of 813C and carbonate content with terrestrial age that may be linked to palaeoclimatic events. Recent recoveries of rare meteorites in Australia Bevan (1992a) listed a number of rare meteorites re¬ covered from Australia. The most significant of the ob¬ served falls (Table 1) are the two carbonaceous chon¬ drites Murchison [CM2] (CM= Mighei type carbon¬ aceous chondrite) and Karoonda [CK4] (CK= Karoonda type carbonaceous chondrite; see Bevan 1992a and refer¬ ences therein). Carlisle Lakes, a previously anomalous and ungrouped chondritic meteorite find from the Nullarbor (Binns & Pooley 1979) has recently been shown to belong to an entirely new group of chondrites including an observed fall, Rumuruti, from Kenya (Schulze et al 1994). The new group of chondrites, known as the 'R' group (after Rumuruti), includes sev¬ eral other meteorites from Antarctica and one from the Reg El Acfer in North Africa (Bischoff et al 1994; Rubin & Kallemeyn 1989, 1993, 1994; Schulze et al 1994). A new group of carbonaceous chondrites, the CR group (named after the type meteorite Renazzo), has been de¬ scribed (i e.g . see Weisberg et al 1993). Spettel et al (1992) have suggested that Loongana 001, an unusual chon¬ drite found in 1990 in the Western Australian Nullarbor is related to the CR group. However, Kallemeyn & Rubin (1995) have shown that the meteorite does not belong to any of the established carbonaceous chondrite groups and along with the Coolidge chondrite found in the USA, forms a distinct grouplet of carbonaceous chondrites related to the CV group (named after the type meteorite Vigarano). Bevan (1992a) noted that out of the main groups or sub-types of meteorites then known, only ten were not represented in collections from Australia. In the last three years, however, several ordinary chondrites of pet¬ rologic type 7 have been described providing examples previously missing (Wlotzka 1994), along with a possi¬ bly new petrologic type 2 (?) member of the CV group of carbonaceous chondrites, Mundrabilla 012 (Ulff- Moller et al 1993). One enstatite chondrite of type 7, Forrest 033, has been recorded (Wlotzka 1994). A new CV3 chondrite, Denman 002, has been described from the South Australian Nullarbor (Dominik & Bussy 1994). Figure 10. Mass (1.536 kg) of the Cook 003 CK4 chondrite found in the South Australian Nullarbor. The knobbly surface is due to large protruding chondrules (photograph by K Brimmell). 39 Journal of the Royal Society of Western Australia, 79(1), March 1996 Sleeper Camp 006 and Cook 003 (Fig 10), two stones found in the Western Australian and South Australian Nullarbor respectively, are additional CK4 chondrite finds. Cook 003 may be paired with an earlier reported discovery, Maralinga (Geiger et al 1992). Two other re¬ coveries from the Nullarbor, Camel Donga 003 and Watson 002 are the first known examples of CK3 chon¬ drites from Australia and their reported find-sites are sufficiently far apart to discount pairing (Wlotzka 1993a, b). In terms of achondrites, Calcalong Creek (Hill et al 1991; Wlotzka 1991), reportedly discovered within the strewn field of the previously known Millbillillie achon- drite (eucrite) in Western Australia, is the first lunar me¬ teorite found outside of Antarctica. Calcalong Creek is a polymict lunar breccia with the highest KREEP compo¬ nent of any known lunar meteorite (Hill et al. 1991). New achondrites from the Nullarbor include several howardites; Camel Donga 004, Hughes 004, Hughes 005, Old Homestead 001, Mundrabilla 018 (Wlotzka 1995) and Muckera 002. Old Homestead 001 and Mundrabilla 018 were found close together, as were Hughes 004 and 005 and these may be fragments of the same fall. The fragments constituting Muckera 002 were found near the site of discovery of the Muckera meteorite (now Muckera 001) which is also a howardite. Eagles Nest found in central Australia and Reid 013 from the Nullarbor, two new examples of brachinaites (olivine- rich achondrites named after the type meteorite found at Brachina in South Australia) have been found (Wlotzka 1992b,1993b), Reid 013 is remarkably similar to Nova 003, a meteorite the locality of which is uncertain (Wlotzka 1993b). Three new ureilites, Hughes 007, Hughes 009 and Nullarbor 010, have been found in the Nullarbor and the latter meteorite is similar to Nova 001, an ureilite originally reported as found in Mexico although now considered to be of uncertain location (Wlotzka 1993a). A new lodranite (an anomalous stony-iron meteorite of rare type possibly related to the ureilite achondrites), named Gibson, has been found in north-western Austra¬ lia and is the first meteorite of its kind found in Austra¬ lia (Wlotzka 1992b). Since 1992, two important new masses of iron meteorites have been reported. Two dis¬ tinct group HE irons, Watson (93 kg) (Olsen et al 1994) and Miles (265 kg), have been reported from the South Australian Nullarbor and Queensland, respectively (Wlotzka 1992a, 1994). A new iron. Hidden Valley weighing 7 kilograms, belonging to chemical group IIIAB was also found in Queensland in 1991 (Wlotzka 1994). These discoveries bring the total number of dis¬ tinct irons reported from Australia to 71. Meteorite groups and types remaining to be reported from Australia include chondrites belonging to EH3-4, EL5, Cl, CK5, C03, and a possibly new 'CH' group of carbonaceous chondrites (Bischoff et al 1993), and achondrites belonging to the calcium-poor groups, aubrites and diogenites. Summary Over the last ten years, the number of meteorites known from Australia has doubled. Even so, the large number of recent recoveries probably represents a small fraction of the meteorites that are available for collection in the country's arid and semi-arid zones. Together with meteorites from other hot and cold deserts of the world, the concentration of meteorites in the Nullarbor (Fig 11) is already providing a valuable research resource. New groups of meteorites are being recovered that are ex¬ tending our knowledge of the early Solar System; statis¬ tical studies are providing information on the flux of meteorites with time; and terrestrial age dating com¬ bined with weathering studies are yielding palaeoclimatic information about the areas of meteorite accumulation. Figure 11. Crusted fragment of the Camel Donga eucrite achon- drite shower at the site of discovery on the Nullarbor Plain (lens cap for scale is 5 cm diameter). Although various dating techniques have been ap¬ plied to a wide variety of terrestrial materials ( e.g . see Lamb 1977), one of the major problems of Quaternary palaeoclimatic research is the paucity of dateable materi¬ als that can provide an absolute chronology for events. Concentrations of meteorites, such as in the Nullarbor Region, provide dateable materials of a variety of terres¬ trial exposure times spanning the accumulation period, and may allow changes in weathering rates through time to be estimated. The pioneering work of Bland et al (1995 a,b), Jull et al. (1995) and others is demonstrating that weathered stony meteorites have the potential to provide useful palaeoclimatic information over the pe¬ riod of their accumulation. The surfaces of ancient stony meteorites from the Nullarbor frequently possess variably thick carbonate coatings (caliches) that have been derived via the calcar¬ eous clay from the limestone country rock (Bevan & Binns 1989 a,b). Additionally, extensively weathered Nullarbor stony meteorites contain veins and pockets of carbonate that have penetrated the fabric of the meteor¬ ite along cracks and pores. Detailed mineralogical and isotopic studies of carbonates from Nullarbor meteorites have yet to be performed. However, measurements of the stable isotopic compositions of these evaporitic de¬ posits may aid in determining the source materials and the mechanisms of calichification, related to tempera¬ tures of deposition. Since these carbonates are likely to have grown much more rapidly than marine carbonates they could be used to derive high resolution tempera¬ ture profiles, and offer possibilities of palaeotemperature 40 Journal of the Royal Society of Western Australia, 79(1), March 1996 assessment over the time span of meteorite accumula¬ tion in the Nullarbor. Meteorites found in Australia are playing an increas¬ ingly important role in fundamental research across a wide spectrum of disciplines both within the country and overseas. There is every reason to believe that dense accumulations of meteorites, as in the Nullarbor, exist throughout the arid zone of Australia. Eventually, Aus¬ tralia may outstrip Antarctica and the USA as a source of meteorite recoveries. This seemingly unlikely source of Quaternary palaeoclimatic information provides yet another aspect to meteorite research in Australia. What studies of ancient meteorite finds from Australia are likely to reveal in detail about past climates is unknown. Like all fundamental scientific research, one never knows how useful it will be until it is done! Acknowledgements: The author thanks an anonymous reviewer for many helpful comments that improved the manuscript. Danielle Hendricks, Pe¬ ter Downes, Jennifer Bevan and Kris Brimmell are thanked for their assis¬ tance in the preparation of the manuscript. John de Laeter is thanked for constant encouragement and help over many years of research into Aus¬ tralian meteorites. References Anderson C 1913 A catalogue and bibliography of Australian Meteorites. Records of the Australian Museum 10:53-76. Aylmer D, Bonnano V, Herzog G F, Weber H, Klein J & Middleton R 1988 A1 and Be production in iron meteorites. Earth Planetary Science Letters 88:107. Barber D J & Hutchison R 1991 The Bencubbin stony-iron mete¬ orite breccia: Electron petrography, shock history and affini¬ ties of a "carbonaceous chondrite" clast. Meteoritics 26:83-95. Benbow M C & Hayball A J 1992 Geological observations of Old Homestead Cave, Nullarbor Plain, Western Australia. Aus¬ tralian Caver 130:3-6. Bevan A W R 1992a Australian meteorites. Records of the Aus¬ tralian Museum Supplement. 15:1-27. Bevan A W R 1992b 1992 WAMET/EUROMET joint expedition to search for meteorites in the Nullarbor Region, Western Australia. Meteoritics 27:202-203. Bevan A W R 1996 Australian crater-forming meteorites. AGSO Journal of Australian Geology & Geophysics 16:421-429. Bevan A W R & Binns R A 1989a Meteorites from the Nullarbor Region, Western Australia: I. A review of past recoveries and a procedure for naming new finds. Meteoritics 24:127-133* Bevan A W R & Binns R A 1989b Meteorites from the Nullarbor Region, Western Australia: II. Recovery and classification of 34 new meteorite finds from the Mundrabilla, Forrest, Reid and Deakin areas. Meteoritics 24:134-141. Bevan A W R & Binns R A 1989c Further meteorite recoveries from the Nullarbor Region, Western Australia. Meteoritics 24:251. Bevan A W R & Griffin B J 1994 Re-examination of the Murchison Downs meteorite: A fragment of the Dalgaranga mesosiderite? Journal of the Royal Society of Western Australia 77:45-49, Bevan A W R, McNamara K J & Barton J C 1988 The Binningup H5 chondrite: A new fall from Western Australia. Meteoritics 23:29-33. Bevan A W R & Pring A 1993 Guidelines for the naming of new meteorite finds from the Nullarbor Region, South Australia. Meteoritics 28:600-602. Binns R A & Pooley G D 1979 Carlisle Lakes (a): a unique oxi¬ dized chondrite. Meteoritics, 14:349-350. Bischoff A, Palme H, Schultz L, Weber D, Weber H, & Spettel B 1993 Acfer 182 and paired samples, an iron-rich carbonaceous chondrite: Similarities with ALH85085 and relationship to CR chondrites. Geochimica et Cosmochimica Acta 57:2631-2648. Bischoff A, Geiger T, Palme H, Spettel B, Schultz L, Scherer P, Loeken T, Bland P, Clayton R N, Mayeda T K, Herpers U, Meltzow B, Michel R & Dittrich-Hannen B 1994 Acfer 217- A new member of the Rumuruti chondrite group (R). Meteorit¬ ics 29:264-274. Bland P A, Berry F J & Pillinger C T 1995a Meteorites as indica¬ tors of palaeoclimatic conditions. Lunar and Planetary Sci¬ ence 26:129-130. Bland P A, Berry F J & Pillinger C T 1995b. Weathering of ordi¬ nary chondrites from Algeria and Australia as a climatic indi¬ cator. Meteoritics 30:487. Cleverly W H 1993 The role of the Western Australian School of Mines in meteorite recovery and research. Western Australian School of Mines Annual Magazine 90 years, Kalgoorlie: pp 59-65. Cooksey T 1897 The Nocoleche meteorite with a catalogue and bibliography of Australian meteorites. Records of the Austra¬ lian Museum 3:51-62. Davey A G, Gray M R, Grimes K G, Hamilton-Smith E, James J M & Spate A P 1992 World Heritage significance of karst and other landforms in the Nullarbor region. Report to the Com¬ monwealth Department of the Arts, Sport, The Environment & Territories, Canberra. 202 pp. De Laeter J R 1972 The Mundrabilla meteorite shower. Meteoritics 7:285-294. De Laeter J R & Bevan A W R 1992 A history of Western Austra¬ lian meteoritics. Records of the Western Australian Museum 15:749-762. De Laeter J R & Cleverly W H 1983 Further finds from the Mundrabilla meteorite shower. Meteoritics 18:29-34. Dodd R T 1981 Meteorites - A Petrologic-Chemical Synthesis. Cambridge University Press, Cambridge. Dominik B & Bussy F 1994 Denman 002: A new CV3 chondrite from South Australia. Meteoritics 29:409-410. Fudali R F & Ford R 1979 Darwin glass and Darwin Crater: a progress report. Meteoritics 14:283-296. Geiger T, Bischoff A, Spettel B & Bevan A W R 1992 Cook 003: A new CK chondrite from the Nullarbor Region, South Austra¬ lia. Lunar and Planetary Science. 23:401. Gibbons GS 1977 Index of Australian Meteorites. Journal of the Geological Society of Australia 24:263-268. Gladman B J, Bums J A, Duncan M, Lee P & Levison H F 1996 The exchange of impact ejecta between terrestrial planets. Science 271:1387-1392. Goede A, Atkinson T C & Rowe P J 1992 A giant Late Pleistocene halite speleothem from Webbs Cave, Nullarbor Plain, south¬ eastern Western Australia. Helictite 30:3-7. Goede A, Flarmon R S, Atkinson T C & Rowe P J 1990 Pleistocene climatic change in southern Australia and its effects on speleothem deposition in some Nullarbor caves. Journal of Quaternary Science 5:29-38. Graaff W J E van de, Crowe R W A, Bunting J A & Jackson M J 1977 Relict early Cainozoic drainages in arid Western Austra¬ lia. Zeitschrift Geomorphology 21:379-400. Graham A L, Bevan A W R & Hutchison R 1985 Catalogue of Meteorites, with special reference to those represented in the collection of the British Museum (Natural History), 4rh ed. British Museum (Natural History), London & University of Arizona Press, Tucson. Grossman J 1996 Meteoritical Bulletin no. 80. Meteoritics 31 (in press). Haidinger W K von 1861 Zwei Meteoreisenmassen in der Nahe von Melbourne in Australien aufgefunden. Sitzungsberichte de Akademie der Wissenschaften, Wien, 43:583-584. Hey M H 1966 Catalogue of Meteorites, with special reference to those represented in the collection of the British Museum (Natural History), 3rd ed., British Museum (Natural History), London. 41 Journal of the Royal Society of Western Australia, 79(1), March 1996 Hill D H, Boynton W V & Haag R A 1991 A lunar meteorite found outside the Antarctic. Nature 352:614-617. Hodge-Smith T 1939 Australian meteorites. Memoirs of the Aus¬ tralian Museum 7:1-84. Ingram B S 1969 Sporomorphs from the desiccated carcases of mammals from Thylacine Hole, Western Australia. Helictite 7:62-66. Jennings J N 1967a Some karst areas of Australia. In: Landform Studies from Australia and New Guinea (eds. J N Jennings & J A Mabbutt). ANU Press, Canberra, 256-292. Jennings J N 1967b The surface and underground geomorphol¬ ogy. In: Caves of the Nullarbor (eds. J R Dunkley & T M L Wigley). The Speleological Research Council, University of Sydney, 13-31. Jennings J N 1983 The disregarded karst of the arid and semi- arid domain. Karstalogia 1:61-73. Jull A J T, Bevan A W R, Cielaszyk E & Donahue D J 1995 14C terrestrial ages and weathering of meteorites from the Nullarbor Region, Western Australia. Lunar and Planetary In¬ stitute Technical Report 95-02:37-38. Jull A J T, Wlotzka F, Palme H & Donahue D J 1990 Distribution of terrestrial age and petrologic type of meteorites from west¬ ern Libya. Geochimica et Cosmochimica Acta 54:2895-2898. Kallemeyn G W & Rubin A E 1995 Coolidge and Loongana 001: A new carbonaceous chondrite grouplet. Meteoritics 30:20-27. Koeberl C, Delisle G & Bevan A 1992 Meteorite aus der Wiiste. Die Geowissenschaften 10:220-225. Lamb H H 1977 Climate: Present, Past and Future. Vol 2. Methuen & Co, London. Lowry D C 1970 Geology of the Western Australian part of the Eucla Basin. Geological Survey of Western Australia Bulletin, Perth. 122. Lowry D C & Jennings J N 1974. The Nullarbor karst, Australia. Zeitschrift Geomorphologie 18:35-81. Martin H A 1973 Palynology and historical ecology of some cave excavations in the Australian Nullarbor. Australian Journal of Botany 21:283-316. Mason B 1974 Notes on Australian meteorites. Records of the Australian Museum 29:169-186. McNaught R H 1993 Fire from the sky. Sky and Space 6:25-38. McSween H Y 1987 Meteorites and their Parent Bodies. Cam¬ bridge University Press, Cambridge. Olsen E, Davis A, Clarke R S Jr., Schultz L, Weber H W, Clayton R, Mayeda T, Jarosewich E, Sylvester P, Grossman L, Wang M-S, Lipschutz M E, Steele I M & Schwade J 1994 Watson: A new link in the HE iron chain. Meteoritics 29:200-213. Prior G T 1923 Catalogue of Meteorites, with special reference to those represented in the collection of the British Museum (Natural History), British Museum, London. Prior GT & Hey M H 1953 Catalogue of Meteorites, with special reference to those represented in the collection of the British Museum (Natural History), second edition, British Museum (Natural History), London. Rubin A E & Kallemeyn G W 1989 Carlisle Lakes and Allan Hills 85151: Members of a new chondrite grouplet. Geochimica et Cosmochimica Acta 53:3035-3044. Rubin A E & Kallemeyn G W 1993 Carlisle Lakes chondrites: Relationship to other chondrite groups. Meteoritics 28:424- 425. Rubin A E & Kallemeyn G W 1994 Pecora Escarpment 91002: A member of the new Rumuruti (R) chondrite group. Meteorit¬ ics 29:255-264. Schulze H, Bischoff A, Palme H, Spettel B, Dreibus G & Otto J 1994 Mineralogy and chemistry of Rumuruti: The first mete¬ orite fall of the new R chondrite group. Meteoritics 29:275- 286. Shoemaker E M & Shoemaker C S 1988 Impact structures of Aus¬ tralia. Lunar and Planetary Science Conference 19:1079-1080. Shoemaker E M & Shoemaker C S 1996 The Proterozoic impact record of Australia. AGSO Journal of Australian Geology & Geophysics: in press. Spettel B, Palme H, Wlotzka F & Bischoff A 1992 Chemical com¬ position of carbonaceous chondrites from Sahara and Nullarbor Plains. Meteoritics 27:290-291. Street-Perrott F A & Harrison S P 1984 Temporal variations in lake levels since 30000 yrs BP - an index of the global hydro- logical cycle. In: Climate Processes and Climate Sensitivity (eds J E Hansen & T Takahashi). American Geophysical Union Geophysical Monograph 29:118-129. Thome A G 1971 The fauna. In: Archaeology of the Gallus Site, Koonalda Cave (ed R V S Wright). Australian Aboriginal Studies No. 26 Australian Institute of Aboriginal Studies, Canberra, 45-47. Ulff-Moller F, Rasmussen K L & Grundvig S 1993 Mundrabilla 012: A new CV2(?) chondrite find from SW Australia. Meteor¬ itics 28:451-452. Weisberg M K, Prinz M, Clayton R N & Mayeda T K 1993 The CR (Renazzo-type) carbonaceous chondrite group and its impli¬ cations. Geochimica et Cosmochimica Acta 57:1567-1586. Weisberg M K, Prinz M, Clayton R N, Mayeda T K, Grady M M & Pillinger C T 1995 The CR chondrite clan. Proceedings of the National Institute of Polar Research Symposium on Ant¬ arctic Meteorites 8:11-32. Wilson R B & Cooney A M 1967 The Mundrabilla meteorite: A new discovery in Western Australia. Nature 213: 274-275. Wlotzka F (Ed) 1991 Meteoritical Bulletin no 71. Meteoritics 24: 255-262. Wlotzka F (Ed) 1992a Meteoritcal Bulletin no 72. Meteoritics 27: 109-116. Wlotzka F (Ed), 1992b Meteoritical Bulletin no 73. Meteoritics 27: 477-484. Wlotzka F (Ed) 1993a Meteoritical Bulletin no 74. Meteoritics 28: 146-153. Wlotzka F (Ed) 1993b Meteoritical Bulletin no 75. Meteoritics 28: 692-703. Wlotzka F (Ed) 1994 Meteoritical Bulletin no 77. Meteoritics 29: 891-897. Wlotzka F (Ed) 1995 Meteoritical Bulletin no 78. Meteoritics 30: 792-796. Wyrwoll K-H 1979 Late Quaternary climates of Western Austra¬ lia: evidence and mechanisms. Journal of the Royal Society of Western Australia 62:129-142. 42 Journal of the Royal Society of Western Australia, 79:43-50, 1996 Isotopic anomalies in extraterrestrial grains T R Ireland Research School of Earth Sciences, Australian National University, Canberra ACT 0200 Abstract Isotopic compositions are referred to as anomalous if the isotopic ratios measured cannot be related to the terrestrial (solar) composition of a given element. While small effects close to the resolution of mass spectrometric techniques can have ambiguous origins, the discovery of large isotopic anomalies in inclusions and grains from primitive meteorites suggests that material from distinct sites of stellar nucleosynthesis has been preserved. Refractory inclusions, which are predominantly composed of the refractory oxides of Al, Ca, Ti, and Mg, in chondritic meteorites commonly have excesses in the heaviest isotopes of Ca, Ti, and Cr which are inferred to have been produced in a supernova. Refractory inclusions also contain excess 2flMg from short lived 2hA\ decay. However, despite the isotopic anomalies indicating the preservation of distinct nucleo- svnthetic sites, refractory inclusions have been processed in the solar system and are not interstel¬ lar grains. Carbon (graphite and diamond) and silicon carbide grains from the same meteorites also have large isotopic anomalies but these phases are not stable in the oxidized solar nebula which suggests that they are presolar and formed in the circumstellar atmospheres of carbon-rich stars. Diamond has a characteristic signature enriched in the lightest and heaviest isotopes of Xe, and graphite shows a wide range in C isotopic compositions. SiC commonly has C and N isotopic signatures which are characteristic of H-burning in the C-N-O cycle in low-mass stars. Heavier elements such as Si, Ti, Xe, Ba, and Nd, carry an isotopic signature of the s-process. A minor population of SiC (known as Grains X, ca. 1 %) are distinct in having decay products of short lived isotopes 26A1 (now 2hMg), -“Ti (now -“Ca), and 49V (now 49Ti), as well as 2*Si excesses which are characteristic of supernova nucleosynthesis. The preservation of these isotopic anoma¬ lies allows the examination of detailed nucleosynthetic pathways in stars. Introduction The study of isotopic anomalies in meteorites offers a direct image of nucleosynthesis in stars and the pro¬ cesses by which dust is dispersed into the interstellar medium. The solar system is composed of a mixture of a variety of nucleosynthetic sources which were homog¬ enized during the solar nebula and further during plan¬ etary formation. Isotopic abundances on Earth are af¬ fected only by radioactive decay and by the relatively small fractionations caused by the differences in the masses of the isotopes during kinetic processes. The largest fractionations are found in H (which has the larg¬ est mass difference between two isotopes) and fraction¬ ation effects become smaller with increasing mass. However, in different nucleosynthetic environments in stars, there are large (orders of magnitude) variations in isotopic production rates. The preservation of material from these distinct environments results in grains with diverse isotopic compositions. The measurement of isotopic compositions generally involves the measurement of a ratio of one isotope to another in a mass spectrometer. For an element with more than three isotopes, an isotopic composition is said to be anomalous if it cannot be related to the terrestrial (and by inference solar) composition through a mass fractionation law that describes the behaviour of the isotopes in physicochemical processes including the © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 measurement process itself. A mass fractionation correc¬ tion can be achieved by using one isotopic ratio to deter¬ mine the fractionation and then removing that degree of fractionation from the other isotopic ratio(s). A frac¬ tional deviation of a fractionation-corrected ratio is often reported such that a deviation of zero is normal, positive indicates the ratio exceeds normal and negative indi¬ cates it is below normal. For an element with only two isotopes, a mass fractionation correction is not possible and the isotopic ratio is said to be anomalous if it ex¬ ceeds the range of that ratio typically found on Earth. A wide variety of mass spectrometers have been used in analyzing isotopic compositions of meteoritic phases. Gas and solid-source mass spectrometers require mutigrain samples to achieve sufficient signal for a pre¬ cise analysis but have clearly characterized various min- eralogical grain types (diamond, silicon carbide, graph¬ ite) as highly anomalous. But perhaps the most telling information is coming from ion microprobe analysis which allows the measurement of single grains (to sub¬ micron sizes) for a number of elements and isotopic sys¬ tems which can then be related to distinct astrophysical settings. This paper gives a brief introduction to nucleosynthe¬ sis in stars, gives an overview of some of the isotopically anomalous extraterrestrial grains found so far, and re¬ lates these to some distinct astrophysical settings which are possible sites for grain formation. It is not intended to be comprehensive review of all knowledge of the iso¬ topic systematics of extraterrestrial grains but has been written to give a more general outline of the field along 43 Journal of the Royal Society of Western Australia, 79(1), March 1996 with some illustrative examples. More detailed reviews and information can be found in Lee (1988) and Clayton et al. (1988) for isotopic systematics of refractory inclu¬ sions, and Anders & Zinner (1993) and Ott (1993) for interstellar grains. The role of ion microprobe mass spectrometry in these fields has been reviewed by Ire¬ land (1995). Nucleosynthesis Apart from 'H, 2H 3He, 4He and 7Li which were formed in Big Bang nucleosynthesis, all other nuclides were produced, and are being produced, in nuclear re¬ actions predominantly in stars, but also in circumstellar environments and the interstellar medium. In light of the abundance curve of the nuclides (Fig 1), Burbidge et al. (1957) specified 8 modes of nucleosynthesis that could account for its features. The two fundamental thermo¬ nuclear reactions driving stars are H-buming and He-buming. Upon exhaustion of the H, the star contracts under its own gravity and if the star has sufficient mass. He will ignite and burn to form C and O. Reactions involving the successive addition of a particles (a -process) to nuclei heavier than 20Ne result in four-structure nuclei (24Mg, 28St 32S, 36 Ar, ^Ca, and possibly ^Ca) at higher abun¬ dances than their neighbours. For a massive star, nucleosynthesis can proceed up to the mass region of the iron group of elements where the nuclear binding energy is at a maximum per nucleon and no further energy can be released by fusion. At this stage, nuclear statistical equilibrium (e-process) is achieved in the reac¬ tions and leads to a build-up of Fe-group-element abun¬ dances. The low-abundance nuclides 9Be, and 10B and 11 B are unstable in stellar interiors and require for¬ mation in low temperature environments. The forma¬ tion of these elements was assigned to the x-process (x for unknown) which is probably largely due to spalla¬ tion. Burbidge et al. (1957) also used the x-process to explain the abundance of D, but this is now assigned to the Big Bang. Elements heavier than the Fe group are mainly pro¬ duced in neutron-capture reactions. In the s-process, the addition of neutrons is slow relative to the likelihood of a P decay, whereas in the r-process the addition of neu¬ trons happens on a rapid time-scale and very neutron- rich nuclei can result. The s-process occurs in asymp¬ totic giant branch (AGB) stars undergoing core He burn¬ ing where neutrons are released through reactions such as 13C(a,n)lf,0 or 22Ne(a,n)25Mg. The r-process requires a large neutron flux and it is likely to be related to super¬ novae. Low abundance nuclei on the low mass side of the valley of stability ( i.e . neutron-poor isotopes) are re¬ ferred to as p-process nuclides and these have probably formed by photonuclear reactions, also in supernovae. A supernova of Type II occurs when a massive star has consumed most of its nuclear fuel. At this stage it has an Fe-rich core and is surrounded by layers still burning the remnants of its fuel. Upon near exhaustion of this fuel, the star can no longer support itself gravita¬ tionally and it implodes. The rebound shock wave ejects the outer layers while the material inside the mass cut collapses into a degenerate neutron star or pulsar. No¬ vae and supernovae of Type I involve binary stars with a white dwarf left after the AGB phase. A nova can occur if the white dwarf accretes more than 10'3 solar masses of H from its companion resulting in explosive H burning. If the accreting white dwarf eventually ex¬ ceeds the Chandrasekhar limit (1.4 Msolar) then the star collapses and explodes as a Type la supernova. The evolution and ultimate state of a star is governed primarily by two parameters, metallicity and mass. The metallicity of a star is the proportion of elements heavier than H and He and it affects nucleosynthesis in that heavier elements can act as seeds for more energetic re¬ actions. The mass of the star limits the ultimate tempera¬ ture and pressure in the stellar core. For a star less than Figure 1. The solar abundances are an average of a number of different nucleosynthetic sites. Burbidge et al (1957) first de¬ scribed the features of the nuclidic abundance curve in terms of eight discrete stellar processes and with only a little refinement the principal tenants are still applicable. H-Buming and He-burn- ing are the fundamental thermonuclear reactions driving stars converting H to He and He to C and O. The relatively high abundances of nuclei in the mass range 20 - 40 with mass num¬ bers divisible by 4 (i.e. 2r'Ne, :4Mg, 2SSi, 32S, 3* Ar, 4"Ca) are due to progressive addition of a-particles (< x-process ). Stars more mas¬ sive than 9 Msolar can ignite the C,0-rich core left after He burn¬ ing and for the most massive stars nucleosynthesis can proceed to equilibrium burning (e-process) at the iron abundance peak. The s- and r-processes refer to progressive addition of neutrons at slow (s-process) and rapid ( r-process ) rates with respect to the competing p decays. Note the peaks in abundance of s- and r- process nuclei around magic number nuclei (N=50, 82, and 126). Low-abundance neutron-poor nuclei are likely due to photo- nuclear reactions ( p-process ). Finally the abundances of the light, low-abundance elements Li, Be, B as well as D, were attributed to an unknown process (x-process). Deuterium is now attributed to Big Bang production and the abundances of Li, Be, and B are likely from spallation of heavier nuclei. 44 Journal of the Royal Society of Western Australia, 79(1), March 1996 9 M^|ar, no nucleosynthesis beyond He-burning occurs and the star decays into a white dwarf. Heavier stars can ignite the C-O core and reactions proceed to build heavier nuclei (up to ^Fe) with the heaviest stars eventu¬ ally exploding in Type II supemovae. In the chemical evolution of the galaxy, material is ejected from a dying star to become available for the next generation. Super¬ novae are extremely efficient at getting a large mass of matter into the interstellar medium, but these events are uncommon since they only occur in the heaviest stars (1 event per 50 years per galaxy producing 2 M^ilar = 0.04 Msoiaryr1)- Most material ejected into the interstellar me¬ dium comes from the winds of AGB stars and ensuing planetary nebula (1 yr1 producing 0.3 Mgolar = 0.3 M^|aryr*’). Novae are more frequent but produce far less material (30 yr'1 producing 10"1 Msolar = 0.003 M^yr1; Truran 1993). Winds carry material away from the star surface into the circumstellar environment where it condenses into grains, and these grains continue to move away from the star and may be entrained into molecular clouds to form the next generation of stars. In this way, our solar system was formed in a molecular cloud that had contri¬ butions from a large number of stars of different genera¬ tions. Isotopic anomalies Isotopic anomalies were first discovered for the noble gases Ne and Xe from chondritic meteorites (Reynolds & Turner 1964; Black & Pepin 1969). However, the sys- tematics of the noble gases were poorly understood and an origin from within the solar system from radiogenic decay of transuranic nuclides and fission could not be dismissed. Furthermore, in the 1960s the current para¬ digm on the evolution of the solar nebula was that it attained sufficiently high temperatures to volatilize most if not all of the interstellar dust falling into it and there¬ fore all isotopic heterogeneity had been removed (Cameron 1962). The discovery of isotopic variations in O, the most abundant lithophile element in the solar system quickly changed the situation (Clayton et al 1973). Refractory Inclusions The oxygen anomalies, excesses in lftO of up to 5 %, were discovered in refractory inclusions from the Allende carbonaceous chondrite. These inclusions are composed of refractory oxides that would condense from a cooling gas of solar composition, and the Allende inclusions commonly include the minerals spinel MgAl204, melilite Ca,Al2Si07 - Ca2MgSi207, pyroxene CaMgSi2Ofi, and perovskite CaTi03. Refractory trace ele¬ ments are enriched in these inclusions at approximately 20 times that found in bulk carbonaceous chondrites in¬ dicating that the refractory inclusions could have con¬ densed as the first 5 % of solar material. These character¬ istics suggested that refractory inclusions would be the best place for a search for isotopic anomalies since these early-formed solids may have crystallized before all iso¬ topic heterogeneities were removed from the solar nebula. With the discovery of oxygen isotopic anomalies, the race was on to find further isotopic effects. Magnesium isotopic abundances were found to give excesses and deficits in 2ftMg after correcting for mass-dependent frac¬ tionation (Gray & Compston 1974; Lee & Papanastassiou 1974), but large excesses of 26Mg which were correlated with Al/Mg in a number of phases from a single refrac¬ tory inclusion established the presence of short lived 26A1 in the early solar system (Lee et al 1977). Excess 26Mg (i.e. extinct 26A1) is commonly present in refractory in¬ clusions at a maximum level of 5 10'5 27 Al. During the final stages of hydrostatic burning in a massive star, the Fe-group elements form under quasi¬ equilibrium conditions since binding energy per nucleon is at a maximum. Howrever, during the supernova the isotopic compositions of the Fe group elements could be affected by large changes in the physical conditions of nucleosynthesis. Furthermore, titanium, on the wing of the Fe abundance peak, is highly refractory and concen¬ trated in the inclusions. Titanium isotopic anomalies were first discovered as a 0.1 %, or 1 %o, enhancement in “Ti, a very small effect barely resolvable at that time (Heydegger et al 1979). Soon, however, titanium isoto¬ pic anomalies became regarded as being ubiquitous in refractory inclusions (Niemeyer & Lugmair, 1981). Fur¬ ther developments in mass spectrometry revealed effects in other Fe-group elements including excesses of ^Ca (Jungck et al 1984), and MCr (Birck & Allegre, 1984). The common feature of these anomalies is that they oc¬ cur in the heaviest nuclides of each element and hence suggest a common nucleosynthetic component is re¬ sponsible. These excesses in ^Ca, “Ti, and ^Cr were successfully modelled as the products of explosive nucleosynthesis around a supernova whereby reactions in the ejecta resulted in neutron-rich compositions (Hartmann et al 1985). Coincidentally, a supernova had earlier been proposed not only as the source for lhO and 26A1 but also as a trigger for the collapse of the molecu¬ lar cloud into the solar system (Cameron & Truran 1977). Refractory inclusions are also found in other meteor¬ ites besides Allende, and in particular the refractory in¬ clusions of CM meteorites such as Murchison and Murray contain corundum A1203 and hibonite CaAl120]9. The presence of these minerals with even higher con¬ densation temperatures than the assemblages of the Allende inclusions suggested that these might contain even larger isotopic anomalies. However, the CM inclu¬ sions were very small, seldom over 0.2 mm. The mass spectrometry of the Allende inclusions involved chemi¬ cal separation of the elements of interest and then load¬ ing the separate onto a filament which was heated to a temperature sufficient to ionize the element. With such small samples from Murchison, the prospect of conven¬ tional thermal ionization mass spectrometry wras not good. However, ion microprobes were beginning to be utilized in the search for isotopic anomalies and the small CM inclusions were an ideal place to start. The ion microprobe utilizes a focussed primary beam of O or Cs+ ions to sputter away the sample. Sputtering with an energetic ion beam results in a small fraction of the ejected atoms being ionized and these can be extracted and passed to a mass analyzer. The benefit of ion probe analysis is that a single inclusion can be analyzed for a 45 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 2. Ca and Ti isotopic compositions in hibonite (Ca[AlMgTi]12Oiy) inclusions display a large range from excesses in the most neutron rich isotopes to deficits while the other isotopes are close to solar proportions (defined as deviations, 8'Ca or 8'Ti, in %o). The most anomalous grains are both from the Murchison meteorite, 13-13 (Ireland 1990) and BB-5 (Hinton et al. 1987) which have large excesses and large depletions respectively in 48Ca and ^'Ti (insets). The 48Ca and ^Ti anomalies are clearly correlated and are inferred to come from neutron-rich supernova ejecta. variety of elements while maintaining the petrographic context. Titanium isotopic analyses of hibonite-bearing inclu¬ sions quickly established the highly anomalous charac¬ teristics of these grains (Fahey et ah 1985; Ireland et ah 1985; Hinton et al. 1987) with anomalies more than an order of magnitude larger than those found in Allende. Both excesses and deficits were discovered in ^Ti, and the excesses or deficits in *°Ti were correlated with ^Ca (Fig 2; Zinner et al 1986; Ireland 1990). While the excesses in these isotopes are consistent with the addi¬ tion of neutron-rich material, deficits are not so readily interpretable. Hinton et al. (1987) proposed that the deficits were due to the solar system being initially de¬ pleted in these neutron-rich isotopes and that the solar system received a late injection of supernova ejecta which included lflO and 2*A1. However, oxygen isotopic analyses revealed that all hibonite inclusions were enriched in lhO relative to terrestrial by 4 to 7 % (Fahey et al. 1987; Ireland et al 1992) precluding a scenario that includes a solar system initially depleted in ,hO. Magne¬ sium isotopic compositions of the hibonite grains were split between inclusions that have 2hAl/27Al at the ca¬ nonical solar system value of 5 10'5 and inclusions that were below 10~5. Surprisingly, a number of morphological, chemical, and isotopic features are found to be correlated in Murchison refractory inclusions. Most of the hibonite inclusions can be divided into two types, single hibonite crystal fragments with low Ti02 concentrations (<2.5 wt%), and hibonite-spinel inclusions with high Ti02 concentrations (3 - 9 wt%). Most of the crystal frag¬ ments show smooth REE patterns with depletions in the more volatile REE Eu and Yb while the spinel hibonite inclusions most commonly have ultrarefractory-REE-de- pleted patterns. Besides the clear correlation in ^Ca and ^Ti anomalies, it has been noted (Clayton et al 1988) that no inclusion with a large “Ti anomaly also has ex¬ cess 2*Mg (at 5 105), and those inclusions that have anomalies in the ultrarefractory elements of their REE patterns do have excess 26Mg unless they have a Ti isoto¬ pic anomaly (Ireland 1990). These correlations and asso¬ ciations clearly indicate that nucleosynthetic anomalies have survived through discrete inclusion-forming events in the early solar system. But, despite containing isotopically anomalous mate¬ rial, refractory inclusions are not interstellar grains. They are too large and do not preserve the effects ex¬ pected of solids exposed to travel in the interstellar me¬ dium. Presolar carbon and carbides The progressive isolation of isotopically distinct noble gas components finally led to the isolation of presolar grains in the laboratory (Fig 3). The main processing involves the acid dissolution of solar system material with further concentration based on other physical prop¬ erties such as density. It is therefore no surprise that the first interstellar grains identified are all carbon com- 46 Journal of the Royal Society of Western Australia, 79(1), March 1996 pounds that are resistant to acid attack. These com¬ pounds are inferred to be of presolar origin because car¬ bon compounds such as diamond, graphite, and SiC are unstable in the solar nebula and the grains contain mate¬ rial that is isotopically anomalous in the extreme. Xe isotope Figure 3. Progressive enrichment of these isotopically exotic Xe components led to the discovery of interstellar diamond and SiC. Interstellar graphite was discovered through isolation of a nearly pure component of ^Ne. The light and heavy components of Xe- HL cannot be produced in the same nucleosynthetic event and are probably the result of mixing (Huss & Lewis 1994). The Xe-S component from SiC reflects a mixture between the composition produced in s-process nucleosynthesis and a near-normal com¬ ponent (Lewis et al. 1994). Diamond was the first type of presolar grain recog¬ nized and has the highest abundance at 400 ppm by weight of bulk meteorite (Lewis et al. 1987). The dia¬ monds have an average size of only 2 nm and a charac¬ teristic enrichment of the heavy and light isotopes of Xe (Fig 3). The small size of the diamonds means that the individual grains contain only a few hundred atoms and the chemical properties are dominated by the bonds on the surface (Bernatowicz et al 1990). It remains unclear as to how the Xe atoms are located in the diamonds but the concentration levels require that there be only one Xe atom per million diamond grains. One mechanism proposed involves ion implantation of the Xe atoms, but the low concentrations of other elements neighbouring Xe in the periodic table is contrary to that expected. The exact formation environment/mechanism of the dia¬ monds remains enigmatic. The diamonds may have been produced in shock waves from supernovae passing through molecular clouds or from chemical vapour deposition because the free energy difference between diamond and graphite is only lkcal/mol. Support for this theory comes from the near-normal nC/l2C ratios in the diamonds (Fig 4), although a much larger amount of graphite should have been preserved in the meteorites in these scenarios. CNO cycle + ■ Mainstream SiC 1000- in solar 272 100 10 ■ Explosive H-burning — i i i i i T 10 He-burning t5-^“ 100 12Q/13C He-burning or Explosive H-burning X4 I 1 i ■ i f r i t 1 1 - r~ 1000 Figure 4. C and N isotopic compositions of presolar diamond, graphite, and silicon carbide. The range of C and N isotopic compositions from these grains is so large that the data is dis¬ played on a plot of the logarithm of the ratios. Diamond has a composition close to solar (origin), while silicon carbide (squares) and graphite (open circles) show large variations. Most SiC oc¬ curs in the CNO-cycle quadrant but a suite of grains known as Grains X, which comprise ca. 1 % of SiC analyzed, have highly unusual characteristics and are plotted with a crossed square. These Grains X have other unusual isotopic characteristics as shown in Figures 5 and 6. Data from Hoppe et al. (1994) and Amari et al. (1990). Silicon carbide was first identified from the Murray CM meteorite by Bernatowicz et al. (1987). It is perhaps the most fruitful of the presolar grains since it contains reasonably high concentrations of a number of trace ele¬ ments and is still reasonably abundant («* 6 ppm). Indi¬ vidual grains range in size from <0.2 to around 20 pm allowing ion microprobe analysis of single grains for a variety of elements (see for example Hoppe et al. 1994). Ion microprobe analysis of both individual grains and bulk SiC indicated large isotopic anomalies in a number of elements, e.g. Si, C, N, Mg, Ti, Ba, Nd, and Sm (Bernatowicz et al 1987; Ireland et al 1991; Zinner et al 1987, 1989, 1991 a,b). The mainstream SiC grains show 12C and 1?N enrichments (Fig 4) consistent with a mix¬ ture of CNO-cycle material with isotopically normal C and N. Most of the heavier elements analyzed from SiC qualitatively show the effects of the s-process with Si isotopic compositions enriched in 2ySi and :i0Si (Fig 5), Ti isotopic compositions enriched in the minor isotopes relative to 4STi (Fig 6), and s-process enrichments in heavier elements such as Xe (Fig 3), Ba, Nd, and Sm. These grains are likely to have formed around C-rich AGB stars and dispersed into the interstellar medium by the strong outflows from these stars. However, there are discrepancies in detail between the predicted and measured abundances. For example, the best-fit line passing through the mainstream SiC Si-isotope data has a slope of 1.3 (Fig 5) which differs from the theoretical value of ca. 0.35 for s-process nucleosynthesis. 47 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 5. Si isotopic compositions of SiC grains. Most grains show elevated 29Si/28Si and ^Si / 28Si ratios relative to terrestrial by up to 200 %o . The enrichment of 29Si and “Si can be ascribed to an s-process however the correlation in the ^Si/28Si and “Si/ 28Si ratios indicates a slope of 1.3 which does not agree with the predicted value of 0.35 for s-process nucleosynthesis. Grains X have 29Si / 2RSi and “Si/^Si ratios below terrestrial (negative delta values) and are inferred to have elevated 28Si abundances which is characteristic of formation in supernova ejecta. Data from Hoppe et al. (1994). Figure 6. Ca and Ti isotopic compositions in SiC. Mainstream SiC grains have Ti isotopic compositions (a) that show enrich¬ ments in all isotopes relative to 4hTi producing a characteristic V- shaped pattern. Ca isotopic compositions of mainstream SiC (b) show no large non-linear effects. Grains X on the other hand, have Ti isotopic compositions enriched in 49Ti and one grain shows a large enrichment of ^Ca These anomalies are inferred to be due to the decay of 49V and ^Ti which are produced only in supernovae. Data from Ireland et al. (1991), Amari et al. (1992), and unpublished observations. A minor fraction of the SiC grains (ca. 1 %) show 13C and 14N enrichments as well as excesses of the decay products of short-lived radionuclides (Fig 6) such as 2bAl (now 26Mg), “Ti (now '“Ca), and 4yV (now 49Ti). It ap¬ pears that these short-lived nuclides are only produced together in supernovae and so Grains X may represent material from the C-rich zones in the outer layers of a presupemova star (Amari et al. 1992). Graphite represents less than 2 ppm (by wt) of mete¬ orite and of this material only a fraction is isotopically anomalous (Amari et al. 1990). Graphite spherules (1 to 10 pm in diameter) show a similar extreme range in C isotopic compositions as that found for SiC, but isotopi¬ cally light compositions are more commonly found. Ni¬ trogen is only present at very low levels (c.f. nearly 1 % in SiC) and so the lack of large anomalies in this element may not reflect the intrinsic composition. The graphite structure does not allow trace-element substitution to the same degree as SiC but the graphite spherules have been found to contain very small inclusions of refractory carbides of elements such as Si, Ti, Zr, Mo, and W (Bematowicz et al. 1991). It is interesting to note that production of Zr, in particular, is characteristic of the s- process suggesting that at least a fraction of the graphite grains also come from AGB stars. Interstellar oxides The isolation of presolar carbon-bearing compounds was aided by their resistance to acid attack and so de¬ struction of all the silicates and oxides in the meteorites leads to increasing concentration of these grains. How¬ ever, our solar system is clearly not derived predomi¬ nantly from C-rich stars but has a predominantly O-rich parentage. Most interstellar grains coming into the so¬ lar nebula must therefore be oxides but their isolation requires separation from isotopically normal solar nebula oxides which have largely the same physical properties. Nevertheless, interstellar corundum grains have been identified using similar techniques to those used in the isolation of the carbides (Hutcheon et al. 1994). Several grains with extreme enrichments of 26Mg and 170 and s-process Ti anomalies have been discov¬ ered, suggesting a source from O-rich AGB stars. How¬ ever, the abundance of the corundum grains is far lower than the carbon-bearing grains suggesting that the dust in the solar nebula was predominantly in the form of silicate rather than oxide. The stars revisited Isotopic anomalies direct from the sites of stellar nucleosynthesis are preserved in solar system materials. Refractory inclusions preserve their isotopic anomalies despite being processed in the solar nebula, but both the chemistry and isotopic systematics of carbide grains sug¬ gest that they have presolar connections. Local chemical fluctuations allowing carbon or carbides to condense could be envisaged in the solar nebula and so their chemical instability, while supporting an exotic origin, does not preclude formation in the solar system. However, the presence of extreme isotopic anomalies in the presolar grains is the strongest indicator that they have a stellar parentage not related to our Sun. Moreover, the 13C/12C 48 Journal of the Royal Society of Western Australia, 79(1), March 1996 ratios measured from SiC and graphite show a similar range and abundance pattern to that measured in car¬ bon stars (see Anders & Zinner 1993). Matching isotopic characteristics with a certain class of star is not straightforward, nor in general does a unique classification become apparent. Clearly, inter¬ stellar grains must be derived from stars with strong outflow winds or that have undergone some form of explosion to distribute material into the interstellar me¬ dium. The main contributor to the SiC and graphite abundances appears to be AGB stars. Perhaps this is not surprising since these stars are common, have strong outflow winds, and hence contribute a large amount of material to the interstellar medium. A small fraction of the SiC grains appears to have been produced in a su¬ pernova. It also appears that Fe-group anomalies in refractory inclusions can be related to supernova nucleo¬ synthesis but their isotopic characteristics appear quite different to those in the Grains X. Specifically, Ti from refractory inclusions is anomalous in ^Ti (and to a lesser extent 49Ti) which is related to a neutron-rich equilib¬ rium process. On the other hand, Ti in grains X has a smooth pattern except for 4yTi anomalies that are related to 49V decay. Thus, Ti in the refractory inclusions ap¬ pears to come from close to the mass cut of a supernova while Ti in Grains X is probably derived from the outer layers of a supernova. A range of nucleosynthetic sites is now available for study in the laboratory and detailed nucleosynthetic calcu¬ lations can now' be compared with actual compositions. In large part, the agreement is quite acceptable at least qualitatively. For example, the predicted s-process com¬ positions of the elements Xe, Ba, and Nd is in good agreement with the theoretical production. There are differences however that cannot be explained at the present time such as the slope of the Si correlation and the correlation between Si and Ti isotopes in mainstream SiC. Measurements of a wider range of elements and refinement of calculations will hopefully yield better in¬ sights into the nucleosynthetic conditions and evolution of stars responsible for the presolar dust grains. Acknowledgments: This paper is based on a talk given at the de Laeter Symposium held in Perth November 1995 to honour the retirement of John de Laeter from the Curtin University of Technology. 1 am grateful to the organizers for inviting me to participate in this memorable occasion. This paper has benefitted considerably from time spent with E Zinner. M Zadnik provided a detailed review which tidied up the text. References Amari S, Anders E, Virag A & Zinner E 1990 Interstellar graphite in meteorites. Nature 345:238-240. Amari S, Hoppe P, Zinner E & Lewis R S 1992 Interstellar SiC with unusual isotopic compositions: grains from a supernova? Astrophysical Journal 394:L43-L46. Anders E & Zinner E 1993 Interstellar grains in primitive meteor¬ ites: diamond, silicon carbide, and graphite. Meteoritics 28:490-514. Bernatowicz T, Fraundorf G, Tang M, Anders E, Wopenka B, Zinner E & Fraundorf P 1987 Evidence for interstellar SiC in the Murray carbonaceous chondrite. Nature 330:728-730. Bernatowicz T J, Gibbons P & Lewis R S 1990 Electron energy loss spectrometry of interstellar diamonds. Astrophysical Journal 359:246-255. Bernatowicz T J, Amari S, Zinner E K & Lewis R S 1991 Interstel¬ lar grains within interstellar grains. Astrophysical Journal Let¬ ters 373:L73-L76. Birck J L & Allegre C J 1984 Chromium isotopic anomalies in Allende refractory inclusions. Geophysical Research Letters 11:943-946. Black D C & Pepin R O 1969 Trapped neon in meteorites. II. Earth and Planetary Science Letters 6:395-405. Burbidge E M, Burbidge G R, Fowler W A & Hoyle F 1957 Syn¬ thesis of the elements in stars. Reviews of Modem Physics 29:547-650. Cameron A G W 1962 The formation of the sun and planets. Icarus 1:13-69. Cameron A G W & Truran J W 1977 The supernova trigger for formation of the solar system. Icarus 30:447-461. Clayton R N, Grossman L & Mayeda T K 1973 A component of primitive nuclear composition in carbonaceous meteorites. Science 182:485-488. Clayton R N, Hinton R W & Davis A M 1988 Isotopic variations in the rock-forming elements in meteorites. Philosophical Transactions of the Royal Society of London A325:483-501. Fahey A J, Goswami J N, McKeegan K D & Zinner E 1985 Evi¬ dence for extreme a,Ti enrichments in primitive meteorites. Astrophysical Journal Letters 296:L17-L20. Fahey A J, Goswami J N, McKeegan K D & Zinner E 1987 lhO excesses in Murchison and Murray hibonites: A case against a late supernova injection origin of isotopic anomalies in O, Mg, Ca, and Ti. Astrophysical Journal Letters 323:L91-L95. Gray C M & Compston W 1974 Excess 2hMg in the Allende mete¬ orite. Nature: 495-497. Hartmann D, Woosley S E & El Eid M F 1985 Nucleosynthesis in neutron-rich supernova ejecta. Astrophysical Journal 297:837- 845. Hinton R W, Davis A M & Scatena-Wachel D E 1987 Large nega¬ tive 5£)Ti anomalies in refractory inclusions from the Murchison carbonaceous chondrite - evidence for incomplete mixing of neutron-rich supernova ejecta into the solar system. Astrophysical Journal 313:420-428. Heydegger H R, Foster J J & Compston W 1979 Evidence of a new isotopic anomaly from titanium isotopic ratios in mete¬ oric materials. Nature 278:704-707. Hoppe P, Amari S, Zinner E, Ireland T & Lewis R 1994 Carbon, nitrogen, magnesium, silicon, and titanium isotopic composi¬ tions of single interstellar silicon carbide grains from the Murchison carbonaceous chondrite. Astrophysical Journal 430:870-890. Huss G R & Lewis R S 1994 Noble gases in presolar diamonds. I. Three distinct components and their implications for diamond origins. Meteoritics 29:791-810. Hutcheon 1 D, Huss G R, Fahey A J & Wasserburg G J 1994 Extreme 2,1Mg and 170 enrichments in an Orgueil corundum: identification of a presolar oxide grain. Astrophysical Journal Letter 425:L97-L100. Ireland T R 1990 Presolar isotopic and chemical signatures in hibonite-bearing refractory inclusions from the Murchison carbonaceous chondrite. Geochimica et Cosmochimica Acta 54:3219-3237. Ireland T R 1995 Ion microprobe mass spectrometry: Techniques and applications in cosmochemistry, geochemistry, and geo¬ chronology. In: Advances in Analytical Geochemistry Vol II (eds M Hyman & M Rowe). JAI Press, Greenwich, 1-118. Ireland T R, Compston W & Heydegger H R 1985 Titanium isoto¬ pic anomalies in hibonites from the Murchison carbonaceous chondrite. Geochimica et Cosmochimica Acta 49:1989-1993. Ireland T R, Zinner E K & Amari S 1991 Isotopically anomalous Ti in presolar SiC from the Murchison meteorite. Astrophysi¬ cal Journal Letters 376:L53-L56. Ireland T R, Zinner E K, Fahey A J & Esat T M 1992 Evidence for distillation in the formation of HAL and related hibonite in¬ clusions. Geochimica et Cosmochimica Acta 56:2503-2520. 49 Journal of the Royal Society of Western Australia, 79(1), March 1996 Jungck M H A, Shimamura T & Lugmair G W 1984 Ca isotope variations in Allende. Geochimica et Cosmochimica Acta 48:2651-2658. Lee T 1988 Implications of isotopic anomalies for nucleosynthe¬ sis. In: Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). University of Arizona Press, Tucson, 1063- 1089. Lee T & Papanastassiou D A 1974 Mg isotopic anomalies in the Allende meteorite and correlation with O and Sr effects. Geo¬ physical Research Letters 1:225-228. Lee T, Papanastassiou D A & Wasserburg G J 1977 Aluminium- 26 in the early solar system: Fossil or fuel? Astrophysical Jour¬ nal Letters 211:L107-L110. Lewis R S, Tang M, Wacker J F, Anders F & Steel E 1987 Interstel¬ lar diamonds in meteorites. Nature 326:160-162. Lewis R S, Amari S & Anders E 1994 Interstellar grains in mete¬ orites: II. SiC and its noble gases. Geochimica et Cosmochimica Acta 58:471-494. Niemeyer S & Lugmair G W 1981 Ubiquitous isotopic anomalies in Ti from normal Allende inclusions. Earth and Planetary Science Letters 53:211-225. Ott U 1993 Interstellar grains in meteorites. Nature 364:25-33. Reynolds J H & Turner G 1964 Rare gases in the chondrite Renazzo. Journal of Geophysical Research 69:3263-3281. Truran J 1993 Interstellar grains in the laboratory. Workshop on Isotopic Anomalies, Washington University, St. Louis, Missouri. Zinner E K, Fahey A J, Goswami J N, Ireland T R & McKeegan K D 1986 Large ^Ca anomalies are associated with ^'Ti anoma¬ lies in Murchison and Murray hibonites. Astrophysical Jour¬ nal Letters 31LL103-L107. Zinner E, Tang M & Anders E 1987 Large isotopic anomalies of Si, C, N and noble gases in interstellar silicon carbide from the Murray meteorite. Nature 330:730-732. Zinner E, Tang M & Anders E 1989 Interstellar SiC in the Murchison and Murray meteorites: Isotopic composition of Ne, Xe, Si, C, and N. Geochim. Cosmochim. Acta 53:3273- 3290. Zinner E, Amari S, Anders E & Lewis R 1991a Large amounts of extinct 2t,Al in interstellar grains from the Murchison carbon¬ aceous chondrite. Nature 349:51-54. Zinner E, Amari S & Lewis R S 1991b s-Process Ba, Nd, and Sm in presolar SiC from the Murchison meteorite. Astrophysical Journal Letters 382:L47-L50. 50 Journal of the Royal Society of Western Australia, 79:51-57, 1996 Solar and solar system abundances of the elements M Ebihara Department of Chemistry, Faculty of Science, Tokyo Metropolitan University, 1-1 Minami-Ohsawa, Hachioji, Tokyo 192-03 Japan Abstract With the recent progress in astrophysical observations, the solar abundances of the elements (elemental composition of the sun) can be profitably compared with the solar system abundances of the elements. In this paper, current solar and solar system abundances of the elements are reviewed and are compared; there seems to be no systematic difference. Solar system abundances of the elements are mostly deduced from the chemical composition of meteorites (mainly Cl chondrites). Compiling and evaluating the data for C2 chondrites, I once inferred that C2 chon¬ drites also were potential candidates for the solar system standard of elemental abundances. The current data show that their abundances correlate with the solar system abundances extremely well. However better data are needed for C2 chondrites in order to acsertain if the correlationis as good as for Cl chondrites. In the near future, the solar wind will be collected by a spacecraft and yield solar abundances with quality comparable to meteorite data. Introduction Solar system abundances of the elements are among the most fundamental quantities in earth and planetary sciences. They are measured quantities, not physical con¬ stants, and their quality (precision and accuracy) has increased with the advances in analytical techniques. So¬ lar system abundances of the elements are mostly based on the chemical analysis of meteorites. On the other hand, solar abundances of the elements are obtained from spectral observation of the sun. In general, the quality of the solar abundances is much poorer than the solar system abundances, although comparison of the two data sets has become possible with the great im¬ provement in the former values in recent years. Solar system abundances of the elements have been repeatedly presented. Among the earliest reports, the table compiled by Goldschmidt (1938) is the most famous. To estimate the solar system abundances of the ele¬ ments, he used both analytical data on meteorites and observational values for the sun. From this table, Suess (1947) later deduced several rules governing the abun¬ dances of the elements in the solar system. He found that the elemental abundances were highly dependent on the stability of nuclides, and so his rules were called nuclear systematics. In 1956, Suess & Urey presented their famous table of element abundances in the solar system. They used not only data then available for meteorites and the sun, but also the nuclear systematics postulated by Suess (1947). This table was the basis of the B2FH model for the nucleosynthesis of the elements (Burbidge et al. 1957). During the next two decades, Cameron periodically compiled the elemental abun¬ dance data for the solar system (e.g. Cameron 1973, 1982). In those days, there were significant developments © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 in analytical techniques (both method and instrument) for meteorites, and so not only the number was increased but also the quality of data were improved. This was a major reason why Cameron revised the abundance table at every opportunity. In 1982, I had the opportunity of compiling solar sys¬ tem abundances of the elements in collaboration with E Anders (Anders & Ebihara 1982). On that occasion, we used our own analytical data for meteorites (mainly Cl chondrites) in addition to the literature data and com¬ piled an abundance table after critical evaluation of the data. This abundance table was widely accepted by the scientific community and updated in 1989 (Anders & Grevesse 1989). There were few significant changes in solar system abundances of the elements between the two tables, because the analytical data of meteorites re¬ ported during that period were limited. In contrast, the solar abundances improved significantly. Reflecting this, Anders & Grevesse (1989) discussed solar and solar sys¬ tem abundances of the elements separately as well as comparatively. Their table, especially for solar system abundances is the current reference for the abundances of elements. As in the past, such compilations need to be revised, although the changes would not be great, espe¬ cially for elemental abundances of the solar system. In this paper, solar and solar system abundances of the elements are reviewed and some future perspectives are presented. Solar abundances of the elements The outer part of the sun consists of the photosphere, chromosphere and corona; the solar photosphere is sur¬ rounded by the chromosphere and further enclosed by the corona. The energy released by the present sun is estimated to be 3.83 1033 erg sec1 and is believed to be supplied by nuclear fusion reactions. Judging from the current temperature and pressure estimated for the cen- 51 Photosphere Journal of the Royal Society of Western Australia, 79(1), March 1996 tre of the sun, it is inferred that the following reactions (named p-p chains) mainly occur inside the sun; p + p -► 2D + e+ + v (99.75 %) p + e- + p - 2D + v (0.25 %) The deuterium produced reacts with a proton to form 3He as follows: 2D + p -» 3He + -y From 3He, 4He is produced via three different paths named PP1, PP2 and PP3 as shown below; [PP1] 3He + 3He - 4He + 2p (86 %) [PP2] 3He + 4He -* 7Be + y 7Be + e* -> 7Li + v 7Li + p - 2 4He (14 %) [PP3] 3He + 4He -> 7Be + y 7Be + p -> 8B -i- y 8B -► 8Be + e+ + v 8 Be - 24He (0.02 %) The end result is that four protons produce 4He. Un¬ der the present conditions, the PP1 reaction is dominant whereas the PP2 and/or PP3 reactions only become sig¬ nificant when temperature reaches 107K. There is an¬ other reaction for producing 4He from protons, called the CNO cycle, in which carbon, nitrogen and oxygen react as catalysts. This reaction becomes important only when the temperature reaches 2 x 107 K. Thus there is the possibility that elemental abundances of hydrogen, helium, lithium and beryllium have changed in the sun. j ' TTTTTTTJ - 1 — 1 1 1 MU i i m mi - 1 — i i i mi i i i imi| i ii i mi -e-r-r TTTTTJ : | j : o : - . . - ° * \ i z o ; o oo 90 . : : E : o : o _ l _ . . ill 11 J _ L_ LIlXUlJ _ 1—1- ■ Corona Figure 1. Coronal abundances versus solar photospheric abun¬ dances of elements. Elements of first ionization potential (FIP) < 10 eV are not fractionated from Si (= 3.55 107), which is marked by a solid circle. Elements of higher FIP are depleted in the co¬ rona by a constant factor, lying on a line parallel to that defined by the elements including Si. Data from Anders & Grevesse (1989). How are the abundances for those elements heavier than beryllium affected in the photosphere? HB could be produced through the PP3 reaction shown above, but it promptly decays to 8Be with a half-life of 0.77 sec. yB, which is the only stable nuclide of boron (Z=5), could be produced by spallation reactions with heavier nuclides according to nuclear synthesis models. In order to produce carbon (Z=6), the following reac¬ tion must occur: 8Be + 4He -+ 12C Since the half-life of KBe is extremely short (1C)'16 sec), the above reaction can be initiated only at high tempera¬ tures (> 108 K) and pressures (high density). Even the center of the present sun does not reach these condi¬ tions. Therefore, those elements heavier than Z=4 (beryl¬ lium) can be scarcely synthesized in the present sun, suggesting that (except for hydrogen, helium, lithium and beryllium) the elemental abundances of the present sun are essentially the same as those of the ancient sun 4.56 10y y ago. There are several way in which the solar abundance of the elements can be determined. These are: (a) photosphere (absorption intensity) (b) corona (emission intensity) (c) solar wind (intensity of low energy ion) (d) solar flare (intensity of high energy ion) The solar photosphere is the outermost layer of the sun, with an average depth of 400 km. It is generally believed that the photosphere has the representative el¬ emental abundances of the sun, because the surface ma¬ terials of the sun are not convected and are scarcely mixed with the inner material, which can be affected by the nuclear reactions (more extensively than the photo¬ sphere). Although the solar photosphere is thus the most suitable source for the solar abundance of the elements, there are several difficulties in obtaining reliable data. Nevertheless, precision of the solar photospheric data has steadily improved over the last two decades. The solar abundances of the elements can be also in¬ ferred from elemental abundances of the corona, ob¬ tained either by observing emission spectra of the co¬ rona or by measuring chemical compositions of the cos¬ mic rays it omits. Coronal emission spectra of different regions of the corona have been observed from space¬ craft; for instance the corona hole, an active area, a quiet area and a prominence. Although emission spectrom¬ etry can potentially yield highly reliable data, the result¬ ing values are not as reproducible as expected and so the solar abundances of the elements deduced from the spectra have relatively large errors (uncertainties). Cosmic rays emitted from the corona are called the solar wind and solar energetic particles (SEP). The solar wind is a current of plasma with high velocity released from the solar corona, and consists mainly of protons and elec¬ trons. Although the number of the elements measured in the solar wind is not large (<10), their abundances are in agreement with those obtained from coronal spec- trometric observations. The SEP has much higher energy (up to MeV) than solar wind (measured in keV) and is emitted from the corona by several different mecha¬ nisms. The elemental abundance data for SEP are superb 52 Journal of the Royal Society of Western Australia, 79(1), March 1996 in both quality and quantity compared with those for solar wind. In Figure 1, the elemental abundances of the corona are compared with those from the photosphere. The coronal abundances are averaged values from different sources but are mostly represented by the SEP values. There seems to be a good agreement between these two sources although the coronal abundances are systemati¬ cally lower than the photospheric data for several ele¬ ments. These elements have a common characteristics; their first ionization potentials are consistently high (> lleV). This suggests that there is elemental fractionation between neutral elements (which have relatively high ionization potentials) and ionic elements (which have relatively low ionization potentials) when the elements are transferred from the chromosphere to the corona. Solar system abundances of the elements The solar system abundances are mostly determined from the chemical analyses of meteorites. Exceptions are hydrogen, carbon, nitrogen, oxygen and noble gas ele¬ ments, which are all highly volatile. Among meteorites, Cl (or Cl) chondrites have been the most extensively analyzed for this purpose. So far, nine meteorites have been recognized to be Cl or Cl-like (Table 1). Of these, four were collected in Antarctica. Antarctic meteorites are susceptible to terrestrial weathering on Antarctica. In fact, Antarctic Cl (or Cl-like) meteorites are not al¬ ways the same as non- Antarctic Cl meteorites in chemi¬ cal composition, mineralogical texture and/or oxygen isotopic composition, indicating that Antarctic CTs are not suitable specimens for measuring solar system abun¬ dances of the elements. The terminology of "chondrite" is due to the presence of chondrules in the meteorite's texture. If this definition is strictly applied, then Cl chondrites can't be called chondrites, because contain no chondrules but consist of 100% matrix materials. This suggests that Cl chondrites are the richest in volatile components among meteorites. Indeed, the water content (measured as HzO) exceeds 20% (weight %) because the matrix is composed of a large amount of hydrous silicates. Non-Antarctic Cl chondrites are all falls (meteorites which were observed to fall and then collected). In contrast, Antarctic CTs are all finds (meteorites which were not observed to fall). Among the 4 Antarctic CTs, Y 82162 might be a real Cl, but the remaining are between Cl and C2. Among non- Antarctic CTs, Revelstoke and Tonk are too small to be subjected to chemical analysis. Of the remaining three non- Antarctic CTs, Orgueil is the largest and the most frequently analyzed. Thus, solar system abundances of the elements are mostly deduced from the chemical composition of only one meteorite. There are large dif¬ ferences (>10 %) for Hf and Hg between the values of Anders & Ebihara (1982) and Anders & Grevesse (1989). For the rest of the elements, the two sets of values esti¬ mated by these authors are within 10 %. Solar abundance values for hydrogen, nitrogen, oxy¬ gen and noble gas elements are deduced from sources other than meteorites. For these elements except noble gases, the solar photospheric data are used. For noble gas elements, each has its own method of estimation. The abundance of helium is calculated using the photo¬ spheric value of hydrogen and the He/H ratio for the Fill region and/or hot stars (outside the solar system). Solar abundances of neon and argon are also derived from observations of the HII region and other extra¬ solar regions. However, the values estimated from mea¬ surements of prominences, the solar wind and SEP are generally in agreement with those obtained from the spectroscopy of the stars outside our solar system. For argon, an estimated (interpolated) value of ^Ar based on the abundances of 28Si and ^Ca is also consistent with the above values. The abundances of krypton and xenon are determined from the abundances of neighboring ele¬ ments. For krypton, there are two approaches; either s- process systematics or interpolations based on 81 Br and ^Rb for 83Kr, and ^Se and ^Sr for ^Kr. The two values so obtained are similar. Characteristics of solar and solar system abundances of the elements Solar versus solar system abundance of the elements The sun and other members of the solar system must have formed almost simultaneously about 4.56 109 y ago. Table 1 Cl and Cl-like chondrites recovered so far. Meteorite Year of Recovery Place of Recovery Fall or Find Original Weight non-Antarctic Alais 1806 France Fall 6 kg Ivuna 1938 Tanzania Fall 705 g Orgueil 1864 France Fall >10 kg Revelstoke 1965 Canada Fall Tonk 1911 India Fall 7-7 g Antarctic * Belgica 7904 1979 Antarctica Find 1.23 kg Yamato 82042 1982 Antarctica Find 37.1 g Yamato 82162 1982 Antarctica Find 39.5 g Yamato 86720 1986 Antarctica Find 859 g *The following Cl properties (shown in parentheses) are confirmed for the Antarctic meteorites; Belgica 7904 (O-isotopes), Yamato 82042 (petrography), Yamato 82162 (O-isotopes, chemistry) and Yamato 86720 (O-isotopes). 53 Journal of the Royal Society of Western Australia, 79(1), March 1996 The heterogeneity of the source material, proto-solar nebula, has been a subject of discussion for many years. There are several lines of evidence indicating that the solar nebula was heterogeneous in isotopic composition. In contrast, there are no data supporting inhomogeneity of the solar nebula in elemental abundances. The sun represents 99.7 % mass of our solar system, indicating that the elemental abundances of the sun are necessarily those of the solar system. As discussed previously, the elemental abundances of the present sun must be the same as those of the sun of 4.56 104 y ago for all the elements except hydrogen, helium, lithium and beryl¬ lium. Because solar system abundances of the elements are based on meteorite data, then solar abundances and solar system abundances are essentially identical for these elements except for several light elements and noble gas elements if meteorites, especially Cl chon¬ drites, preserve the elemental abundances of the proto¬ solar nebula. In Figure 2, solar and solar system abundances of the elements are compared; elemental abundances (the num¬ ber of elements) are all normalized to Si = 10h atoms. As is readily apparent, elements having relative abundances of more than eight orders of magnitudes form a straight line with a slope of 0.988 (r = 0.977). Whenever either solar abundances or solar system abundances of the ele¬ ments have been revised, such a graph has been repeat¬ edly drawn and the extent of correlation has increased with the time. Considering that the precision of meteor¬ ite data, and especially photospheric data, has greatly improved in recent years and that the agreement be¬ tween solar abundances and solar system abundances of the elements has been enhanced in recent years, it is clear that elemental abundances of meteorites, especially Cl chondrites, are essentially equal to those of our solar system. Cl chondrite Figure 2. Solar photospheric abundances versus Cl chondrite abundances of elements; abundances are normalized to Si = 106. Only Li is located largely apart from the correlation line. From Ebihara (1992). The scale on both axes are logarithm. Suess' nuclear systematics for solar system abundances of the elements Suess (1947), evaluating the then available solar sys¬ tem abundances of the elements, found the following systematics for the abundances of the solar system nu¬ clides: (a) abundances of the nuclides having an odd mass number change smoothly with increasing mass number A for the region of A>50; in this case, abun¬ dances of the isobars (nuclides having the same A) are combined; (b) when A is even, the following variables change smoothly with increasing A; (i) for the nuclides of A<90, the total abundances of the nuclides having the same I, where I is defined as I = A - 2Z; (ii) for the nuclides of A>90, the total abundances of the isobaric nuclides; (c) for isobaric nuclides, the nuclide having a larger I is more abundant than that having a smaller I for the region of A<70; for the region of A>70, the above relation is reversed; (d) when the numbers of nucleons (proton and neutron) correspond to so-called "magic numbers", some ir¬ regularities arise in the relationships mentioned above. When Suess & Urey (1956) summarized the cosmic abundances* of the elements, they used the nuclear sys¬ tematics postulated by Suess (1947). They revised and even infered several values of solar abundances based on these systematics. In later years, these revised or in¬ ferred values were replaced with the values measured in meteorites, mainly those of Cl. This gave great credit to the Suess' nuclear systematics, even though it is an em¬ pirical rule. Thereafter, Suess' nuclear systematics, espe¬ cially the first rule, has often been used in evaluating solar system abundances of the elements. Abundances of odd-mass nuclides in Cl chondrites are shown in Figure 3A for A = 59 to 139, and Figure 3B for A = 131 to 209. Elemental abundances of Cl chon¬ drites compiled by Anders & Grevesse (1989) and stable isotope abundances recommended by the International Union of Pure and Applied Chemistry (IUPAC 1991) are used to calculate relative abundances of nuclides having odd A. There appear to be some peaks in Figure 3A. The peaks at A = 89 and A = 117 to 119 correspond to the irregularity due to "magic numbers"; N (neutron num¬ ber) = 50 for the former and Z = 50 (Sn) for the latter. A peak appearing around 130 can be explained by a rela¬ tively high probability of nuclide formation produced by p-process nucleosynthesis. In both Figures 3A and 3B, we see very smooth changes in abundances. In 1982, Anders & Ebihara (1982) identified irregularities at the *In those days, the term "cosmic abundances of the nuclides" was often used in place of the term "solar system abundances of the elements", possibly with a tacit understanding that there was no significant difference between the two abundances. Strictly speaking, cosmic abundances of the elements cannot be obtained by analysis of any material. However, as our solar system is not special in the astronomical sense, we can assume that the solar system abundances of the elements are representative of the cos¬ mic abundances of the elements, to a first order approximation. 54 Abundance / Si (10 atoms) Journal of the Royal Society of Western Australia, 79(1), March 1996 1 — ' — I — • — 1 — • — 1 — ' — 1 — ' — 1 — ' — 1 — 1 — 1 ' 1 ' I ' 1 ' 1 ' 1 ' 1 ' ' ; • -• — 1 » 1 — ' — i — ' — 1 — ' r— ' ; 8 Cl chondrite 8 ° C2 chondrite • o o • o • o • cm m • • • • » • - O o ° ° o o • O r • o • ° • a • • # 8 • • • r ° ° o • * ? • • • *• • • ° ° o O ° O i 3 ■ i ■ ■ . i _ _ _ l _ . _ i — . — I — . — I — - — I — • — i — ‘ — 1 — ' — 1 — 1 — 1 — 1 — 1 — 1 — L • ■ - _ 1 _ _ — 1 — 1 — 1 — * — 1 — * — 10" io£ 10 1 V) E o CO CO 10C Qi U c co T3 C D n < 10 - 1 10' 10' 59 63 67 71 75 79 83 87 91 Mass Number io 1 10 v- 10 - 1 10' 10' » ■ -i — • — i — ' — i — ' — i — ■ — i — ' — r — ' i ' i 1 i ' 1 1 1 ' 1 ' 1 ' 1 • - 1 - ^ - 1 - ' - 1 - * - 1 - - - 1 - - - - Cl chondrite O C2 chondrite ► • • 8 8 8 ‘ • ^ o - • 8 o •o O • 9 • 0 ® ® m o o : • o o 8 * 8 • cm cm 8g" ® § 8 8 o o *••28 3 1 ■ . . 1 1 I 1 1 ' ' ‘ L x i _ _ _ i _ . — i — . — i — . — i — • — 131 135 Figure 3A,B Mass Number . Abundances of odd-mass nuclides from Co to La (A) and from Xe to Bi (B) in Cl and C2 chondrites. From Ebihara (1992). 55 Journal of the Royal Society of Western Australia, 79(1), March 1996 Ag-Cd region and the Nd-Sm-Eu-Gd region. The former irregularity almost disappears in Figure 3A, where new and more accurate values of Ag for Cl chondrites ob¬ tained by using mass spectrometric isotope dilution technique (Loss et al 1984) are used. In contrast, the irregularity at the light lanthanoid region remains unre¬ solved. Anders & Grevesse (1989) concluded that this irregularity suggested the limitation of the Suess' nuclear systematics in evaluating solar system abun¬ dances of the elements. Future perspective for the study on solar and solar system abundances of the elements Do Cl chondrites yield the solar system abundances of the elements? Cl chondrites are believed to be the most primitive meteorites and to have experienced no thermal activity since their formation at 4.56 109 y ago. This is supported by the observation that Cl chondrites are the richest in volatile compounds (including organic materials) and el¬ ements* However, mineralogical observations clearly show that Cl chondrite parent bodies have experienced a significant degree of aqueous alteration. (Tomeoka & Busek 1988). Chemical analyses of Cl chondrites reveal a positive correlation between the abundances of several pairs of elements (Ebihara et al 1982). This correlation was interpreted to reflect aqueous alteration of the Cl chondrite parent body. Furthermore, it was recently dis¬ covered that refractory siderophile elements (Re, Os and Ir) were fractionated among ordinary and carbonaceous chondrites including Cl chondrites (Ebihara & Ozaki 1995); Si-normalized values for these elements in ordi¬ nary and carbonaceous chondrites are summarized in Table 2. It can be seen that the Os and Ir values of H chondrites are similar to those of the Cl chondrites esti¬ mated by Anders & Ebihara (1982). The Os and Ir values of C2 and C3 chondrites are nearly equal, being 5-6 % higher than those of H and Cl chondrites. For the abun¬ dance of Re, H chondrites give the highest value whereas Cl chondrites yield the lowest value. Thus, it is apparent that Re is fractionated from Os and Ir among ordinary and carbonaceous chondrites. These elements are cosmochemically grouped as the most refractory ele¬ ments, and so their fractionation on the parent body Table 2 Atomic abundances of Re, Os and Ir in carbonaceous and ordi¬ nary chondrites (relative to 106 Si atoms)3. Meteorite groups Re Os Ir Carbonaceous Cl A&Eb 0.0507 0.717 0.660 A&GC 0.0517 0.675 0.661 C2 0.0589 0.740 0.685 C3 0.0619 0.769 0.715 Ordinary H 0.0665 0.723 0.669 L 0.0353 0.403 0.366 LL 0.0258 0.324 0.270 3 Ebihara & Ozaki (1995); bAnders & Ebihara (1982); cAnders & Grevesse (1989). seems to be very unlikely. As Re is the most refractory of the siderophile elements, condensation and/or accre¬ tion processes must be responsible for this fractionation. Thus, it appears that Cl chondrites are not as primitive and pristine as they were once thought to be. C2 (or CM) chondrites are much closer to Cl chon¬ drites than any other group of meteorites (Ebihara 1992). C2 chondrites have apparently experienced aqueous al¬ teration, to a lesser degree than Cl. Elemental abun¬ dances of C2 chondrites are compared with the solar photospheric data in Figure 4. There is a high correlation between two abundances (albeit less than in Figure 2). The Suess plot for C2 chondrites is shown in Figures 3A and 3B, and is compared with that for Cl chondrites. C2 chondrite abundances change as smoothly as those of Cl. The quality of the data for Cl chondrites are much better than that for C2 chondrites, because Cl chondrites have been analyzed much more extensively than C2 chondrites, and there has been greater scrutiny of data. C2 chondrite Figure 4. Solar photospheric abundances versus C2 chondrite abundances of elements; abundances are normalized to Si=106. From Ebihara (1992). The scale on both axes are logarithm. From the comparison of Cl and C2 abundances on Figures 2,3 and 4, it can be concluded that C2 chondrites may also be an alternative choice as a solar system stan¬ dard. The number of C2 falls is several times the num¬ ber of Cl falls. Unfortunately, the quality of analytical data for C2 chondrites doesn't seem to be as high as that for Cl. The essential question is, which meteorites are the most suitable for evaluating the solar system abun¬ dances of the elements? Can Cl chondrites still claim the status as the solar system standard? Or, can C2 chon¬ drites be substituted for Cl chondrites? The propriety for being the solar system standard should be judged by the consistency with solar abundances of the elements rather than the abundance of volatile materials/ele¬ ments. It is regrettable that the quality of data for solar abundances is much poorer that that for meteorites even at present, and it is unlikely that the situation will 56 Journal of the Royal Society of Western Australia, 79(1), March 1996 change in the near future. Data sources for solar abun¬ dances of the elements which can be analyzed with qual¬ ity comparable with that for meteorites are required. Solar wind as a promising source of the solar system abundance of the elements Solar wind is a source of data for solar abundances of the elements which is independent of the solar photo¬ sphere. Only a few elements (hydrogen, helium, carbon, nitrogen, oxygen, neon, silicon, argon and iron) have so far been determined in the solar wind. He/H ratios are relatively easily determined by observation from space¬ craft. It is well known that the He/H ratio of solar wind varies over minutes to days by a factor of more than 100. With a systematic variation between solar minimum and solar maximum, the mean value for half a year be¬ comes constant, 0.053 to 0.050; this is considerably smaller than the value adopted for the solar system (~0.1). Solar wind abundances of noble gas elements having small Z (helium, neon and argon) were deter¬ mined by the Apollo mission (Geiss et al 1972) which exposed to solar wind aluminum and platinum foils spread on the surface of the moon for several hours to days. These foils were then returned to earth and ana¬ lyzed by noble gas mass spectrometry. The results showed that the particle flux changed in the short term, but that the relative abundances of elements and iso¬ topes were essentially constant. Elemental abundances of helium, neon and argon deduced from observation of the HII region are in agreement with those obtained from solar wind and SEP. For carbon, nitrogen, oxygen, silicon and iron, the data obtained by the ISEE-3 satellite are available; the data for oxygen and iron are fairly reliable while those for the remaining elements cannot be regarded as representative of the solar wind. It seems to be impossible to improve the quality of the data for solar wind using the present techniques and methods. To overcome this, a long-term collection of solar wind and/or a high sensitivity analytical method for chemical and isotopic measurements is needed. For the first alternative, a proposal is the collection of solar wind by spacecraft, similar to the Apollo moon foil ex¬ periments. For such measurements a spacecraft is re¬ quired to stay in space as long as possible, and finally return to the earth. It is highly probable that this will be realized early in the 21st century. Acknowledgments: The author thanks the symposium committee for the opportunity to present this paper on the occasion of the de Laeter Sympo¬ sium on Isotope Science. This work was supported in part by Grant-in- Aids for Scientific Research of the Ministry of Education, Science and Cul¬ ture, Japan (No. 0545012). References Anders E & Ebihara M 1982 Solar-system abundances of the ele¬ ments. Geochimica et Cosmochimica Acta 46:2363-2380. Anders E & Grevesse N 1989 Abundances of the elements: Mete- oritic and solar. Geochimica et Cosmochimica Acta 53:197- 214 Burbidge E M, Burbidge G R, Fowler W A & Hoyle F 1957 Syn¬ thesis of the elements in stars. Review of Modern Physics 29:547-650. Cameron A G W 1973 Abundances of the elements in the solar system. Space Science Review 15:121-146. Cameron A G W 1982 Elementary and nuclidic abundances in the solar system. In: Essays in Nuclear Astrophysics (eds C A Barnes, D N Schramm & D D Clayton). Cambridge University Press, Cambridge, 23-43. Ebihara M 1992 Evaluation of the Cl and other carbonaceous chondrites as the solar abundance standards. In: Unstable Nu¬ clei in Astrophysics (eds S Kubono & T Kajino). World Scien¬ tific, Singapore, 161-169. Ebihara M & Ozaki H 1995 Re, Os and Ir in Antarctic unequilibrated ordinary chondrites and implications for the solar system abundance of Re. Geophysical Research Letters 22:2167-2170. Ebihara M, Wolf R & Anders E 1982 Are Cl chondrites chemi¬ cally fractionated? A trace element study. Geochimica et Cosmochimica Acta 46:1849-1861. Geiss J, Buehler F, Cerutti H, Eberhardt P & Filleux C 1972 Solar wind composition experiment. In: Apollo 16 Preliminary Sci¬ ence Report. NASA SP-315:14-1 - 14-10. Goldschmidt V M 1938 Geochemische verteilungsgesetze der Elemente. IX. Die Mengenverhaltnisse der elemente und der atom-arten. Skrifter utgit av det Norske Videnskaps-Akademi I Oslo. I. Math-naturv. Klasse, 1937, 4:1-148. IUPAC Inorganic Chemistry Division 1991 Isotopic Compositions of the Elements 1989. Pure and Applied Chemistry 63:991- 1002. Loss R D, Rosman K J R & de Laeter J R 1984 Mass spectrometric isotope dilution analyses of palladium, silver, cadmium and tellurium in carbonaceous chondrites. Geochimica et Cosmochimica Acta 48:367-388. Suess H E 1947 Uber kosmische Kernhaufigkeiten. I. Einige Haufigkeitsregeln und ihre Anwendung bei der Abschatzung der Hciufigkeitswerte fur die mittelschweren und schweren Elemente. Zeitschnft fur Naturforschung 2A:311-321. Suess H E & Urey H C 1956 Abundances of the elements. Review of Modern Physics 28:53-74. Tomeoka K & Busek P R 1988 Matrix mineralogy of the Orgueil carbonaceous chondrite. Geochimica et Cosmochimica Acta 52:1627-1640. 57 Journal of the Royal Society of Western Australia, 79:59-65, 1996 Origin of the terrestrial planets and the moon S R Taylor Department of Nuclear Physics, Research School of Physical Sciences Australian National University, Canberra, ACT 2000 Abstract Our ideas about the origin and evolution of the solar system have advanced significantly as a result of the past 25 years of space exploration. Metal-sulfide-silicate partitioning seems to have been present in the early dust components of the solar nebula, prior to chondrule formation. The inner solar nebula was depleted in volatile elements by early solar activity. The early formation of the gas giant, Jupiter, affected the subsequent development of inner solar system and is responsible for the existence of the asteroid belt, and the small size of Mars. The Earth and the other terrestrial planets accreted in a gas-free environment, mostly from volatile-depleted planetesimals which were already differentiated into metallic cores and silicate mantles. The origin of the Moon by a single massive impact with a body larger than Mars explains the angular momentum, orbital characteristics and unique nature of the Earth-Moon system. The density and chemical differences between the Earth and Moon are accounted for by deriving the Moon from the mantle of the impactor. The relation of the terrestrial planets to the solar system The early history of the rocky terrestrial planets has to be placed in the broader perspective of the evolution of the solar system. They constitute such a tiny proportion of the original solar nebula that to a first approximation they could be ignored, except that we are standing on one of them. A basic question is whether the Earth and the other planets were formed by breakup of the gaseous solar nebula, or assembled "brick by brick" from smaller bodies. Were they formed in the nebula while the main gaseous and icy constituents were present, or had the hydrogen, helium, water, methane and ammonia, that constituted 99.5% of the primordial nebula, been dis¬ persed before the formation of the inner planets? Why is Jupiter so large and what effect has it had on the rest of the system? Why is Mars so tiny, compared not only with the Earth, but to massive Jupiter? Why is there such a small amount of matter in the asteroid belt? Why is the Earth, and apparently Venus and Mars, depleted in vola¬ tile elements? Is this a local or more widespread phenonomen? Did the Earth accrete from a local zone in the nebula, or was there widespread mixing and homog¬ enization in the early nebula? What is the relationship of the Moon to the Earth and what effect did the Moon¬ forming event have on the early Earth? The solar nebula The solar nebula initially separated as a small, slowly rotating, fragment of a molecular cloud. This allowed the formation of a single star surrounded by a rotating disk instead of the more common formation of a double star system, that constitute about 80% of all stars. Probably the disk was noma xi symmetric, which would allow both the inward flow of mass and the outward transfer of © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 angular mometum ( e.g . Boss 1988). Astrophysical evi¬ dence suggests lifetimes of a few million years before the nebula is dispersed. The composition of Cl carbonaceous chondrites is very close to that of the solar photosphere for non-gaseous elements, and so is probably the best estimate available for the composition of the earliest con¬ densed material from the solar nebula. Recent work has resolved the previous outstanding anomaly of distinctly different iron abundances: the new solar values now match the Cl iron abundances (Holweger et al. 1990). Once the Sun has acquired about one third of its present mass, temperature and pressure conditions in the interior allowed H burning to begin. Observations on young stars suggest that the Sun underwent violent T Tauri and FU Orionis outbursts as it proceeded on its evolutionary path toward the main sequence. Strong stel¬ lar winds began to disperse the nebula, thus limiting the ultimate size of the Sun (e.g. Shu et al. 1987). Early vio¬ lent solar activity may sweep away not only the H, He and other gaseous elements, but also ices and volatile elements not condensed or trapped in planetesimals large enough (metre-km size?) to remain in the inner nebula. Loss of volatile Rb relative to refractory Sr and of vola¬ tile Pb relative to refractory U and Th appears to be wide¬ spread in the inner portions of the early nebula. Venus, Earth and Mars all appear to be depleted in volatile ele¬ ments, as shown by their low K/U ratios, and by the U/ Pb and Rb/Sr isotopic systematics in the case of the Earth. This depletion appears to be typical of the entire inner solar system out to perhaps 3 AU, at which dis¬ tance more primitive asteroids begin to dominate the as¬ teroid belt (Bell et al 1989; Gaffey 1990). The age and initial Sr isotopic data from meteorites (Tilton 1988) record a single massive loss of volatile Rb relative to refactory Sr effectively at To, so that this was a nebular¬ wide event rather than being connected with isotopic evolution in individual parent bodies. The depletion of 59 Journal of the Royal Society of Western Australia, 79(1), March 1996 volatile elements must have occurred through physical processes (e.g. sweeping out of fine material by early so¬ lar winds) at relatively low temperatures, so that the nebula was cool at that stage. Chondrules Among the earliest events in the solar system was the formation of chondrules. These mm-size quenched sili¬ cate droplets form one component of chondritic meteorites. The other main constituent is the fine-grained matrix, complementary in composition to the chondrules, being notably enriched in Fe. Thus, the chondrules are not simply remelted matrix, and some segregation of the iron-rich matrix from the iron-poor chondrule precursor material must have occurred prior to the chondrule-forming events. Formation of chondrules occurred in the nebula, rather than in any type of planetesimal or asteroidal en¬ vironment (Taylor et al. 1983). They formed by a fast flash-melting type of process that did not change the composition of the parent material significantly (Lofgren & Russell, 1986; Radomdsky & Hewins, 1988; Hewins 1992). It is not possible to cool molten drops so rapidly in a hot nebula, so the process must have been highly localised in an overall cool environment (Wood 1987). Various scenarios exist to explain these observations of which the nebular flare model appears to be most consis¬ tent with observations (Levy & Araki, 1989), What was the nature of the chondrule precursor material? Metal, sulfide and silicate phases existed already in the early nebula, either as interstellar dust grains, or condensed from nebular gas (Grossman et al. 1988). How was the silicate dust melted preferentially, without involving the metal and sulfide phases to more than a minor extent? Perhaps silicate dust was separated from metal and sul¬ fide, either by differential gravitational settling or mag¬ netically in the case of the metal, or perhaps the silicates stuck together more efficiently (Scott et al. 1988). The planetesimal hypothesis A fundamental question about the origin of the Earth concerns the state of the precursor material prior to planet formation, and the mode of accretion of the plan¬ ets. Did the terrestrial planets form directly from the dispersed dust and gas of the nebula or were they built up brick by brick from planetesimals? The concept that the Earth accreted cold from fine-grained dust and gas was long postulated by Urey fag- Urey 1952). The strik¬ ing chemical heterogeneity of the planets and asteroids (Wasson 1985; Taylor 1988) argues against simple con¬ densation models. Several observations suggest that the inner planets are the end products of a hierarchical accretionary process that first produced a large number of planetesimals which were later accreted to form the larger planets (Safronov 1969; Wetherill 1986). What sort of evidence do we have for these now vanished objects? A major piece of evidence comes from the tilt or inclination of the planets to their axis of rotation. The largest impact is required to account for Uranus. Calculations show that a body the size of the Earth, crashing into that planet would be needed to tip it through 90° (Benz & Cameron 1989). Smaller collisions are needed to account for the tilt of the other planets, but a few very large objects must have been responsible, since the impacts of many small bodies will average out (Benz et al. 1989). The variable rotation rates of the planets may also be a consequence of giant impacts late in their accretional history. Venus, in contrast to the Earth, has a low obliq¬ uity, and is rotating slowly backwards. These properties may result from the accretion of Venus from many small bodies, and from the lack of a giant impact on that planet (Wood 1986; Wood, pers. comm.). It is also usually con¬ sidered that the absence of a primitive terrestrial atmo¬ sphere is due to its early collisional removal. In this in¬ terpretation, Venus has retained a massive atmosphere due to the lack of large atmosphere-removing collisions with that planet. The high metal/silicate ratio of Mer¬ cury is best explained by stripping of much of the silicate mantle during a large collisional event, other hypotheses encountering many difficulties (Benz et al. 1988). Finally, the long-standing problem of the origin of the Moon is resolved by the impact of an already differentiated mas¬ sive (0.14 earth mass) body with the Earth, the material making up the Moon being mostly (>80%) derived from the silicate mantle of the impactor (Benz et al. 1989). How many objects were there and how big were they? Prior to the final sweep-up into the four terrestrial plan¬ ets, Wetherill (1986) calculates that 100 objects of lunar mass, ten with masses exceeding that of Mercury, and several exceeding the mass of Mars should form. He further estimates that perhaps one-third of these objects, which would provide a total of 50-75% of present Earth mass, struck the Earth. The formation of Jupiter Early formation of Jupiter (318 Earth-masses) appears to be required for several reasons. The planet forms early enough to deplete the asteroid belt (which now contains only 5% of lunar mass) in material, and to be responsible for the small mass of Mars (0.11 Earth- mass). This low density region of the nebula seems unlikely to have been a primary feaure of the nebular disk even if the disk was non-axisymmetric. Jupiter must also form before the gas¬ eous components of the nebula were dispersed. Other models such as the giant gaseous protoplanet hypothesis call for the formation of the planets by fragmentation of the primordial solar nebula. Jupiter should be the prime example of such a process. However, there are two prin¬ cipal objections. The moment of inertia data for Jupiter show that it possesses a central core of 15-20 earth masses. At the prevailing conditions in the center of Jupi¬ ter (40 mbars, 20000 K) rock and ice will be miscible with the gaseous components (Stevenson 1985). It will thus not be possible for a core to "rain-out" in the manner of a terrestrial planetary metallic core, where there are both significant density differences and metal-silicate immisci- bility at the temperatures and pressures within the Earth (3.5 mbars, 5000 K; Stevenson 1985). Thus it is necessary to form a massive core first, which can then collect the gas by gravitational attraction. A second objection is that Jupiter does not possess the solar bulk composition that would be expected if Jupiter were derived from a fragment of the primordial nebula; this gas giant has a (rock+ice)/(H+He) ratio about 10 times that of the Sun. Both these properties are readily explicable in terms of the planetesimal hypothesis. 60 Journal of the Royal Society of Western Australia, 79(1), March 1996 However, it is first necessary to form a central core of 15-20 Earth masses, which can then collect the H and He envelope by gravitational attraction. How did such a large nucleus form so rapidly and so early at 5 AU from the Sun? A plausible scenario has been suggested by Lissauer (1987). As early strong solar winds associated with the T Tauri stage of stellar evolution sweep out the uncondensed components from the inner nebula, water ice will condense at about 5 AU at which location the nebular temperature falls below about 160 K. This con¬ densation causes a local increase in particle density of the nebula at such a "'snow line", which will also act as a "cold trap" for other components. Rapid accretion of a large ice and rock core can thus occur at this unique location, and act as a nucleus to collect a hydrogen and helium envelope. The low gas /(ice + rock) ratio in Jupi¬ ter implies that by the time that the core of Jupiter had grown large enough to collect a gaseous envelope, the gaseous nebula was already being dispersed, and that Jupiter simply ran out of material. Accretion of the inner planets in a gas-free environment Once Jupiter has formed, this massive planet domi¬ nates subsequent evolution of the solar system. After a few million years, the gas is gone and the ices and other volatiles have been swept away, so that the inner planets accreted from the left-over rocky debris. Depletion of ma¬ terial in the asteroid belt occurs both from accretion of material to Jupiter, and subsequent pumping up of ec¬ centricities and inclinations of the asteroids remaining, so that the survivors have been unable to collect them¬ selves into a planet. Others are tossed out of the system entirely. The asteroid belt appears to have existed from the earliest times. Thus the belt was not a very good quarry from which to obtain material for the inner plan¬ ets. The accretion of Mars took place in a zone depleted in planetesimals from the same cause (early formation of Jupiter) and this region, at 1.5 AU again does not seem capable of supplying much material for Venus or the Earth. Differentiation of precursor planetesimals What was the history of the planetesimals prior to their incorporation into the inner planets? Some of the largest, the size of Mars, would have made respectable planets in their own right if fate had taken a different course. Were they already differentiated into silicate mantles and metallic cores before they came to a violent end as they were swept up into Earth or Venus? Based on evidence from meteorites, even some rela¬ tively small planetesimals underwent internal differen¬ tiation into metallic cores and silicate mantles within a few million years of Tu (4570 my). The larger planetesi¬ mals had already gone through a melting episode, with silicate mantle and metallic core formation, before they were accreted by the inner planets (Gaffey 1990; Taylor & Norman 1990). Such bodies of course may have been broken up by collisions and reaccreted in differing pro¬ portions of metal and silicate fractions, so that much di¬ versity of composition among the accreting bodies can be expected. This question of heat supply for early planetesimal melting and metamorphism is essentially unresolved. Two principal mechanisms are currently discussed. If 2hAl (t^ = 730 000 years) was present in the early solar system (Podosek & Swindle 1988), it could have consti¬ tuted an important heat source. The second possibility is by inductive heating during the early intense T Tauri and FU Orionis stages of solar activity. Both of these mechanisms encounter difficulty and early planetesimal heating may be the result of processes not presently un¬ derstood (J A Wood, pers. comm.). A crucial question for the terrestrial planets is the width of the feeding zones from which they accumulated (Wetherill 1985). The limited data for Venus show simi¬ lar K/U ratios to the Earth of about 10\ This, coupled with the similar uncompressed density (about 4.0 g cm 3) for the two planets, their similar size and their small separation of about 0.3 AU suggests that they accreted from a similar suite of planetesimals. Mars is less dense (uncompressed density 3.75 g cm 3) and has a high obliq¬ uity and fast rotation rate, indicative of collisions with large objects. It is more volatile- rich than either Earth or Venus, having a K/U ratio of about 1.5 103. Thus Mars, about equidistant from the Earth and the main asteroid belt, appears to be distinct from both, suggesting that there was very little mixing within the nebula over dis¬ tances greater than about 0.5 AU The survival of zoning in the asteroidal belt also points toward rather limited mixing. Other evidence includes the rarity of xenoliths of one class of meteorite in another. The great diversity in oxygen isotopic compositions (Thiemens 1988) including that of the chondrules (Grossman et al. 1988) is also strongly indicative of very limited mixing. In addition, the various classes of chondrites do not show simple chemical interrelationships which might indicate a helio¬ centric variation in composition. The general failure to identify specific classes of meteorites as building blocks for the terrestrial planets (e.g. Taylor, 1988) suggests that the inner planets accreted from rather narrow zones in the nebula, without incorporating much material from the location of the present asteroid belt. The terrestrial Mg/Si ratio The upper mantle of the Earth is depleted in Si and has an enhanced Mg/Si ratio relative to that of the primi¬ tive solar nebula. The bulk Mg/Si ratio of the Earth is uncertain since we do not know the Mg /Si ratio of the lower mantle. The debate over this question is unre¬ solved (e.g. Anderson 1989). Recent suggestions that mantle plumes, responsible for hot-spot volcanism, are derived from the core-mantle boundary (Griffiths & Campbell 1991; Sleep 1992), imply that the whole mantle is involved and that there is significant mixing between upper and lower mantle. If the lower mantle of the Earth has the same Mg/Si ratio as the upper mantle, then the implications for the accretion of the Earth are consider¬ able. In this event, the Earth accreted from a set of plan¬ etesimals with non-CI Mg/Si ratios. The variation in Mg/Si in chondrites covers such a wide range that the existence of planetesimals with higher Mg/Si ratios seems possible. This would imply a very large reservoir of planetesimals at about one AU with Mg/Si ratios sig¬ nificantly higher than solar. Core-mantle relationships The highly siderophile elements would have been effi¬ ciently extracted into the metal core under equilibrium 61 journal of the Royal Society of Western Australia, 79(1), March 1996 conditions. However the present upper mantle was ap¬ parently never in equilibrium with the core, for the abun¬ dances of Re, Au, Ni, Co and the platinum group ele¬ ments (PGE= Ru, Rh, Pd, Os, Ir, Pt), although low, are higher than predicted (Arculus & Delano 1981; Delano 1986; Newsom & Palme 1984; Newsom 1986). Late accre¬ tion of Cl planetesimals rich in PGE is a common expla¬ nation for their over-abundance in the upper mantle. The addition of the metallic core of the impactor responsible, in the single impact hypothesis, for the origin of the Moon (Benz et at 1989) is another possible source of ma¬ terial. A cometary influx might be an equally viable source, although the high impact velocities of comets de¬ rived from the outer solar system may cause removal rather than addition of material. The 'predestination' scenario (Taylor 1983; Taylor & Norman 1985; Murthy & Karato in press) in which the terrestrial planets accrete from planetesimals which were already mostly differentiated into metallic, silicate and sulfide phases implies little further reaction between metal and silicate once these bodies accreted to the Earth. In this scenario the core mantle relationships were mostly established at low, and not high pressures. A further consequence may be noted. The metallic core of the Earth contains about 10% of a light element. The two current contenders are oxygen and sulfur. Al¬ though meteorites are not a perfect analogue for the ter¬ restrial precursor planetesimals, they do tell us that el¬ emental and mineralogical fractionation was endemic in the early nebula. If silicate, sulfide and metal phases, formed under low-pressure equilibrium conditions, were already present in the accreting planetesimals, separa¬ tion of these phases may occur concomitantly with accre¬ tion and thus there may be little high-pressure equilibra¬ tion between core and mantle in the Earth. Thus it seems unlikely that oxygen entered the core, since this requires megabar pressures. Sulfur then becomes the most viable candidate for the light element in the earth's core. Since metal-sulfide-silicate equilibria was accomplished pre¬ dominantly at low pressures in precursor planetesimals, troilite will be the main source of sulfur. Late Veneers A number of possible effects of late additions to the Earth have been proposed. Thus comets are often in¬ voked as a source of water (e.g. Chyba 1987). A cometary source may also account for the difference in the atmo¬ spheric abundances of the rare gases in the Earth, Venus and Mars (Owen et ah 1992). This concept is attractive since, in the scenario developed here, the inner solar sys¬ tem is depleted in water and other volatiles. Further¬ more, the terrestrial water budget, although uncertain, probably constitutes less than 500 ppm of the mass of the Earth (Bell & Rossman 1992; Thompson 1992). This is less than V1000 of the water budget in the primitive nebula and could readily be suppplied by a few large comets. Such stochastic processes also rather conve¬ niently account for the differences among the terrestrial planets: the vexed question of the missing water on Ve¬ nus is simplified if that planet never had any to begin with. Comets, however, may be a fickle source of atmo¬ spheres and hydrospheres, since they impact with rela¬ tively high velocities and thus may remove as much ma¬ terial as they contribute (e.g. Melosh & Vickery 1989). There are various other unresolved problems with the concept of late veneers. The Moon shows no evidence of such events and the Moon remains "bone-dry". Mercury A large impact is probably responsible for the strange fact that Mercury has such a small rocky mantle and such a large iron core, and an inclined orbit so close to the sun. Two explanations are current. The first pro¬ poses that that the silicate was boiled away in some early high temperature event, connected with early solar activ¬ ity (the surface temperature on the present sunlit side of Mercury is 425 C, and hot enough to melt lead). How¬ ever, extremely high temperatures of several thousand degrees are required to boil off the rocky mantle. The alternative explanation, is that Mercury was struck by a body about V6 of its mass at a late stage in its accretion. The collision fragmented the planet with most of the sili¬ cate lost to space but the iron core surviving to reaccrete with a depleted silicate mantle (Benz et ah 1988). If Mer¬ cury has a plagioclase-rich crust analogous to the lunar highlands, then it is likely to be depleted in the more volatile elements, since flotation of such a crust in a magma ocean requires a water content less than 0.1% (Walker & Hays 1977). Attempts to secure a K/U ratio for Mercury, which would shed some light on these in¬ teresting problems, should be accorded a high priority. The origin of the Moon The broad aspects of lunar evolution are well under¬ stood. The moon was partially or wholly melted at, or shortly after, accretion. This vast mass of molten silicate has been termed the "magma ocean" and a high tem¬ perature and rapid mode of origin for the moon is required to account for it. The crystallisation of the magma ocean is understood in principle (e.g. Taylor 1982; Warren 1985). Feldspar was an early phase to crystallise. It floated, due to the low density of the feldspar crystals and the anhydrous nature of the silicate melt, and formed a thick feldspathic crust by 4440 my. Convection during cooling may have swept "rockbergs" of feldspar together, accounting for the variations in crustal thick¬ ness. A small lunar iron core about 2-5% by volume formed in the centre. This sequestered the siderophile elements. The lunar mantle was fully crystallised by about 4400 my, and resulted in a zoned silicate mineral¬ ogy, from which the mare basalts were derived much later by partial melting. This cumulate hypothesis for the source region of mare basalts is well established (e.g. Tay¬ lor & Jakes 1974; Fujimaki & Tatsumoto 1984). As the silicate minerals crystallised, those trace elements which were excluded from their crystal lattices were concen¬ trated in the residual melt. The final stage of magma ocean evolution was the intrusion of this residual liquid into the feldspathic highland crust. The fluid was en¬ riched in elements such as Th, U, Zr, Hf, Nb, K, REE, P (from which the acronym KREEP has been coined) and is responsible for the extraordinary near surface abundance of elements such as K, U, Th, and REE, which may be concentrated by factors of several hundred relative to bulk moon or primitive nebula values. It pervaded the crust, with which it was intimately mixed by the continu¬ ing meteoritic bombardment. The final event in crustal evolution was the intrusion of an Mg and KREEP-rich suite of rocks, produced perhaps by sub-crustal melting 62 Journal of the Royal Society of Western Australia, 79(1), March 1996 induced by the impacts of giant planetesimals. Bulk moon models which contain more than 5% A1203 pro¬ vide the best match to the seismic velocity profile, im¬ plying that the moon is enriched in refractory elements relative both to the Earth and to primitive solar nebula levels. This conclusion has been confirmed by data from the Clementine mission (Lucey et al. 1995) Hypotheses of lunar origin The major models for the origin of the Moon can be grouped into five separate categories, which include; (a) capture from an independent orbit; (b) fission from a rapidly rotating Earth; (c) formation as a double planet; (d) disintegration of incoming planetesimals; and (e) Earth impact by a Mars-sized planetesimal. All fail to account for the unique nature of the Earth- moon system except the last. This process accounts for the high angular momentum (3.45 1041 rad g cm2 sec'1) of the Earth-Moon system and the non-equatorial lunar or¬ bit as well as providing extreme temperature conditions which can produce an initially molten Moon and the bone-dry features of lunar geochemistry. Computer simulations of the giant impact hypothesis under condi¬ tions that form a lunar mass depleted in metallic iron in terrestrial orbit clearly indicate that it is mostly the mate¬ rial from the silicate mantle of the impactor that finishes up in the Moon (e.g. Cameron & Benz 1991). This con¬ clusion is reinforced by the geochemical problem of the failure to match Earth mantle and lunar compositions for a number of crucial elements ( e.g . Taylor 1986a, b; Newsom & Taylor 1989). The similarity in oxygen isotopes between the Earth and Moon indicate derivation of both the Earth and the impactor from the same region of the nebula, thus ex¬ cluding models that derive the impactor from the outer reaches of the solar system. The similarity in 53Cr/52Cr ratios (53Cr is derived in part from short-lived n1Mn) be¬ tween the Earth and the Moon and their contrast with higher meteoritic values (Lugmair & Maclsaac 1995) car¬ ries the same implication of derivation of lunar material from around 1 AU. A third constraint is the relatively low collision velocity (Benz et al 1989; Cameron & Benz 1991) required to produce a Moon-sized body, which again restricts the impactor to be a nearby object. If the material in the Moon is derived from the impactor, then that body had a lower Rb/Cs ratio than the Earth. The primitive lunar initial ^Sr/^Sr ratios indicate that the im¬ pactor must have been depleted in Rb relative to Sr very close to To. Current models assume that core-mantle separation occurred in both the impactor and the Earth before im¬ pact, to account for the lunar siderophile element abun¬ dances and the lunar depletion in iron (13% FeO) relative to primordial solar nebula volatile-free abundance levels (as shown by the Cl meteorites) of 36%. The abundance of FeO in the mantle of the impactor must however have been greater than that of the terrestrial mantle (8% FeO), since the bulk Moon contains a much higher abundance. Mars, in contrast, has a mantle FeO content of 18%. Effects on the Earth of the Moon-forming impact The important consequence of the single giant impact event for the Earth was that the energies involved are sufficient to melt the Earth However, such melting is probably inevitable if the Earth was accreted from a hier¬ archical suite of planetesimals, regardless of w'hether the Moon-forming event occurred. Any primitive atmo¬ sphere is removed, which probably accounts for the very much lower 36 Ar content (by about two orders of magni¬ tude) of the terrestrial atmosphere compared with that of Venus. The lack of geochemical evidence for early differentia¬ tion of the Earth (e.g. McFarlane & Drake 1990) analo¬ gous to that shown by small-scale terrestrial layered in¬ trusions (e.g. Skaergaard, Stillwater) or by the Moon may be due to the scale of the event. Thus a molten terrestrial mantle may be turbulent, and crystals may not have had the opportunity to settle, thus precluding large-scale frac¬ tionation (Tonks & Melosh, 1990) In addition to the accretion of the impactor's core, about 10% of the mass of the Earth's mantle is added from the impactor's mantle. The models of Benz et al. (1989) indicate that most of the metal core ends up in the Earth, with the metal penetrating the mantle and ending up wrapped about the Earth's core. Such an event would not disturb siderophile abundance patterns already present in the Earth's mantle. However, a significant amount of material from the impactor's core, enriched in siderophile elements, will probably be vaporized and re¬ distributed into the mantle. Conclusions 1. Depletion of volatile elements in the inner nebula oc¬ curred effectively at To before the chondrules were formed and affected the solar nebula out to about 3 AU The probable mechanism was dispersal of uncondensed volatiles by early strong stellar winds during the T Tauri stage of solar evolution. 2. Jupiter formed early before the gas component in the nebula was totally depleted. 3. Accretion of the Earth, inner planets and the asteroid belt took place in a gas-free environment in the inner solar system following the formation of Jupiter. 4. The terrestrial planets were built from precursor plan¬ etesimals that had survived the clearing of the inner solar nebula. The larger ones had already formed me¬ tallic cores and silicate mantles and had already expe¬ rienced at least one episode of melting and differentia¬ tion. 5. Because this metal-sulfide-silicate fractionation oc¬ curred largely at low pressures, the geochemistry of the core and mantle may instead be dominated by the low-pressure equilibria established in the precursor planetesimals. Sulfur becomes a viable candidate for the light element in the core 6. Only limited mixing occurred in the inner nebula dur¬ ing planetary formation, with accretionary zones per¬ haps 0.3 AU wide. Very little material from the aster¬ oid belt was incorporated in the Earth. 7. The Moon formed as the result of a single giant im¬ pact of a Mars-sized body with the Earth. Most of the 63 Journal of the Royal Society of Western Australia, 79(1), March 1996 material in the Moon came from the mantle of the impactor. 8. The Earth was melted either as a result of the Moon¬ forming event, or as a consequence of its accretion from a hierarchical sequence of planetesimals. References Anderson D L 1989 Theory of the Earth. Blackwell Scientific Publications, Cambridge. Arculus R J & Delano ] W 1981 Siderophile element abundances in the upper mantle: evidence for a sulfide signature and equilibrium with the core. Geochimica et Cosmochimica Acta 45:1331-1343. Bell D R and Rossman G R 1992 Water in Earth's mantle: The role of normally anhydrous minerals. Science 255:1391-1397. Bell J F, Davis D R, Hartmann W K & Gaffey M ] 1989 Aster¬ oids: The big picture. In: Asteroids II (eds R P Binzel, T Gehrels & M S Matthews). Arizona University Press, Tucson, 921-945. Benz W & Cameron A G W 1989 Tilting Uranus in a giant impact abstract. Bulletin of the American Astronomical Soci¬ ety. 21:916. Benz W, Cameron A G W & Melosh H J 1988 The origin of the Moon: Further studies of the giant impact abstract. Lunar and Planetary Science 19:61-62. Benz W, Cameron A G W & Melosh H j 1989 The origin of the Moon and the single impact hypothesis III. Icarus 81:113- 131. Boss A P 1988 Protostellar formation in rotating interstellar clouds VII. Opacity and fragmentation. Astrophysical Jour¬ nal 331:370-376. Cameron A G W & Benz W 1991 The origin of the Moon and the single impact hypothesis IV. Icarus 92:204-216. Chyba C F 1987 The cometary contribution to primitive oceans. Nature 330:632-635. Delano J W 1986 Abundances of cobalt, nickel, and volatiles in the silicate portion of the Moon, in Origin of the Moon (eds. W K Hartmann, R J Phillips &l G J Taylor). Lunar Planetary Institute, Houston, 231-247. Fujimaki H & Tatsumoto M 1984 Lu-Hf constraints on the evo¬ lution of lunar basalts. Journal of Geophysical Research 89:B445-B458. Gaffey M J 1990 Thermal history of the asteroid belt: Implica¬ tions for the accretion of the terrestrial planets. In: Origin of the Earth (eds H E Newsom & J H Jones). Oxford University Press, 17-28. Griffiths R W & Campbell I H 1991 On the dynamics of long- lived plume conduits in the convecting mantle. Earth and Planetary Science Letters 103:214-227. Grossman J N 1988 Formation of chond rules, in Meteorites and the Early Solar System (eds. J F Kerridge & M S Matthews). University Arizona Press, Tucson, 680-696 Hewins R H 1992 Chondrule formation. Meteoritics 27:232-233. Holweger H, Heise C & Koch M 1990 The abundance of iron in the Sun derived from the photospheric Fell lines. Astronomy and Astrophysics 232:510-515. Levy E H & Araki S 1989 Magnetic reconnection flares in the protoplanetary nebula and the possible origin of meteorite chondrules. Icarus 81:74-91. Lissauer J 1987 Timescales for planetary accretion and the struc¬ ture of the protoplanetary disk. Icarus 69:249-265. Lofgren G & Russell W T 1986 Dynamic crystallization of chon¬ drule melts of porphyrtic and radial pyroxene composition. Geochimica et Cosmochimica Acta 50:1715-1726 Lucey P G, Taylor G J & Malaret E 1995 The abundance and distribution of iron on the Moon: Implications for crustal differentiation, structure and the origin of the Moon. Science 268:1150-1153. Lugmair G W & Maclsaac Ch 1995 Radial heterogeneity of 53Mn in the early solar system? Lunar and Planetary Science 26:879-880. MacFarlane, E. A. & Drake, M. J. 1990 Element partitioning and the early thermal history of the earth, in Origin of the Earth (eds: H E Newsom & j H Jones). Oxford University Press, Oxford, 135-150. Melosh H J & Vickery A M 1989 Impact erosion of the primor¬ dial atmosphere of Mars. Nature 338:487-489. Murthy V R & Karato S 1996 Core forandmation & chemical equilibrium in the Earth - II: Chemical consequences for the mantle and the core. Physics of the Earth and Planetary Inte¬ riors: In press. Newsom H E 1986 Constraints on the origin of the Moon from the abundance of molybdenum and other siderophile ele¬ ments, in Origin of the Moon (eds. W K Hartmann, R J Phillips & G J Taylor). Lunar Planetary Institute, Houston, 203-229. Newsom H E & Palme H 1984 The depletion of siderophile elements in the Earth's mantle: New evidence from molyb¬ denum and tungsten. Earth and Planetary Science Letters 69:354-364. Newsom H E & Taylor S R 1989 Geochemical implications of the formation of the Moon by a single giant impact. Nature 338:29-34. Owen T, Bar-Nun A & Kleinfeld I 1992 Possible cometary origin of heavy noble gases in the atmospheres of Venus, Earth and Mars. Nature 358:43-46. Podosek F A & Swindle T D 1988 Extinct radionuclides. In: Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). University Arizona Press, Tucson, 1093-1113. Radomsky P M & Hewins R H 1988 Chondrule texture/compo¬ sition relations revisited: Constraints on the thermal condi¬ tions in the chondrule forming region. Meteoritics 23:297- 298. Safronov V 1969 Evolution of the protoplanetary cloud and for¬ mation of the Earth and planets. NASA TT F-677 1972. Scott E R D, Barber D J, Alexander C M, Hutchison R & Peck J A 1988 Primitive material sunn v in g in chondrites: matrix. In: Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). University Arizona Press, Tucson, 718-745. Shu F H, Adams FC & Lizano S 1987 Star formation in molecu¬ lar clouds: Obsen'ations and theory. Annual Review of As¬ tronomy and Astrophysics 25:21-81. Sleep N J 1992 Hotspot volcanism and mantle plumes. Annual Review of Earth and Planetary Sciences 20:19-43. Stevenson D J 1985 Cosmochemistry and structure of the giant planets and their satellites. Icarus 62:4-15. Taylor G J, Scott E R D & Keil K 1983 Cosmic setting for chon¬ drule formation. In: Chondrules and their Origins (ed E A King). Lunar and Planetary Institute, Houston, Texas, 262- 278. Taylor S R 1982 Planetary science: a lunar perspective. Lunar and Planetary Institute, Houston. Taylor S R 1983 Element fractionation in the solar nebula and planetary compositions: A "predestination" scenario ab¬ stract. Lunar and Planetary Science 14:779-780. Taylor S R 1986a The origin of the Moon: Geochemical consider¬ ations. In: Origin of the Moon (eds W K Hartmann, R J Phillips & G J Taylor). Lunar and Planetary Institute, Hous¬ ton, 125-144. Taylor S R 1986b Cutting the Gordian Knot: Lunar composi¬ tions and Mars-sized impactors. Lunar and Planetary Science 17:881-882. Taylor S R 1988 Planetary compositions. In: Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). Uni¬ versity of Arizona Press, Tucson. 64 Journal of the Royal Society of Western Australia, 79(1), March 1996 Taylor S R 1992 Solar System Evolution: A New Perspective. Lunar and Planetary Institute, Houston and Cambridge Uni¬ versity Press, Cambridge. Taylor S R & Jakes P 1974 The geochemical evolution of the moon. Proceedings of Lunar Science Conference 5:1287- 1305. Taylor S R & Norman M D 1990 Accretion of differentiated planetesimals to the Earth. In: Origin of the Earth (eds H E Newsom & J H Jones). Oxford University Press, Oxford, 29- 43. Thiemens M H 1988 Heterogeneity in the nebula: Evidence from stable isotopes, in Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). University of Arizona Press, Tucson, 899-923. Thompson A B 1992 Water in the Earth's upper mantle. Nature 358:295-302. Tilton G W 1988 Age of the solar system, in Meteorites and the Early Solar System (eds J F Kerridge & M S Matthews). Uni¬ versity of Arizona Press, Tucson, 259-275. Tonks W B & Melosh H J 1990 The physics of crystal settling and suspension in a turbulent magma ocean. In: Origin of the Earth (eds H E Newsom & J H Jones). Oxford University Press, Oxford, 151-174. Urey H C 1952 The Planets, Their Origin and Development. Yale University Press, Yale. Walker D & J F Hays 1977 Plagioclase flotation and lunar crust formation. Geology 5:425-428. Warren P H 1985 The magma ocean concept and lunar evolu¬ tion. Annual Review of Earth and Planetary Science 13:201- 240. Wasson ] T 1985 Meteorites. Freeman, New York. Wetherill G W 1985 Occurrence of giant impacts during the growth of the terrestrial planets. Science 228:877-879. Wetherill G W 1986 Accumulation of the terrestrial planets and implications concerning lunar origin. In: Origin of the Moon (eds W K Hartmann, R J Phillips & G J Taylor). Lunar and Planetary Institute, Houston, 519-550. Wood J A 1986 Moon over Mauna Loa: A review of hypotheses of formation of Earth's Moon. In: Origin of the Moon (eds W K Hartmann, R J Phillips & G j Taylor). Lunar and Planetary Institute, Houston, Texas, 17-53. Wood J A 1987 Was chondritic material formed during large- scale, protracted nebular evolution or by transient local events in the nebula? Lunar and Planetary Science 18: 1100- 1101. 65 Journal of the Royal Society of Western Australia, 79:67-71, 1996 Cosmogenic noble gases in silicate inclusions of iron meteorites: Effects of bulk composition on elemental production rates. O Jentsch & L Schultz Max-Planck-Institut fur Chemie, 55020 Mainz Germany Abstract The influence of bulk chemical composition on elemental production rates of neon has been studied experimentally by investigating mineral samples separated from the iron meteorites Four Corners, Landes, Linwood and Zagora. The neon in these samples is purely cosmogenic and its isotopic composition indicates an enhanced production of 21Ne. This is due to the production of additional secondary neutrons caused by a larger multiplicity in meteorites of high-Z bulk chemistry. Introduction Meteoroids as meter-sized bodies are irradiated in space by high-energy cosmic-ray particles that penetrate these bodies and interact with its matter. Bauer (1947) and Huntley (1948) independently proposed that mea¬ surable amounts of He should be produced in iron mete¬ orites. Paneth et al. (1952) showed that this was indeed true and that cosmic-ray produced 3He/4He ratio is in the range 0.2 to 0.3 and were thus very different from atmospheric He, Soon thereafter, also cosmogenic Ne and Ar isotopes were discovered (Reasbeck & Mayne 1955; Gentner & Zahringer 1957). Since than, cosmic-ray pro¬ duced ("cosmogonic") stable or radioactive reaction products have been studied in great detail to investigate the irradiation history of extraterrestrial matter or the nature of cosmic rays in the past (see e.g. Anders 1962; Lai 1972; Reedy et al 1983; Vogt et al 1990). The production rate of cosmogenic nuclides is depen¬ dent on the chemical composition of the target material, the properties of the incident particle irradiation, and the shielding conditions (pre-atmospheric size of the meteor¬ oid and sample's location within the meteoroid). The ef¬ fects of shielding are mainly caused by the fact that in nuclear reactions initiated by primary cosmic rays sec¬ ondary particles are also produced. These secondary par¬ ticles, mostly neutrons of smaller energy, cause further nuclear reactions. Elemental or isotopic nuclide ratios produced by these low-energy particles are different from those of the primary high-energy irradiation. Thus, the shielding of a sample is characterized by elemental or isotopic ratios of specific cosmogenic nuclides produced in different percentages by primary and secondary par¬ ticles. For example, for shielding corrections of produc¬ tion rates of stony meteorites, cosmogenic 22Ne/2lNe is commonly used {e.g. Eugster 1988), or for iron meteor¬ ites, the 4He/38Ar ratio {e.g. Voshage & Feldmann 1978) is a measure for shielding. The production rates of cosmogenic nuclides depend twofold on the bulk chemical composition of the meteor- © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology/ Perth, 1995 ite. First, the production of a particular nuclide is given by the concentration-weighted sum of the individual el¬ emental production rates from each target element in the meteorite and, secondly, the elemental production rates themselves depend on it since spectra and intensities of secondary particles inside the meteoroid are influenced by the mean mass number and the mean atomic number of the bulk meteorite. In particular, the multiplicity for the production of secondary neutrons is greater in metal- rich meteorites compared to stony meteorites. The latter ("matrix") effect was first suggested by Begemann & Schultz (1988). This idea is based on mea¬ surements of cosmogenic noble gases in metal and sili¬ cates of mesosiderites (Begemann et al. 1976). These au¬ thors had observed that the concentration of cosmogenic -^Ar produced from Ca in silicates (mainly by secondary neutrons) is larger than expected from the 38 A r produced in metal by the primary irradiation. Begemann and Schultz (1988) show that the production rate ratio of 38Ar from Ca to that of 3SAr from Fe is correlated with the total bulk metal content of mesosiderites. Also, the 22Ne/ 21 Ne ratio in these silicate samples is relatively low. This is explained by an enhanced production of 21 Ne via sec¬ ondary neutrons and the reaction 24Mg(n,a)2,Ne. The matrix effect was also treated theoretically by Michel et al (1990) and Masarik & Reedy (1994) using model calculations which combine spectra of primary and secondary particles derived from Monte Carlo calcu¬ lations of intra- and internuclear cascades with thin-tar- get cross sections of the underlying nuclear reactions. These calculations show, for example, that the matrix ef¬ fect is most pronounced in the production rate of 21 Ne from Mg, and compared to stony meteorites the 21 Ne pro¬ duction is enhanced by a factor of up to two in silicate inclusions of iron meteorites. We report here measurements of concentration and isotopic composition of neon in silicates separated from three IA iron meteorites with silicate inclusions (Landes, Linwood and Zagora) and from the IB iron Four Cor¬ ners. This work was carried out to obtain information on the matrix effect in iron meteorites which are more metal- rich than the previously investigated mesosiderites. 67 Journal of the Royal Society of Western Australia, 79(1), March 1996 Experimental Methods and Results Carbon and silicate phases of Landes and Zagora were separated from the meteorite by anodic dissolution of the FeNi. After several ultrasonic treatments and quenching steps with liquid N2, a further separation of different grain size fractions took place using density and mag¬ netic separation techniques. The Four Corners and Linwood separates were obtained from R S Clarke (Washington DC) and are leftovers from investigations by G P Merill and E P Henderson. The concentration of the elements Mg, Al, Si, K, Ca, Ti and Fe of selected separates was determined using x-ray fluorescence (XRF) techniques. The isotopic composition and concentrations of noble gases were measured using apparatus and experimental procedures as described by Schultz et al (1991). Results are given in Table 1 together with estimates of the mineral composition of the samples calculated from its chemical composition. Figure 1 shows isotopic ratios of neon in a 3-isotope plot. Cosmogenic ^Ne/^Ne ratios in ordinary chondrites are rather independent of shielding (20Ne/22Ne = 0.84 ± 0.02) and are found in a narrow area (Fig 1; hatched, marked "Chondrites"). If the measured neon is a mixture of cosmogenic gas and a trapped component (solar or planetary gas, atmospheric contamination) the data points would be shifted to higher 20Ne/22Ne values. The low 20Ne/22Ne values indicate, however, that no signifi¬ cant abundance of trapped Ne is present in spite of trapped Ar, Kr and Xe that is commonly observed in silicate inclusions of iron meteorites (Hintenberger et al 1969; Niemeyer 1979; Crabb 1983; Jentsch & Schultz 1987). The displacement of many data points to higher 21Ne/22Ne values is discussed below. The three feldspar-rich samples are characterized by smaller 20Ne/22Ne and 21Ne/22Ne ratios. This is due to the relatively high concentration of Na in these samples. So¬ dium is an element that has a high production rate for 22Ne compared to that of the other two neon isotopes. Smith &z Huneke (1975) observed 20Ne/22Ne ratios < 0.7 in plagioclase separated from ordinary chondrites. Our feldspar measurement of Zagora is very similar to this value. The two Landes samples with a smaller percent¬ age of feldspar (and thus Na) are located on the mixing line between the "pure" feldspar neon ratio and "nor¬ mal" cosmogenic neon. Unfortunately, the concentration of Na was not measured directly in these samples be¬ cause NaN03was used to make the tablets for XRF-mea- surements. Table 1. Chemical and mineralogical information on the investigated separates as well as on their concentration and isotopic composition of neon. Meteorite Mg Al Si K Ca Ti Fe Mg/(Mg+Al+Si) Mineral 20Ne 21 Ne 22Ne 22Ne/2,Ne and sample # [wt %] [wt.% ratio] composition [in 10Am3STP/el Opx Ol Cpx Fsp [wt.%] Four Comers 401 19.5 1.22 26.5 0.075 2.50 0.13 3.8 0.414 67 12 12 9 348.1 407.9 394.6 0.967 Landes 1364 22.4 0.11 26.0 0.022 1.20 0.12 4.0 0.462 81 14 5 - 78.5 89.4 91.3 1.021 1344 21.9 0.35 25.7 0.021 1.59 0.13 4.3 0.456 79 14 7 - 75.2 85.5 88.0 1.029 1363 26.9 0.09 23.0 0.016 .78 0.07 3.8 0.539 42 54 4 - 84.8 97.7 98.5 1.008 1355 4.6 9.90 31.2 0.36 1.95 0.07 2.9 0.100 22 - - 78 15.2 17.1 19.9 1.161 1 1345 22.2 0.37 25.6 0.027 1.42 0.13 4.3 0.460 78 15 7 - 79.7 91.2 92.8 1.017 1 1365 23.1 0.23 25.3 0.022 1.14 0.17 4.1 0.474 74 21 5 - 60.5 68.9 70.3 1.021 1 1366 26.0 0.11 23.7 0.015 0.69 0.08 3.7 0.522 53 44 3 - 89.9 103.2 103.8 1.006 II353 5.4 9.40 31.2 0.33 1.98 0.07 2.6 0.117 26 - - 74 20.2 22.7 26.4 1.162 III343 21.4 0.44 25.1 0.021 2.38 0.13 5.0 0.456 70 17 13 - 75.3 85.8 87.6 1.021 III347 25.2 0.21 23.7 0.010 1.49 0.10 3.9 0.513 49 43 8 - 86.5 99.8 100.9 1.011 Linwood 425 11.5 3.49 30.8 0.21 3.23 0.14 2.7 0.252 43 - 13 44 32.3 36.6 36.1 0.987 Zagora 1329 23.0 0.22 25.4 0.015 1.12 0.13 4.1 0.473 76 19 5 - 91.6 108.2 104.3 0.964 1417 24.4 0.15 24.6 0.015 0.99 0.12 3.9 0.496 65 31 4 - 118.0 139.1 133.5 0.960 1328 22.2 0.24 25.5 0.023 1.21 0.14 4.8 0.464 63 32 5 - 113.7 133.7 129.2 0.966 II322 23.5 0.07 24.3 0.013 0.81 0.11 5.9 0.490 70 27 3 - 110.8 130.2 125.4 0.963 1 1320 24.3 0.09 24.1 0.017 0.86 0.11 5.1 0.501 64 33 3 - 110.1 129.9 124.0 0.955 II319 23.7 - 23.3 0.026 1.05 0.10 7.1 0.504 59 36 5 - 112.6 132.6 127.5 0.962 1 1325 25.3 0.08 22.8 0.020 2.54 0.09 4.3 0.525 33 52 15 - 119.6 141.9 135.1 0.952 III386 0.94 11.8 32.8 0.63 1.82 0.05 2.1 0.021 9 - - 91 20.8 23.1 35.2 1.523 68 Journal of the Royal Society of Western Australia, 79(1), March 1996 1.5 1.4 1.3 1.2 * 1.1 CN Z 0.8 r O 0.7 - 0.6 - • 0.5 - 1 - 1 - 1 - 1 - * 0.6 0.7 0.8 0.9 1.0 1.1 21Ne/22Ne Figure 1. A three-isotope plot of neon showing the position of cosmogenic neon in chondrites (hatched area) and those of ana¬ lyzed minerals separated from iron meteorites. Trapped compo¬ nents or atmospheric contamination, characterized by high 2(,Ne/22Ne and low 21Ne/22Ne values, are not present in appre¬ ciable amounts. Data points below the chondritic values are feldspar-rich samples with higher concentrations of sodium that cause higher production rates of 22Ne. • Zagora o Four Comers Linwood o Landes Chondrites o Discussion An enhanced flux of secondary neutrons in FeNi-rich meteorites will produce lower cosmogenic 22Ne/2lNe due to the reaction 24Mg(n,a)21Ne. To show this effect one must refer to a similar Mg concentration between normal chondritic samples and silicate inclusions of iron meteor¬ ites. In addition, the minimum values of cosmogenic 22Ne/21Ne for chondrites must be known. For stony meteorites the shielding effect has been studied on individual samples as well as on samples taken from drill cores of meteorites. Eberhardt et al. (1966) found a relation between cosmogenic 3He/21Ne and 22Ne/2lNe that is widely used for shielding correc¬ tion of production rates. This relation - originally ob¬ served in samples of different chondrites - is shown in Figure 2 together with a number of measurements for samples taken from defined locations in larger meteor¬ ites or , as in the case of Kokubunji, taken from different samples of a meteorite shower. A lower limit of observed cosmogenic 2*Ne/21Ne ratios is about 1.06. This value was also confirmed in measurements of the largest stony me¬ teorite analyzed so far, the Jilin chondrite (Begemann et al. 1995). These data are shown in Figure 3. Jilin has had a two-stage exposure history, a short 4tt irradiation as a meteoroid and an extended one in 2ir geometry (Begemann et al. 1985). All data have "NeA'Ne > 1.06 but the measured concentrations of 21 Ne vary by a factor of about 6. They do not follow the expected variation of the production rate with shielding (e.g. Eugster 1988) as indicated by the heavy curve in Figure 3. The Jilin data clearly show that the lower limit for 22Ne/21Ne in chon¬ drites is about 1.06. The calculation of exposure ages using 22Ne/21Ne as a shielding indicator may yield a lower limit only if cosmogenic 22Ne/21Ne is less than about 1.08. 22Ne/21Ne Figure 2, Cosmogenic TTe/21Ne versus 22Ne/21Ne in chondrites. The Bern-Line (Eberhardt et al. 1966) is obtained from measure¬ ments of different meteorites. Other lines are measurements of samples taken from defined locations within one meteorite or from individual specimens of a meteorite shower (adapted from Loeken et al. 1992; references are given therein). 22Ne/21Ne Figure 3. Concentration of cosmogenic 21Ne as function of cos¬ mogenic 22Ne/21Ne in samples taken from the Jilin chondrite (Begemann et al. 1995). The solid curve is the expected trend of the shielding effect on the production rate of 2lNe (Eugster 1988; normalized to 21Ne = 1.8 x 10H cm3STPg and 22Ne/21Ne = 1.06). These measurements show clearly that for 22Ne/21Ne<1.06 the predicted relation is not valid and that 22Ne/21Ne = 1.06 is the lower limit of cosmogenic neon in chondrites. 69 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 4. Cosmogenic 22Ne/21Ne ratios of silicates from chon¬ drites and mesosiderates that have values smaller than 1.06 (adapted from Begemann & Schultz 1988). Data points are con¬ nected when different mineral separates from the same meteor¬ ite have been analyzed. 0.7 0.6 ^ 0.5 < GO 0 4 + cn S 0.3 D> ** 0.2 0.1 0.0 - 1 - 1 - i— : - 0.90 1.00 1.10 1.20 1.30 22Ne/21Ne Figure 5. :2Ne/:iNe ratios of silicate separates from iron meteor¬ ites as function of their Mg/(Mg+Si+Al) ratio. The position of chondrites is shown by the hatched area. For the same chemical composition the silicates of iron meteorites have lower 22Ne/ 21Ne ratios due to an enhanced flux of secondary neutrons. Samples with low Mg concentrations are feldspar-rich with rela¬ tively high sodium contents. Their 22Ne/21Ne is influenced by higher production rates of 22Ne from Na. — r □ \ • Landes o Zagora ■ Linwood □ Four Comers -Bulk. \ \ Chondrites ©- Figure 4 (adopted from Begemann & Schultz 1988) shows 22Ne/21Ne as a function of the Mg(Mg+Si+Al) ra¬ tio. More than 95% of cosmogenic neon is produced in chondrites under normal shielding conditions from these three elements, about 80% alone from Mg. There is a small dependence of the 22Ne/21Ne on the chemical com¬ position as indicated by the lines connecting measure¬ ments of mineral separates of individual meteorites. Re¬ ferring to the same chemistry [= Mg/(Mg+Si+Al)] the lowest 22Ne/21Ne of mesosiderites is about 1.01 and defi¬ nitely lower than the lower limit for chondrites of 1.06. Figure 5 shows a similar diagram for the measure¬ ments of silicate inclusions from iron meteorites. The mineral separates of Landes and Zagora show a similar dependence of 22Ne/21Ne ratios on chemistry as observed for mesosiderites; however, the feldspar samples do not follow this trend because of the ^Ne produced from Na (see above). The 22Ne/21Ne - if taken at the chondrite chemical composition - is 1.03 for Landes and 0.97 for Zagora. Those of Linwood and Four Corners cannot be deduced with certainty from this graph because only one sample of each wras measured. However, assuming simi¬ lar relations between chemistry and 22Ne/21Ne as ob¬ served for Landes and Zagora, the cosmogenic 22Ne/21Ne is even smaller than 0.97. The underlying physical reason for smaller cosmo¬ genic 22Ne/21Ne ratios in iron-rich meteorites compared to chondrites is the higher multiplicity for the production of secondary neutrons from high-Z elements like Fe and Ni as compared to Mg, Si or A1 (Begemann & Schultz 1988). These additional neutrons enhance the production rate of 21Ne via the reaction 24Mg(n,a)2lNe and, thus, lower the cosmogenic 22Ne/21Ne ratios. This effect complicates their use as a shielding indica¬ tor of silicates in iron-rich meteorites because 22Ne/21Ne depends on the following: (1) Chemical composition of the analyzed sample, es¬ pecially the Mg concentration. This effect is seen by the dashed lines in Figure 5. (2) The chemical composition of the bulk meteorite because a high-Z matrix will change the second¬ ary neutron flux. The difference between indi¬ vidual iron meteorites in Figure 5, after taking into account the chemical composition of the min¬ erals, is presumably partly due to their different percentage of silicates, graphite and troilite. (3) Shielding as a function of preatmospheric size and location of the analyzed sample. This complicated relationship makes it difficult to use the cosmogenic 22Ne/2lNe in silicates separated from iron rich meteorites as a shielding correction factor of cosmo¬ genic nuclide production rates. Furthermore, model cal¬ culations or empirical models describing production rates of cosmogenic nuclides in iron meteorites (e.g. Nagai et al. 1993) must consider this matrix effect. It should also be noted that in the metal phase of iron meteorites, cosmogenic 22Ne/21Ne is generally greater than 1.06 (see compilation by Schultz & Kruse 1989). Acknowledgments: Wc dedicate this paper to J de Laeter and we thank him for his continuing interest in our work and for his wise council in many questions concerning atomic weights and isotopic ratios. We thank¬ fully acknowledge the advice of K Rosman in regard to this paper and thank R Clarke for providing meteoritic samples. 70 Journal of the Royal Society of Western Australia, 79(1), March 1996 References Anders E 1962 Meteorite Ages. Reviews of Modern Physics 34:287-325. Bauer C A 1947 Production of helium in meteorites by cosmic radiation. Physical Review 72:354-355. Begemann F & Schultz L 1988 The influence of bulk chemical composition on the production rate of cosmogenic nuclides in meteorites. Lunar and Planetary Science 19:51-52. Begemann F, Weber H W, Vilcsek E & Hintenberger H 1976 Rare gases and 36C1 in stony-iron meteorites: Cosmogenic elemental production rates, exposure ages, diffusion losses and thermal histories. Geochimica et Cosmochimica Acta 40:353-368. Begemann F, Zhaohui Li, Schmitt-Strecker S, Weber H W & Zitu Xu 1985 Noble gases and the history of Jilin meteorite. Earth and Planetary Science Letters 72:247-262. Begemann F, Fan Caiyun, Weber H W & Wang Xianbin 1996 Light noble gases in Jilin: More of the same and something new. Meteoritics and Planetary Science: in press. Crabb J 1983 On the siting of noble gases in silicate inclusions of the El Taco iron Meteorite. Lunar and Planetary Science 14:123-135. Eberhardt P, Eugster O, Geiss J & Marti K 1966 Rare gas mea¬ surements in 30 stone meteorites. Zeitschrift fur Naturforschung 21 a:41 6-426. Eugster O 1988 Cosmic-ray production rates for 3He, 21 Ne, 38 Ar, 83Kr, and ,2hXe in chondrites based on 81Kr-Kr ages. Geochimica et Cosmochimica Acta 52:1649-1662. Gentner W & Zahringer J 1957 Argon und Helium als Kernreaktionsprodukte in Meteoriten. Geochimica et Cosmochimica Acta 11:60-71. Hintenberger H, Schultz L & Weber H 1969 Rare gases in the iron and in the inclusions of the Campo del Cielo meteorite, El Taco. In: Meteorite Research (ed P M Millman). Reidel, Dortrecht, 895-900. Huntley H E 1948 Production of helium by cosmic rays. Nature 161:356. Jentsch O & Schultz L 1987 Trapped noble gases in silicate in¬ clusions of the Landes iron meteorite. Meteoritics 22:418-419. Lai D 1972 Hard rock cosmic archeology- Space Science Re¬ views 14:3-102. Loeken T, Scherer P, Weber H W & Schultz L 1992 Noble gases in eighteen stone meteorites. Chemie der Erde 52:249-259. Masarik J & Reedy R C 1994 Effects of bulk composition on nuclide production processes in meteorites. Geochimica et Cosmochimica Acta 58:5307-5317. Michel R, Dragovitsch P & Filges D 1990 On the dependence of cosmogenic nuclide production rates in meteoroids on bulk chemical composition. Meteoritics 25:386-387. Nagai H, Honda M, Imamura M & Kobayashi K 1993 Cosmo¬ genic l0Be and :,'A1 in metal, carbon, and silicate of meteor¬ ites. Geochimica et Cosmochimica Acta 57:3705-3723. Niemeyer S 1979 I-Xe dating of silicate and troilite from IAB iron meteorites. Geochimica et Cosmochimica Acta 43:843- 860. Paneth F A, Reasbeck P & Mayne K I 1952 Helium 3 content and age of meteorites. Geochimica et Cosmochimica Acta 2:300-303. Reasbeck P & Mayne K I 1955 Cosmic radiation effects in mete¬ orites. Nature 176:733-734. Reedy R C, Arnold J R & Lai D 1983 Cosmic-rav record in solar system matter. Annual Review of Nuclear Particle Science 33:505-557. Schultz L & Kruse H 1989 Helium, neon, and argon in meteor¬ ites - a data compilation. Meteoritics 24:155-172. Schultz L, Weber H W & Begemann F 1991 Noble gases in H- chondrites and potential differences between Antarctic and non-Antarctic meteorites. Geochimica et Cosmochimica Acta 55:59-66. Smith S P & Huneke J C 1975 Cosmogenic neon produced from sodium in meteoritic minerals. Earth and Planetary Science Letters 27:191-199. Vogt S, Herzog G F & Reedy R C 1990 Cosmogenic nuclides in extraterrestrial materials. Reviews of Geophysics 28:253-275. Voshage H & Feldmann H 1978 Investigations on cosmic-ray- produced nuclides in iron meteorites. 1. The measurement and interpretation of rare gas concentrations. Earth and Plan¬ etary Science Letters 39:25-36. 71 Journal of the Royal Society of Western Australia, 79:73-79, 1996 Aspects of low energy nuclear fission J W Boldeman Physics Division, Australian Nuclear Science and Technology Organisation, Private Mailbag 1, Menai NSW 2234 Abstract Since the discovery of nuclear fission in 1939 there have been numerous studies of the fission process. This paper considers several areas of research to illustrate some of the contributions that have been made. The emphasis is on neutron-induced fission at relatively modest excitation energies with some consideration of spontaneous fission. Introduction The fission process is defined as the division of a heavy excited nucleus into two fragments similar in mass. Since the discovery of fission (Hahn & Strassman 1939) there have been tens of thousands of research pa¬ pers devoted to different aspects of the process. The two primary reasons for these extensive studies are obviously the generation of accurate data and understanding of the process for the design and subsequent safe operation of power and research reactors and the contribution that such studies can make to an improved understanding of nuclear structure. Because of the enormous range of top¬ ics and the plethora of details, this paper can address at most only a minute fraction of the current literature. The emphasis here will be on low energy neutron induced fission. Over the years there have been many excellent reviews. An extremely comprehensive book by Vandenbosch & Huizenga (1973) covered the studies to that date. A very recent book on the topic was edited by Wagemans (1991). The basis of selection of material for the various figures has been to provide a simple picture of the various processes. In order to understand what is happening in the fis¬ sion process, it is clear that there are a number of ques¬ tions that must be answered. Some of the more obvious are listed below; - if fission can occur what is it that stops all nuclei from breaking apart spontaneously?; - what determines the fission probability?; - when fission does occur what can be learnt from the angular distribution of the fragments?; - how is the energy shared after fission?; - what is the mass distribution of the fragments?; - how are the neutrons emitted in the fission process and what are their energy distribution?; and - what can be learnt from the yields in the symmetric region of the mass yield curve? © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 The Fission Barrier The starting point in any discussion of the fission bar¬ rier is the liquid drop model introduced before fission was actually discovered but modified almost immedi¬ ately by Bohr & Wheeler (1939). In this model, the nucleus is assumed to be equivalent to a liquid drop in which the short range nuclear forces are idealised by the surface tension of the drop and the Coulomb repulsive forces are included by assuming the drop to be uniformly charged throughout its volume. Fission occurs within this model if sufficient excitation is given to the system to allow the internal vibrations to overcome the attractive surface tension of the drop. Although the liquid drop model provided, in principle, a reasonable approxima¬ tion to the real world, it failed in accurately predicting many of the detailed systemmatics of the fission process. A major advance occurred in the mid 1960's, when the model was improved dramatically by the inclusion of shell effects (Strutinsky 1967; Strutinsky & Pauli 1969). The consequence of these corrections is illustrated by plotting the potential energy hindering fission as a func¬ tion of the symmetric axis deformation (Fig 1); also shown is the smooth potential barrier as originally pre¬ dicted by the liquid drop model. It is seen that the single smooth barrier of the simple liquid drop model becomes double humped for nuclei in the vicinity of the uranium. This extension of the model allows many unexpected fea¬ tures of the fission process to be explained. It also ex¬ plains why the ground state for those nuclei represented by the potential barrier in the figure (nuclei in the vicin¬ ity of the uranium nuclei) is deformed. It is now important to introduce the concept of fission channels. In 1956, A Bohr first suggested that for a fis¬ sioning nucleus with excitation only slightly more than the fission threshold i.e with an energy only slightly ex¬ ceeding the potential energy barrier in Fig 1, the nucleus at a deformation corresponding to the highest point in Fig 2 is cold with respect to internal excitation, all energy is bound up in potential energy of deformation and the only nuclear states at the peak of the fission barrier (the saddle point) via which fission can proceed will be col¬ lective states similar to those of the heavy deformed nu¬ clei near their ground states i.e in the first well of the potential barrier in Fig 1. The transition states will be characterised by the quantum numbers I and K, where I is the total spin of the compound nucleus and K is its 73 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 1. Double humped fission barrier, compared with the liquid drop model (modified from Strutinksky (1967). Figure 2. Comparison of transition states at the fission barrier with the spectrum of low lying states for deformed nuclei (modified from Griffin 1965). 74 Journal of the Royal Society of Western Australia, 79(1), March 1996 projection on the symmetric axis i.c. along the axis of deformation specified by the deformation parameter in Fig 1. Extensive research, particularly in the early days of fission studies, identified a spectrum of different quan¬ tum states at the saddle point which were necessary to explain many of the details of the fission process. Figure 2, modified from Griffin (1965) illustrates the wealth of data that has been obtained principally from studies of fission fragment angular distributions. Although this in¬ formation was derived originally for a simple single humped barrier, to a first approximation Fig 2 represents the spectrum of double humped transition states. As indicated previously, the angular distribution of the fission fragments can be used to derive information regarding the spectrum of states at the saddle point of the fissioning nucleus (Fig 2). In an even-even nucleus, the spectrum of transition states at the saddle point de¬ formation is expected to be quite similar to that found at low excitation in its permanently deformed equilibrium configuration. The ground state has total angular mo¬ mentum and parity Itt = 0+ and has projection of angular momentum on the nuclear symmetry axis K = 0. The excited states correspond to the simple vibrational modes of a liquid drop. The lowest mass-asymmetric mode has K = 0, but negative parity, whereas the axial-symmetric gamma vibration of lowest energy has Ktt = 2+. In addi¬ tion, a bending mode with Ktt = 1- is expected to occur at moderate excitation at the saddle point, although it has not been identified at equilibrium deformation. Since the fissioning nucleus is non-spherical, rotational bands of increasing I are built on each of these vibrational states. Simple superposition of these fundamental modes leads to further fission channels with increasingly greater exci¬ tation until, at about 1.5 MeV, sufficient energy is avail¬ able to break nucleon pairs and thereafter single particle excitations combined with collective vibrations lead to a rapid increase in the complexity of the transition state spectrum. Fission Fragment Angular Distributions A basic assumption of the Bohr model is that the quantum number K remains a constant of the motion between saddle point and scission. Thus the measured angular anisotropy W(0°)/W(90°) can be used to provide information on the properties of the transition states at the saddle point. The most extensively studied even- even fission system studied is neutron fission of U235. Figure 3, modified from the recent study by Straede et al. (1987) shows the anisotropy i.e. W(0°)/W(90°) as a func¬ tion of neutron energy. The data is readily explained in terms of the model discussed above. For neutron fission of an even target nucleus leading to an odd fissioning system, the lowest lying fission channels are expected to be essentially single particle states which should be identifiable with the appropriate Nilsson orbits. The excess angular momentum appears as a rotation about an axis perpendicular to the symmetric axis. Thus with each intrinsic state there is associated a rotational band with energy given by E(Ih)=Ed£! [l(I+l)-2KJ+5^a(-l)^(I+1/2)] (eq. 1) where j is the moment of inertia associated with the band and a is the decoupling constant for the K= Vi bands. The structure in the anisotropy becomes much larger. As an example, measurements of the anisotropy for neutron fis¬ sion of 230Th near the large resonance in the cross section at 715 keV are shown in Fig 4. Neutron Energy (MeV) Figure 3. The 235U(n,f) fission fragment anisotropy (modified from Straede et al 1987). 75 Journal of the Royal Society of Western Australia, 79(1), March 1996 ' ■+— - 1 - 1 - 1 - t- 680 700 720 740 760 Neutron Energy (keV) Figure 4. Fission fragment anisotropy1'1 for neutron fission of ™Th (adapted from Boldeman & Walsh, 1986). Sub-barrier Fission Cross-sections For nuclei represented by the fission barrier in Fig 1, fission can only take place if sufficient energy is given to the system to allow it to proceed either over the barrier or to tunnel through. For sub-barrier fission, the cross section is given by an expression which contains reaction information multiplied by the penetrability through the barrier. The penetrability through the simple liquid drop fission barrier is given by the expression P = [1+ exp(2Tr(Ei + EIKt! - E)/fuu)] 1 (eq. 2) where Ei is the height of the barrier, EIK" is the energy of level (IKtt), E is the excitation energy and hu) is the curvature of the fission barrier. With a double humped barrier the expression becomes a little more complex TfK = + (1- rK - tK) PA(PA + PB) (eq. 3) where tK and rK are the transmission and reflection coefficients which relate to the amplitudes of the fission wave functions inside and outside the nuclear potential and PA and PB are the penetrabilities for the two barriers defined as before. Some unexpected features of the fission cross-section follow from the double humped barrier shape. The po¬ tential well between the two humps of the double humped barrier will clearly contain states similar in char¬ acter to the states in the first well. However the level density at a specific excitation will be different because of the different heights above the ground state. This produces some interesting structure in the sub-barrier cross-section. Figure 5, adapted from Migneto & Theobald (1968), illustrates the influence of the coupling of states in the first well to states in the second well. The gross structure in the sub-barrier neutron fission cross section of 240Pu reflects the spacing of levels in the second well of the fission barrier while the fine structure is re¬ lated to the level density in the first well. Although many nuclei have barrier shapes characterised by the doubled humped shape of Fig 1, it was predicted by Moller & Nix (1974) that, if asymmetric Figure 5. The neutron fission cross section of 240Pu between 500 eV and 3000 eV (adapted from Migneco & Theobald 1968). 76 Journal of the Royal Society of Western Australia, 79(1), March 1996 NEUTRON ENERGY (keV) Figure 6. Calculated neutron fission cross section compared with the experimental data of Blons et al. (1978). Also shown are the fission cross section data from the angular distribution mea¬ surements (Boldeman & Walsh 1986). deformations are taken into account for nuclei in the vi¬ cinity of thorium, then the second barrier should itself split into two separate peaks making the barrier triple humped in character. The consequences of this charac¬ teristic are shown in the sub-barrier neutron fission cross section of 230Th as measured by Blons et al. (1978). The broad sub-barrier resonance in the cross-section at 715 keV is interpreted as fission through a pure vibrational state in the third well of a triple humped fission barrier. Some controversy regarding the detailed interpretation of the data exists. One interpretation (Boldeman & Walsh 1986) involving a simultaneous fit to the cross sec¬ tion in Fig 6 and the anisotropy data in Fig 4 is presented below. Fission Fragment Yields and Associated Neutron Emission One of the most extensively studied aspects of the fis¬ sion process has been the systematics of the mass divi¬ sion. In this context, it is important to distinguish be¬ tween the fission fragments and fission products and ul¬ timately accumulated product yields. At the instant of scission, the two fission fragments are highly deformed with a high level of internal excitation, although a large proportion of the energy emitted in fission is contained in the kinetic energy of the fission fragments resulting from Coulomb repulsion. Since the fragments are neu¬ tron rich with respect to the stability line, this deforma¬ tion energy is emitted primarily by neutron evaporation and by prompt gamma emission. The term fission prod¬ ucts is used to described the resulting nuclei. Such nu¬ clei are still radioactive and undergo further radioactive decay leading to what are called cumulative yields. Fig¬ ure 7 illustrates the well known asymmetric division that is typical for low excitation neutron fission of nuclei near uranium (modified from Wahl 1965). The dominating in¬ fluence in the determination of the shape of the curve is the existence of the magic number N=50 at mass 80 on the lower end of the distribution and the doubly magic nuclei Z=50, N=82 in the vicinity of mass 128. It is also known that the symmetric region between the two peaks of the mass yield curve starts to fill as the excitation in¬ creases. This can be explained quite simply as the reduc¬ tion of shell effects as the excitation energy increases. Neutron emission from the individual fission fragments is a function of the mass of the fragment (Fig 8; from Nifenecker et al 1973). The resulting "saw tooth" shape of the neutron emission is also consistent with the expla¬ nation presented above. The minimum in the saw tooth curve occurs in the vicinity of the closed shell nuclei. These nuclei are stiff with respect to deformation and therefore less deformation occurs in these fragments. On the contrary, near the peak of the neutron yield curve, the fragments are soft with respect to deformation and a larger proportion of the total fission energy is involved in the deformation of these fragments. Fission Product Yields in the Symmetric Region Studies of fission product yields in the symmetric re¬ gion for thermal neutron fission of the uranium nuclei are important for at least three reasons. Although the yields are low, in a power reactor significant masses of such nuclei are produced and the management of the reactors requires this knowledge. A second reason fol¬ lows from suggestions of fine structure in the fission fragment /(product) yields. The symmetric region pro¬ vides an opportunity to carry out such studies since there are several elements with large numbers of isotopes (e.g. Sn) which therefore simplifies the chemistry. In addi- 77 Journal of the Royal Society of Western Australia, 79(1), March 1996 B Fragment Mass (amu) Figure 8. Average neutron emission per fragment as a function of fragment for (a) thermal neutron fission of 23?U and (b) spontaneous fission of 252Cf (from Nifenecker et al, 1974). 78 Journal of the Royal Society of Western Australia, 79(1), March 1996 117 118 119 120 122 124 126 Mass Number Figure 9. Relative fission yields for isotopes of tin for neutron fission of 235U (thermal and epithermal) normalised to ,26Sn (from Rosman et al. 1986). tion, there are extensive studies of earth science where spontaneous fission of the uranium nuclei produce el¬ emental contaminations that need correction. A series of studies by Rosman et al (1983) has produced consider¬ able symmetric mass yield data. Figure 9 shows their data for the Sn isotopes in the symmetric mass yield re¬ gion; an absence of fine structure is apparent. Recent Developments of the Theory In very recent times, there have been some important developments of the theory of fission particularly at low excitation. These developments have been summarised by Brosa et al (1990). Effectively, they involve a more complete consideration of the later stages of the process effectively from the saddle point to scission. These new discoveries show that for low excitation fission there are several exit channels on paths to scission. Furthermore, when the nucleus is close to scission there is some ran¬ domness in where the neck connecting the elongated nucleus may actually rupture. These considerations are beyond the context of this paper and the reader is re¬ ferred to Brosa et al (1990) and other papers mentioned in that text. References Blons J, Mazur C, Paya D, Ribrag M & Weigmann H 1978 Rota¬ tional bands in asymmetrically deformed 23lTh. Physics Re¬ view Letters 41:1282-1285. Bohr A 1956 On the theory of nuclear fission. 1st International Conference on Peaceful Uses of Atomic Energy, New York, Vol 2, 151-154. Bohr N & Wheeler 1939 The mechanism of nuclear fission. Physics Review 56:426-450. Boldeman J W & Walsh R L 1986 Neutron fission of 230Th revisted. Radiation Effects 92/1-4:317-322. Brosa U, Grossmann S & Muller A 1990 Nuclear scission. Phys¬ ics Reports (A Review Section of Physics Letters) 197:167- 262. Griffin J J 1965 Transition states at the fission barrier. Physics and Chemistry of Fission 1:23-38. Hahn O & Strassmann F 1939 Uber den nachweis und das verhalten der bei der bestrahlung des urans mittels neutronen entstehenden erdalkalimetalle. Naturwissenschaften 27:11. Migneco E & Theobald J P 1968 Resonance grouping structure in neutron induced subthreshold fission of 241lPu. Proceedings of the International National Bureau of Standards, Specialist Publication 299, Vol 1:527-540. Moller P & Nix J R 1974 Calculation of fission barriers. Third IAEA Symposium on the Physics and Chemistry of Fission, Rochester, Vol 1:103-143. Nifenecker H, Signarbieux C, Babinet R & Poitou J 1974 Neu¬ tron and gamma emission in fission. Proc. Third IAEA Sym¬ posium on Physics and Chemistry in Fission, Rochester, Vol 2:117-178. Rosman K J R, De Laeter JR, Boldeman J W & Thode HG 1983 Cumulative yields of stable and long lived isotopes of tin in neutron-induced fission. Canadian Journal of Physics 61:1490-1497. Straede Ch, Budtz-Jorgensen C & Knitter H H 1987 The 235tJ(n,t) fragment mass-, kinetic energy- and angular distributions for incident neutron energies between thermal and 6 MeV. Nuclear Physics A462:85-103. Strutinsky V M 1967 Shell effects in nuclear masses and defor¬ mation energies. Journal of Nuclear Physics A95:420-442. Strutinsky V M & Pauli H C 1969 Shell-structure effects in the fissioning nucleus. Second International Symposium on the Physics and Chemistry of Fission, Vienna, Vol 1, 155-181. Vandenbosch R & Huizenga J 1973 Nuclear Fission. Academic Press, New York. Wagemans C 1991 The Nuclear Fission Process. CRC Press, Boca Raton. Wahl A C 1965 Mass and charge distribution in low-energy fission. Physics and Chemistry of Fission 1:317-331. 79 Journal of the Royal Society of Western Australia, 79:81-83, 1996 Isotopic studies of the Oklo fossil reactors and the feasibility of geological nuclear waste disposal R D Loss Department of Applied Physics, Curtin University of Technology, Perth, WA 6102 Abstract In 1972, six 2000 million year old natural fossil nuclear fission reactors were discovered at a uranium orebody at Oklo in Gabon, South West Africa. Over a period of about one million years the reactors produced some hundreds of tonnes of fission products with characteristic but unnatu¬ ral isotopic compositions. These isotopic compositions have enabled the nuclear characteristics of the reactors and the extent to which they have retained their fission products to be determined. Since their discovery another dozen fossil fission reactors have been found in Gabon. This paper summarizes the results obtained from both the earlier and recent discoveries and their implications for the storage of anthropogenically generated fission waste in geological environments. Introduction: The existence of natural fossil chain fission reactors is perhaps one of the most unusual discoveries in the com¬ paratively short history of nuclear science. When Fermi and colleagues initiated the first artificial self sustaining chain fission reactions they did not realize that nature had assembled and operated a similar type of reactor 2000 million years before. Fourteen years after Fermi's achievements, Kuroda (1956) published the details of the conditions under which such reactors could exist. How¬ ever, despite an extensive survey of the 23SU/23SU ratio during the 1960s, no variation greater than analytical un¬ certainty (±<0.1% ) was found. In 1972, routine measure¬ ments of 235U/238U in a French uranium processing plant revealed a depletion in this ratio of 0.43%. At first it was thought that this was caused by "spent fuel" from a con¬ ventional fission reactor but it was eventually traced io uranium from a sedimentary type uranium deposit at Oklo in Gabon, South West Africa. A survey of the exist¬ ing Oklo mine site revealed six metre long and tens of centimetres thick lenses of highly enriched U which con¬ tained both depleted 235U/23SU ratios and substantial quantities of nuclear fission products. The circumstance giving rise to these fossil nuclear reactors has subse¬ quently become known as the "Oklo Phenomenon". The potential significance of the Oklo fossil reactors as natural analogs for geological repositories of radioactive waste was quickly recognized by the scientific commu¬ nity. Although the possibility of geological storage had been considered prior to the discovery of the Oklo reac¬ tors, the methods of assessing waste retention over the time required for it to decay to "safe" levels were ex¬ tremely limited. The long term behaviour of such waste had to be extrapolated from extremely short term experi¬ ments under conditions which poorly approximated those likely to be found in real repositories. Direct ex¬ perimentation of sites through trial deposition of hun- © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 dred of tonnes of fission waste was also potentially envi¬ ronmentally irresponsible. Besides, the waste problem needs to be tackled now and cannot wait for the time required to assess the potential of geological storage. The role of isotopic studies: Isotopic studies have proved to be the master key for unlocking the many secrets of the Oklo phenomenon. Although non-isotopic studies of the Oklo reactors may reveal unusual amounts of some elements it is only by measuring their isotopic composition that the existence of nuclear fission can be confirmed. Isotopic studies by mass spectrometry are especially powerful because they allow the fission products to be distinguished from the natural elements and their amounts measured in any rock sample. The wealth of information available from isotope stud¬ ies can be illustrated using tellurium as an example. Fig¬ ure 1 shows the relative isotopic compositions of five of the major isotopes of tellurium as measured in a range of samples. For comparison all isotope abundances are ref¬ erenced to the 130Te abundance taken as unity. The solar system values shown are those present in all terrestrial sources, meteorites and the moon (Smith et al., 1978). Samples ORZ9-028 and ORZ9-006 represent Te from samples from inside reactor zone 9 while OHR9-002 is from a host rock sample taken 1.4 m away from this reactor zone. The isotopic abundances of 12HTe and 130Te in ORZ-028 are essentially the same as that produced by the fission of 235U. Since no ,24Te is produced by fission, the absence of this isotope indicates that no terrestrial Te is present in the reactors and that all of the Te present in the reactors is fissiogenic.The 128Te abundance in OHR9-002 indicates that this sample contains a mixture of both fissiogenic and solar system Te The other isotope of significance is 126Te which in excess in samples ORZ9-006 and in par¬ ticular OHR9-002. This effect arises from the parent iso¬ tope, 126Sn which has a half life 105 years indicating that radioactive tin is mobilized inside and in the immediate 81 Journal of the Royal Society of Western Australia, 79(1), March 1996 0) o c cc "O c 3 n co o > re Q) CC 1 .2 0.8 0.6 0.4 0.2 CRZ9-02 8 E3 CRZ9-006 CHR9-002 dSola System tyf~ ■*— f- 124 125 1 26 127 28 129 30 Mass Number Figure 1. The isotopic composition of tellurium in a range of Oklo and other samples (data from Curtis et al. 1989 and Loss et al. 1989) vicinity of the reactors during the reactions and within one million years (10 half lives) of the last reaction. The abundance of fission product Te present in OHR9-002 was also measured to be 3 orders of magnitude lower than that in the reactor zone while at a distance of six metres from the reactor zone the abundance of fission product Te is 1 part in 10s of that inside the reactors. Besides confirming the occurrence of the fission pro¬ cess at Oklo, the isotopic compositions of the fission products enabled the nuclear characteristics of the reac¬ tors themselves to be determined (Naudet 1975 and Ruffenach et al., 1980). The parameters of interest include the integrated neutron flux (fluence), the degree of ther- malization of the neutron spectrum, and the fractional fission contribution from 238U and 239Pu. From these pa¬ rameters, other physical characteristics such as the power output and temperature of the reactors and surrounding host rocks can be determined. These parameters, together with the age of the reactors, the degree of U enrichment and other geochemical parameters have been used to cal¬ culate the total quantity of fission products expected to be produced by the reactors. When this is compared to what is present today a retention factor for the fission products can be determined. For example, reactor zone 9 has retained more than 80% of its fissiogenic Te - a re¬ markable achievement for a 2000 million year old ura¬ nium oxide mineral assemblage located inside a highly porous sandstone matrix. Isotopic studies have shown that many fission pro¬ duced elements have been retained inside the reactor zones (Curtis et al., 1989). The rocks surrounding the re¬ actors have also retained other fission products within a few metres of the reactor zones (Loss et al., 1989). These results are in excellent agreement with those predicted by thermodynamics and geochemistry (Brookins, 1975). Flowever, there are still many aspects of the reactors that remain to be studied. Recent discoveries Since 1972, another dozen or so fossil fission reactors have been discovered at Oklo, Oklobondo and at Bangombe, some 30 km further south of Oklo. These re¬ actors are significant discoveries in their own right since they include a wider range of geological conditions than those represented by the original reactors. One of these reactors is located within metres of a 1000 million year old doleritic dike which cut through the centre of the Oklo uranium ore deposit. This has provided the oppor¬ tunity to study the affect of dike emplacement on the retention of fission products within a reactor. Although only a limited number of studies of these newer reactors have been undertaken the retention of fission products appears to be similar to those for the original reactors. Amongst the most significant new results has been the detection of previously unstudied but very important long lived radiogenic and highly mobile nuclei, 137Cs (de¬ caying to l37Ba) and wSr (decaying to wZr) within metres of reactor zones. These elements were previously thought to have been completely lost from the vicinity of the reac¬ tors (Hidaka et al., 1993 and 1994). Another unusual dis¬ covery has been the formation of unusual refractory and highly insoluble alloys which have trapped almost pure fissiogenic elements within the original uraninite ore grains (Gauthier-Lafaye et al 1996). Studies of this type of alloy may prove useful in developing a suitable stor¬ age matrix for some types of spent fuel. Perhaps the most unusual of the new reactors are those at Bangombe which are much closer to the surface and considerably more weathered and oxidized than those at either Oklo or Oklobondo. Studies of this type of reactor will be particularly useful since future under¬ ground waste sites may eventually be exposed to surface conditions and more aggressive geochemical environ¬ ments. Isotopic studies of minerals within and surround¬ ing this reactor zone has identified clays and phosphates 82 Journal of the Royal Society of Western Australia, 79(1), March 1996 as important in retaining rare earth element fission prod¬ ucts. A case in point is the detection plutonium retention identified by a higher than solar system 235U/238U ratio (Bros et al. 1993). This has occurred in what has essen¬ tially been a high rainfall equatorial environment for per¬ haps millions of years. Conclusions The most important lesson that Oklo teaches us is that nature is able to isolate fissiogenic waste in geological environments for extremely long periods of time even though the reactors were contained in a highly fractured, porous and physically "leaky" matrix. This geophysical environment is one that would hardly be considered as ideal for a nuclear waste repository. The reason for the retention of fissiogenic materials under these conditions was largely due to the local geochemical conditions. The oxidizing potential and acidity of the ground waters together with the presence of scavenging clays and other minerals played a critical role in this process. However it can be concluded that the concept of multiple barrier geological containment including iron oxides and clays which are able to buffer possible changes in ground wa¬ ter conditions indicates that the effective containment of man-made nuclear waste is feasible. References Brookins D G 1977 Applications of the Eh-pH diagrams to problems of retention and/or migration of fissiogenic elements at Oklo. International Symposium of the Oklo Phenomenon, Libreville, Gabon. IAEA TC-1 19/33:243-265. Bros R, Turpin L, Gauthier-Lafaye F, Holliger P & Stille P 1993 Occurrence of naturally enriched 235U: Implications for pluto¬ nium behaviour in natural environments. Geochemie et Cosmochimie Acta 57:1351-1356. Curtis D B, Benjamin T M, Gancarz A L, Loss R D, Rosman K J R, de Laeter J R, Delmore J E, & Maeck W J 1989. Fission product retention in the Oklo natural reactors. Applied Geo¬ chemistry 4:49-62. Gauthier-Lafaye F, Holliger P & Blanc P-L 1996 Natural fission reactors in the Franceville Basin, Gabon: A review of the conditions and results of a “critical event" in a geological system. Geochim etCosmochim Acta: in press. Hidaka H, Holliger P & Masuda A 1993 Evidence of fissiogenic Cs estimated from isotopic deviations in a Oklo natural Re¬ actor zone. Earth and Planetary Science Letters, 114:391-396. Hidaka H, Sugiyama K, Ebihara M. & Hollinger P. 1994 Isoto¬ pic evidence for the retention of 90Sr inferred from the ex¬ cess of 90Zr in the Oklo natural fission reactors. Earth and Planetary Science Letters, 122:173 -182. Kuroda P 1956 On the nuclear Stability of Uranium Minerals. Journal Chemical Physics 25:781 -782. Loss R D , Rosman K J R, de Laeter J R, Curtis D B, Benjamin T M, Gancarz A L, Maeck W J & Delmore J E 1989 Fission product retentivity in peripheral rocks at the Oklo natural fission reactors, Gabon. Chemical Geology 76:71-84. Naudet R. 1975 Mechanismes de regulation des reactions nucleaires. In: The Oklo Phenomenon. Symposium Proceed¬ ings, Libreville, IAEA, Vienna 589-600. Ruffenach J C, Hagemann R & Roth E 1980 Isotopic Abundances measurements - a key to understanding the Oklo Phenom¬ enon. Zeitschrift fuer Naturforsch ung 35A:171-179. Smith C L Rosman K J R & de Laeter J R 1978 The isotopic composition of tellurium. International Journal of Mass Spectromatry and Ion Physics 28:7-17. 83 Journal of the Royal Society of Western Australia, 79:85-90, 1996 Isotopic signatures: An important tool in today’s world D J Rokop, D W Efurd, T M Benjamin, J H Cappis, J W Chamberlin, H Poths, & F R Roensch Chemical Science and Technology Division, Los Alamos National Laboratory, MS J-514, Los Alamos, NM 87545 USA Abstract High-sensitivity, high-accuracy actinide measurement techniques developed to support weap¬ ons diagnostic capabilities at the Los Alamos National Laboratory are now being used for environ¬ mental monitoring. The measurement techniques used are thermal ionization mass spectrometry, alpha spectrometry, and high resolution gamma spectrometry. These techniques are used to ad¬ dress a wide variety of actinide inventory issues; environmental surveillance, site characteriza¬ tions, food chain member determination, sedimentary records of activities, and treaty compliance concerns. As little as 10 femtograms of plutonium can be detected in samples, and isotopic signatures determined for samples containing sub-100 femtogram amounts. Uranium, present in all environmental samples, can generally yield isotopic signatures of anthropogenic origin when present at the 40 picogram g'1 level. Solid samples (soils, sediments, fauna, and tissue) can range from a few particles to several kilograms in size. Water samples can range from a few mL to as much as 200 L. Introduction The moratorium on weapons testing has made it nec¬ essary to find new applications for the radiochemical di¬ agnostic capabilities of the Chemical Science and Tech¬ nology Division at Los Alamos. Many were explored, but the three which seemed to hold the highest potential for doing both good science and having sufficient funds to support them were environmental monitoring of ac¬ tinide elements in support of clean-up efforts, providing tracers for industrial applications, and supporting the International Atomic Energy Agency (IAEA) in its efforts to monitor the Non-Proliferation Treaty (NPT). We are actively involved in the first and third areas and aggres¬ sively pursuing the second. Environmental actinide monitoring has historically been done with inexpensive survey techniques, princi¬ pally alpha spectrometry (AS). The technique, which is fast and relatively simple, suffers one major drawback; it has inadequate resolution to separate anthropogenic source terms from natural background or fall-out. AS also cannot attribute single versus multiple source terms. Mass spectrometric isotopic dilution (MS1D), a tool long used by isotope geologists and the nuclear industry, had not until recently been applied in a broad way to envi¬ ronmental problems. Our first effort using this method was an actinide inventory of the Pajarito Plateau, the area surrounding the Los Alamos National Laboratory (LANL). The method quickly illuminated and expanded the scope of actinides in the environment. Actinide con¬ centrations, previously found to be below health limits and therefore not requiring remedial action or attribu¬ tion, can now be detected with 100 to 1000 times more © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 sensitivity than historic techniques and can frequently be attributed to a specific source. While this capability is a powerful new tool for environmental researchers, help¬ ing determine transport phenomena and remediation ef¬ fectiveness, it has also created a communication problem of enormous magnitude. How do you explain to con¬ cerned citizens, with little scientific background, the tre¬ mendous sensitivity of the method, while assuring them of the corresponding negligible health impact? The nature of any material processing, including nuclear, causes minute amounts of various constituents to be lost during handling. For example, certain fission products, such as Ru, I, and Tc, have very volatile chemi¬ cal forms which are difficult to contain. Their detection can provide significant information about the processing nature. The large volume of the material potentially handled by proliferants makes actinides and the associ¬ ated fission products easy to detect. For this reason, the IAEA has decided to adopt environmental monitoring techniques to assist inspectors in assessing member coun¬ tries compliance with the NPT and other applicable trea¬ ties. A former LANL staff member is in Vienna to help the IAEA build a clean laboratory suitable for environ¬ mental measurements. An ongoing program is training chemists and mass spectroscopists from the IAEA in low level environmental measurement techniques. LANL also participated, along with other laboratories, in field trial experiments which were used to demonstrate the merits of the MSID method. LANL is expecting to par¬ ticipate, along with other network laboratories, in the forthcoming environmental monitoring program. This paper will describe the physical requirements necessary to apply the MSID actinide method at envi¬ ronmental levels, the actinide measurement sensitivities of MSID, and specific applications in various matrices. 85 Journal of the Royal Society of Western Australia, 79(1), March 1996 Clean chemistry and mass spectrometry A 1400 m2 clean chemistry and mass spectrometry fa¬ cility was built at Los Alamos to accommodate the pro¬ cessing of samples for weapons diagnostics. This facility permits very small samples of the actinides to be handled such that the integrity of the field sample can be pro¬ tected from contamination to very low levels. Plutonium process blanks are <3 10 IS grams and uranium blanks <1 10"9 grams. This facility has the capacity to handle >2000 actinide samples annually at this low blank level. Lami¬ nar flow 'clean" air passed through hepa-filters directly over work surfaces and rigid protocols for sample con¬ tent and handling help provide this integrity. A com¬ plete description of these techniques can be found in sev¬ eral publications (Efurd et al 1992,1993; Rokop et. al. 1982). Because of it's high crustal abundance, -3 10^ g g1, natural U in the environment dilutes anthropogenic source terms and makes U a high accuracy measurement. An anthropogenic source term is detected by slight changes to the 235/238 ratio (natural ratio is 0.007254) and the presence of 236U. Plutonium, on the other hand, occurs at extremely low concentrations and requires a high-sensitivity measurement. Almost all surface mate¬ rial now contains fallout plutonium in the 10-50 femtogram per -20 g soil sample range. The 240/239 ratio for fallout is subject to slight changes from local signature overlay. The area near Denver, Colorado, with Rocky Flats providing the local overlay, has a 240/239 fallout ratio of 0.169 ± 0.005 (Efurd et al 1994) while sediment samples collected from the Irish Sea have a 240/239 ratio of 0.24 ± 0.01 (unpublished observations; Efurd et al 1995). The integrated average world-wide ratio is 0T8 (Krey, et al. 1976). Given "clean" chemistry and mass spectrometric ca¬ pability, there are two different philosophies for making measurements if the goal is detecting undeclared or illicit activity. These are the "bulk" and "particle" methods. The "bulk" method involves total dissolution of the ma¬ trix or leaching the desired material from the matrix fol¬ lowed by chemical purification and TIMS. The "particle" method involves collecting particles on swipes or filters, separating the particles and depositing them on a slide with collodian, irradiation in contact with Lexan (for re¬ cording the fission tracks created by actinide particles) reregistering the slide and lexan, cutting out the desired particle, and, finally, TIMS. While the "particle" method generates less ambiguous results and has little or no background from natural or fallout sources, it is slow, much more expensive (5x) and highly labor intensive. It also suffers from difficulty at trying to sort out the par¬ ticular or desired indicator of process. IAEA field trials also demonstrated a lower percentage of "hits" (positive response of anthropoyenic origin) than "bulk" (inte¬ grated, analysis of entire sample). The best situation would utilize the "bulk" method and if a "hit" was re¬ corded and less ambiguity was desired, this would be followed up with the "particle" method. As more data is gathered from environmental samples, the question of how to interpret the results is becoming very important. LANL uses algorithms devel¬ oped for weapons diagnostics to separate multiple actin¬ ide sources. We plan to work with the IAEA to refine these models and make them more useful for their spe¬ cific purposes. Table 1 lists a selection of some of the different matri¬ ces analyzed by the Chemical Science and Technology Division (Los Alamos National Laboratory) for determin¬ ing environmental actinides. While this list is incom¬ plete, it serves to indicate the wide application of "bulk" analysis. Particularly effective are filter papers of swipes taken from non-technical areas of nuclear facilities; these give almost complete indication of the process nature. Table 1 Examples of sample materials for which environmental actinide determinations have been made. NTS Debris Soils Sediments Sea Water Fresh Water Filter Papers Vegetation pine needles grass apples feces (sheep) Soft Tissues whale liver whale blubber star fish salmon chubs trout Bones Table 2 shows the measurement sensitivities of the MSID "bulk" analysis method. It can be seen that ap¬ proximately 1000 times more sample is needed for a sig¬ nature than is required for detection. A caveat here is that recent chemical purification improvements would indicate a lowering of atoms needed for signature by ’ lOx. A general rule is that the more complicated the matrix is, the more difficult it is to purify the actinide, and, the less sensitive the measurement. Table 2 Measurement sensitivities for actinides in various matrices and typical sample minations. sizes required for minimum fallout level deter- Elemental Isotopic Concentrations Fingerprints U 5 x 106 atoms 5 x 109 atoms Pu 5 x 105 atoms 5 x 108 atoms Am 2 x 105 atoms 2 x 108 atoms Np 2 x 105 atoms 2 x 108 atoms Separation Chemistry Sample Matrix Sample Size % Recovery Water 2000 ml >90 Soil 100 g >90 Biotissue 100 g >90 86 Journal of the Royal Society of Western Australia, 79(1), March 1996 Environmental surveillance, the Pajarito Plateau In support of the Environmental Surveillance Group at Los Alamos National Laboratory, an actinide inven¬ tory of ground waters and sediments from the surround¬ ing Pajarito Plateau was performed. Uranium and plu¬ tonium were measured in each sample. Although an¬ thropogenic sources were found for both U and Pu in some samples, no correlation was found between the two. One interesting fact which emerged from the study was that no modern weapons grade plutonium was found; only pre-1960 material was found. All known sources of plutonium on the plateau were considered; pre-1960 weapons grade (240/239 ^ 0.01), modern weap¬ ons grade (240/239 = 0.060 ± 0.005), and fallout (240/239 = 0.18 ± 0.02). Only a few samples had modern 240/239 ratios, and they were shown by mixing models to be a mixture of pre-1960s and fallout. This strongly indicates that modern weapons grade material is not present in the samples submitted for analysis. Concentration Concentration Concentration Measurement Isotope Atom Percent atoms/L ng/L pCi/L Error(2 °) U 234 0.010% 1.69E+ 11 6.58E-02 4.10E-01 8.72E-03 U 235 0.675% 1.12E+ 13 4.36E+ 00 9.42E-03 2.18 E-0 5 U 236 0.000% 4.31E+ 09 1.69E-03 1.09E-04 3. 36 E-0 5 U 238 99.315% 1.64E+ 15 6.50E+ 02 2.18 E-0 1 5.39E-04 Total pCi/L sample: 6. 38 E-0 1 Error(± ) in pCi/L: 8.74E-03 Total ng/L sample: 6.54E+ 02 Total atoms/L sample: 1.65E+ 15 U 238 U 236 U 235 U 234 0 00% 10.00% 20.00% 30 00% 40 00% 50 00% 60 00% 70.00% 80.00% 90.00% 100.00% Atom Percent (%) Expanded View for U 234, U 235, U 236 U 236 U 235 U 234 o. Atom Percent (%) ■ Sample UNatural Uranium ** The U 236 concentration indicates a neutron radiation history *’*' Comparing the U 235 to natural Uranium indicates that this sample contains a depleted component * * The U 234 in this sample is in radioactive dis-equilibrium Figure 1. Example of a report for uranium analysis (of groundwater); uranium isotopic and activity levels and bar graph representation of field sample in comparison with natural uranium. 87 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 1 shows a uranium report from this study. The isotopic composition is a mixture of natural uranium, found in all samples as explained above, and depleted uranium. This sample is from groundwater, which is a good medium in which to find an anthropogenic ura¬ nium source term as soils and sediments can dilute out unnatural sources very quickly because of their high natural uranium content. A bar graph representation was chosen to compare natural uranium with the isoto¬ pic composition found in the sample, to insure that full utilization of the information by a customer with little familiarity with isotopic signatures was possible. Com¬ munications with the customer base has become a chal¬ lenge in itself. Figure 2 shows a plutonium report from this study. The bar graph shows the three known source terms for the Pajarito Plateau, old pre-1960's Pu, modern Pu, and fallout (see above for isotopic details). This sample is from ^40 grams of sediment obtained from a canyon bottom. The Pu isotopic signature is clearly dominated by the old weapons grade material. This sample has a relatively high concentration of Pu but still shows the effect of mixing with fallout material. A very accurate history of Los Alamos, at least in its earlier days, could be written from information obtained by environmental monitoring efforts. These efforts also demonstrate the ability to trace a unknown contaminant to its source term. Isotope Atom Percent Concentration atoms/g Concentration ng/g Concentration pCi/g Measurerment Error(2 a) Pu 239 Pu 240 Pu 241 98.724% 1.273% 0.003% 2.56E+10 3.31E+08 8.82E+05 1.02E-02 1.32E-04 3.53E-07 6.32E-01 2.99E-02 3.65E-02 6.95E-04 3.64E-04 1.65E-02 Total pCi/g sample: Error(±) in pCi/g: Total ng/g sample: Total atoms/g sample: 6.98E-01 1.65E-02 1.03E-02 2.60E+10 Pu 240/Pu 239 Ratio Comparison Chart * Note: Shaded Region Indicates Where Sample Range Must Fall for that Category Figure 2. Example of a report for plutonium analysis (of sediment); plutonium isotopic and activity levels and 240/239 signature comparison to known source terms on the Pajarito Plateau. 88 Journal of the Royal Society of Western Australia, 79(1), March 1996 IAEA 93+2 Field trial experiments The IAEA 93+2 program was started to survey avail¬ able measurement techniques which could determine compliance of signees to the NPT. The 93+2 program seeks support in; a) reliable analytical techniques for the determination of undeclared activities; b) training of staff in sampling techniques; c) training of staff in selected analytical techniques; d) search for and obtain access to environmental data bases; and e) to find network laboratories from member states to help relieve the heavy analytical load of the Safe¬ guards Analytical Laboratory (SAL). The selected techniques should be easily implemented into the IAEA structure and partially supported by mem¬ ber states. The first step in the process is the construc¬ tion of a "clean" laboratory for the chemical preparation and measurement of environmental samples. The latter task was accomplished with the support of US funds. The IAEA, with the support the NN-44, the US De¬ partment of Energy Office for International Safeguards, conducted a series of field trial experiments at twelve facilities around the world. The experiments were de¬ signed to demonstrate whether environmental monitor¬ ing could provide evidence of activities conducted within those facilities. These analytical techniques included; ra¬ diometric measurements, isotopic measurements by both bulk and particle, accelerator mass spectrometry, and el¬ emental screening. Several laboratories participated, us¬ ing various analytical techniques. The consensus opin¬ ion of the evaluators was that environmental monitoring could, indeed, provide useful evidence. Accordingly, the IAEA board of governors has approved the implementa¬ tion of environmental monitoring as one means of strengthening safeguards. Of interest to LANL was the fact that actinides, removed from the vicinity of nuclear facilities, in any one of several matrices, could show very clearly the specific activities of that particular facil¬ ity. Reactors, reprocessing plants, and enrichment plants were monitored. One item of particular importance emerged; could our QA/QC program (Cappis et al 1994) hold up to the scru¬ tiny of the IAEA if we found evidence of a member states non-compliance with a treaty? We expressed our con¬ cern to the IAEA, which was expressing some of the same doubts. This lead to confirmation that suitable per¬ formance based evaluation standards are not available, so the IAEA initiated a program to produce standards of their own. Our long history of actinide chemistry and measurement will be utilized in this program along with the participation of laboratories in at least three other countries. Rocky flats environmental monitoring The Rocky Flats Plant (RFP) near Denver, Colorado, built nuclear weapons components for the USDOE. While the plant was in operation, inadvertent releases of U and Pu occurred. LANL was commissioned to investigate the results of very intense remediation and containment efforts. Retention ponds, diversion structures, and ditches were used for containment. LANL measured the U and Pu content of waters, soils and sediments over a three year period to determine the success of these ef¬ forts. It should be noted that both inflow and outflow from a creek which passed through the plant boundaries was also monitored. The creek contained fallout Pu and natu¬ ral uranium coming in to the plant, with slightly reduced concentration of U on the outflow. The outflow did, however, contain an altered isotopic signature from fall¬ out. It should also be noted that the Pu 240/239 ratio characterized for fallout (0.169 versus 0.18), was slightly different from the world integrated average due to local inputs (Efurd et al 1994; Krey et al 1976). Another observation made during this experiment was that, as expected, anthropogenic U was much more visible in surface waters than in soils or sediments. As pointed out previously, the natural U background in soil quickly dilutes out an anthropogenic source. Pu, how¬ ever, was more visible in the sediments. There was ap¬ proximately 50X more Pu in a gram of sediment than in one liter of water (Rokop et al 1994). This is explained by the fact that Pu is adsorbed onto clay particles in the water, which settle as sediments with time. It is unclear whether the Pu is fixed into the sedimentary layers or if it can move from position to position. Experience with sampling from sedimentary clay cores suggests that it stays fixed, at least somewhat, as sedimentary layers from pre-nuclear history, before 1935, do not show Pu from fallout or any other source (Rokop, Efurd, Roensch & Poths, unpublished studies). While the retention ponds and ditches showed the presence of Rocky Flats product, the remediation efforts were indeed very successful. The concentration of RFP actinides, with one exception , were all far below the very low standards set by the State of Colorado. The one exception, summer-time increase in Pu content of one of the retention ponds, turned out to be due to a near-by volleyball game disturbing soil previously contaminated. This water was retained and remediated until testing demonstrated the removal of Pu. No excursions above, or even close to, Colorado limits were observed outside the plant boundaries. Conclusion Isotopic signatures are a powerful tool that can be uti¬ lized for a wide variety of new applications. Environ¬ mental, industrial, and political spheres will be strongly influenced by their use. This tool is extremely flexible and its use is limited only by imagination. Acknowledgements: The authors would like to dedicate this paper, and the oral presentation it follows, to Professor John de Laeter of Curtin University of Technology, for his long and distinguished career. Three words come to mind when we think of John, gentleman, eloquent, and integrity. References Efurd D W, Rokop D J & Perrin R E 1992 Actinide determina¬ tion and analytical support for water characterization and treatment studies at rocky flats, LA-UR-93-917. Annual 89 Journal of the Royal Society of Western Australia, 79(1), March 1996 Report. Los Alamos national Laboratory, Los Alamos, New Mexico, 29-75. Edfurd D W, Rokop D J & Perrin R E 1993 Characterization of the radioactivity in surface- waters and sediments collected at Rocky Flats Facility, LA-UR-4373. Annual Report. Los Alamos National Laboratory, Los Alamos, New Mexico, 7- 10. Rokop D J, Perrin R E, Knobeloch G W, Armijo V M & Shields W R 1982 Thermal ionization mass spectrometry of uranium with electrodespostion as a loading technique. Analytical Chemistry 54:957-960. Edfurd D W, Rokop D J & Roensch F R 1994 Measurement of 240Pu/239/Pu and 241Pu/239Pu atom ratios in soils repre¬ sentative of global fallout in Colorado, LA-UR-94-4200. Los Alamos National Laboratory, Los Alamos, New Mexico, 8. Edfurd D W, Poths H, Rokop D J, Roensch F R & Olson R L 1995 Isotopic fingerpriting of plutonium in surface soil samples collected in Colorado. Los Alamos National Labora¬ tory, Los Alamos, New Mexico. Krey P W, Hardy E P, Pachucki C, Rourke F, Coluzza J & Benson W K 1976 Mass isotopic composition of global fall¬ out plutonium in soil, in Proceedings of a symposium on transuranium nuclides in the environment. Atomic Energy Agency Report STI/PU410 San Francisco 17-21. Cappis J H, Rokop D J & Dalton S A 1994 LANL-NWAL-QA, LA-UR-94-1133, Los Alamos National Laboratory, Los Alamos, New Mexico, 3-40. Rokop D J, Edfurd D W & Perrin R E 1994 Actinide determina¬ tion and analytical support for the characterization of envi¬ ronmental samples, IAEA-SM-333/99, Vienna, 475-483. 90 Journal of the Royal Society of Western Australia, 79:91-96, 1996 Stable heavy isotopes in human health B L Gulson Graduate School of the Environment, Macquarie University, Sydney NSW 2109: CSIRO/EM, North Ryde, NSW 2113 Abstract Applications of heavy stable isotopes in the field of human health are rather limited compared with applications employing radioactive isotope tracers. The elements reviewed in this paper encompass zinc, copper, iron, calcium, selenium and lead. Most research for the first five metals has addressed metabolic aspects, specifically fractional absorption ("bioavailability"). Lead iso¬ topes, although employed for pioneering pharmacokinetic modelling research in the early 1970s, have not enjoyed widespread acceptance until more recent times. An example is given of the use of lead isotopes to detect changes to blood lead from dietary sources and the contribution of skeletal lead to blood lead. Introduction Most emphasis on isotopic methods in human health has been directed towards the use of light stable isotopes of carbon, nitrogen, oxygen and hydrogen for body com¬ position, energy metabolism and macronutrient metabo¬ lism (Jones 1990). So-called "mineral" (?heavy) stable isotopes have been commonly employed as short-lived radioactive tracers. Limited use of thermal ionisation mass spectrometry (TIMS) and inductively coupled plasma mass spectrometry (ICP-MS) as well as fast atom bombardment mass spectrometry (FAB-MS) has also largely been directed towards metabolic studies, specifi¬ cally bioavailability and, in the case of lead, evaluation of sources and pathways. Part of the reason for the limited use of these methods may be a narrowing of research into highly specialised fields, poor communication be¬ tween disciplines, and arrogance of certain groups. An¬ other reason for the limited use of isotopic methods in the health field is because of the perceived high cost of the measurements as a result of the high cost of the en¬ riched isotopes. This problem is exacerbated by the use of low precision mass spectrometry (ICP-MS), which makes it necessary to use large doses of enriched iso¬ topes. In this paper, 1 will briefly review the use of stable heavy isotopes in human health, although the classifica¬ tion of the elements discussed may be disputed by some as not "heavy" and perhaps the term "mineral should be employed. Elements reviewed are calcium, copper, iron, selenium and zinc with respect to; bodily functions, isotopes, methods of analysis, isotopic research and main references. Information on bodily functions is mainly taken from Florence & Setright (1994). © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 Calcium (Ca) Bodily Functions: An adult body contains 1 kg Ca, 99% of which is located in bones and teeth in the form of apatite. The other —1% is found in intra- and extracellu¬ lar fluids, where it plays a vital role in directing cell func¬ tions and nerve impulses. The Ca concentration in blood serum is critical and lies between narrow limits of 90-100 mg/L1. Parathyroid hormone and Vitamin D regulate Ca intake. If the intake is too low, Ca is "resorbed" from bone stores. If bone depletion continues for extended periods, this condition can give rise to osteoporosis. Maintenance of Ca homeostasis during pregnancy and lactation is critically important for females. Isotope* 40 42 43 44 46 49 % Abundance 96.9 0.65 0.14 2.08 0.03 0.19 * Isotopes in bold italic are those primarily used in investiga¬ tions Methods: TIMS/ Fast Atom Bombardment-MS. Isotopic Research: Fractional Absorption (FA) especially in the study of bone turnover status/ bone loss and osteoporosis. In metabolic investigations of bioavailability (fractional absorption, FA; the amount of the substance taken up by the body compared with the total amount introduced to the body) employing dual isotopic meth¬ ods, one of the isotopes is introduced orally and the other by intravenous injection. For example, ^Ca may be in¬ troduced orally in milk and, a short time later, the 42Ca isotope is injected. Blood and urine samples are col¬ lected serially, commonly over 24 hours. For example, in a study of Chinese children on a low Ca diet of 359 mg d ', the FA was 63%; with a high Ca diet of 862 mg d1, the FA was 55% (Lee et al 1994). In Caucasian children, the FA was --*40% (Lee et al 1994). In a study of one male and one female subject, Price et al (1990) measured a FA of 70%. Abrams (1993) employed TIMS to determine the rate of Ca deposition in bone and the size of the ex¬ changeable Ca pool in bone in girls at puberty, the time of maximum growth associated with the peak of bone mass accumulation and Ca retention. 91 Journal of the Royal Society of Western Australia, 79(1), March 1996 Main References: Abrams et al (1991, 1993, 1994), TIMS; Eastel et al (1989), TIMS; Lee et al (1994), TIMS; Miller et al. (1989), FAB-MS; Moore et al (1985), TIMS; Price et al (1990), TIMS; Smith et al (1985), FAB-MS; Yergey et al (1987,1990,1994), TIMS. Copper (Cu) Bodily Functions: A 70-kg human contains about 80 mg Cu, mostly located in muscle and liver. Copper (and Fe) are the primary oxygen carriers in cells. A Fe-Cu-Zn enzyme, cytochrome oxidase, is present in mitochondria and is responsible for the catalysis of oxygen to water, an impotant step in cellular metabolism. Other Zn enzymes in mitochondria act as antioxidants to remove free radi¬ cals, which could otherwise promote cellular damage. Deficiency in Cu can result in anaemia, osteoporosis, re¬ productive failure, and heart failure. Isotope 63 65 % Abundance 69.1 30.9 Method of analysis: TIMS. Isotopic Research: Bioavailability (FA) for an adequate dietary intake of 1.68 mg d ’, was 36%; on a low intake of 0.79 mg d1, the FA was 56%; and on a high intake of 7.5 mg d_l, the FA was 12%. There is a Cu balance (ho¬ meostasis) with 0.8 mg Cu d'1, and the regulation of Cu absorption and endogenous loss gives protection from Cu deficiency on the one hand, and toxicity on the other. Main Player: Tumlund et al (1989). Iron (Fe) Bodily Functions: In an adult male, *=4 g Fe is distrib¬ uted among the protein, haemoglobin (73%), ferritin and hemosiderin (12%), and myoglobin (14%). Haemoglobin is the main constituent of red blood cells. Its main func¬ tion is to transport oxygen from the lungs in arterial blood to various organs and tissues, and return in ve¬ nous blood carrying some carbon dioxide. Isotope 54 56 57 58 % Abundance 5.8 91.7 2.14 0.28 Methods of Analysis: TIMS/ICP-MS/FAB-MS. Isotopic Research: When added to the diet, there was a FA of ^9% in the elderly, using faecal composites as the sampling medium (Tumlund 1983). Iron bioavailability in blood from Fe tablets was measured by Hansen et al (1992). Iron distribution among the protein species trans¬ ferrin, ferritin, and other haemoproteins was measured in liver and heart by Stuhne-Sekalec et al (1992). Main References: Tumlund (1983), TIMS; Stuhne-Sekalec et al (1992), ICP-MS; Hansen et al (1992), FAB-MS. Lead (Pb) Bodily Function: No natural physiological use. Isotope 204 206 207 208 % Abundance 1.4 24.1 22.1 52.4 Methods of Analysis: TIMS/ ICP-MS. Main References: Pharmacokinetic Modelling, Rabinowitz et al (1976), Heusler-Bitschy et al (1988); pregnancy/ personal monitoring, Manton (1985,1992); efficacy of che¬ lating agents. Smith et al (1994); source of Pb in humans, Yaffee et al (1983), Tera et al (1985), Faccheti (1989), Campbell & Delves (1989); Biokinetics of lead in preg¬ nancy, Gulson et al (1995 a,c, 1996c), Gulson & Calder (1995). Example of the Use of Lead Isotopes: In a project on the "biokinetics of lead in human pregnancy", the primary objective was to establish if lead is released from the ma¬ ternal skeleton during pregnancy and lactation (Gulson et al 1995c). The main cohort were subjects from coun¬ tries outside of Australia because the lead isotopic ratios in their skeleton are different from those in the "long¬ term" Australian population. The Australian population has 206Pb/204Pb ratios generally less than 17.0, whereas people from other countries (and especially Eastern Eu¬ rope), have ratios >18.0. If a subject conceives and East¬ ern European lead, for example, is observed in the blood, then this is an indication that the lead is derived from maternal skeletal stores and not from Australian envi¬ ronmental sources such as diet , soil or dust. The subjects in this study who conceive had a matched non-pregnant control. During pregnancy, the subjects and controls were monitored monthly for blood and urine samples, and quarterly for environmental samples of house dust, drinking water, and a 6-day duplicate diet. In one control subject, blood samples exhibited a sud¬ den change in isotopic composition and increased lead concentration (Figure 1), in the direction which was pre¬ dicted for a pregnant subject. Upon detailed questioning about dietary changes, the only plausible change was daily consumption of several glasses of beverages made with water from a newly-acquired Russian samovar. Testing of the samovar water showed it to have lead concentrations at least 20 times above fully-flushed kitchen tap water and isotope ratios consistent with Rus¬ sian values. Besides negating her role as a control, the other dis¬ concerting aspect of these results was the rapidity of change in both isotopic composition and blood lead con¬ centration. The implication of these rapid changes is that the bioavailability of the lead in the water is very high. The other features of the data shown in Figure 1 are the changes in lead in blood when moving from one country to another. The blood lead concentrations in this subject are low when compared with a value of <10 pg Pb dU, the Australian National Health and Medical Re¬ search Council National Goal for all Australians. There was little change in Pb until the samovar incident. In contrast, the isotopic composition showed a rapid de¬ crease for all subjects from Eastern Europe when they 92 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 1. Time-series variation of isotope ratio, expressed as the abundance of the 206Pb to 204Pb, and blood lead (PbB) of a subject from Eastern Europe. The marked increase in 2t)6Pb/2()4Pb ratio and PbB was traced to consumption of water from a Russian samovar. arrive in Australia, arising from the exchange of Euro¬ pean lead with that in the Australian environment (Gulson et al 1995c). The results of this study demon¬ strated, for the first time, the quantitative contribution of skeletal lead to blood lead. In the case of the subject in Figure 1, skeletal lead is estimated to have a 206Pb/204Pb ratio of 18.0 and this is exchanging with Australian lead with a 2(*Pb/2(nPb of 17.0. At 200-300 days, the 2<*Pb/204Pb in the blood of this subject was ^17.4, which means that approximately 40% of the lead in her blood was deriving from skeletal sources. Selenium Bodily Functions: Selenium is an important antioxidant. For example, glutathione peroxidase converts 1 1202 to wa¬ ter, reducing cell destruction and hence combating de¬ generative diseases. There are 13 Se-containing proteins involved in the correct functioning of the thyroid gland. Selenium is also a strong inhibitor of platelet aggrega¬ tion, a process involved in strokes, heart disease and can¬ cer metastases. Isotope 74 76 77 78 80 82 % Abundance 0.9 9.0 7.5 23.5 50 9.0 Methods of Analysis: ICP-MS. Isotopic Research: FA is up to 70% when present as selenomethionine (organic form in food) compared with 40% in inorganic form. Selenium acts synergistically with Vitamin C as an antioxidant. Main References: Kasper et al (1984), Martin et al. (1988, 1989a, b), Sirichakwal et al (1985), Solomons et al (1986). Zinc Bodily Functions: In an adult, there is ~2g Zn in muscle, bone and tissues. It is essential for bone formation, wound healing, immune system fuction, ageing, sexual function, anexoria nervosa, and reduction of cadmium toxicity. Over 200 enzyme systems require Zn for struc¬ tural integrity or catalysis, including synthesis of DNA and RNA. Isotope 64 66 67 68 70 % Abundance 48.8 27.8 4.1 18.6 0.62 Methods of Analysis: TTMS/ICP-MS/FAB-MS. Isotopic Research: A study by August et al (1999) indi¬ cated that for a diet with Cu addition, FA was «=40% in the young and ^=21% in the elderly; for a diet with no addition of Cu, FA was «=31% in the young and ==17% in the elderly (Turnlund et al 1986). Later work, however, by Couzy et al (1993) using radioisotope methods indi¬ cated that age differences were not significant for Zn FA. In another study, Fairweather-Tait et al (1992) showed that there was no difference in the FA (^30%) of Zn for white compared with wholemeal bread. Jackson et al (1984) and Lowe et al (1993) investigated Zn biokinetic modelling (definition of Zn compartments in the body). Main Players: Turnlund et al (1984, 1986, 1991), TIMS; Jackson et al (1984), TIMS; Couzy et al (1993), radio¬ isotopes; Knudsen et al (1995), ICP-MS; Eagles et al (1989); Janghorbani et al (1984, 1990), ICP-MS. Concluding comments This brief review illustrates that there is considerable scope for investigations in human health using isotopic techniques with stable isotopes. One concern that I have is whether or not many of the metabolic experiments for bioavailability are really applicable to "real life". The elements are commonly introduced as pure compounds and these species may not be in the appropriate bioavailable form. For example, the FA of Fe was mea¬ sured in one experiment as ^9% in elderly subjects. It is fairly well established that the bioavailability of Haeme 93 Journal of the Royal Society of Western Australia, 79(1), March 1996 Fe varies according to the food item: in meat it is ^20%, in fish =6%, in cereals and vegetables **2-5%, and in breast milk the bioavailability is ^50% versus <5% for formula. Furthermore, drinking alcohol with meals en¬ hances bioavailability of Fe (Florence & Setright, 1994). Acknowledgements: The Biokinetics of Lead in Human Pregnancy study is largely financed by a US National Institute of Environmental Health Sciences Grant NO1-ES-05292. The research at Broken Hill was partly supported by the NSW Environment Protection Authority. References General References Florence T M & Setright R T 1994 The Handbook of Preventive Medicine. Kingsclear Books, Sydney. Jones P J H 1990 Stable isotopes in nutrition research: historical perspective and overview. Canadian Journal of Physiology and Pharmacol 68:935-940. Calcium Abrams S A 1993 Pubertal changes in calcium kinetics in girls assessed using 42Ca. Pediatrics Research 34: 455-459. Abrams S A, Yergey A L & Heaney R P 1994 Relationship be¬ tween balance and dual tracer isotopic measurements of cal¬ cium absorption and excretion. Journal of Endocrinology and Metabolism 79:965-969. Abrams S A, Sidbury J B, Muenzer J, Esteban N V, Vieira N E & Yergey A L 1991 Stable isotopic measurement of endogenous fecal calcium excretion in children. Journal of Pediatric Gas¬ troenterology and Nutrition 12:469-473. Abrams S A, Esteban N V, Vieira N E, Sidbury J B, Specker B L & Yergey A L 1992 Developmental changes in calcium kinet¬ ics in children assessed using stable isotopes. Journal of Bone and Mineral Research. 7(3):287-293. Abrams S A, Schanler R J, Yergey A L, Vieira N E & Bronner F 1994 Compartmental analysis of calcium metabolism in very- low-birth-weight infants. Pediatric Research 36:424-428. Blumsohn A, Morris B & Eastell R 1994 Stable strontium absorp¬ tion as a measure of intestinal calcium absorption: compari¬ son with the double-radiotracer calcium absorption test. Clinical Science Colch 87: 363-369. Bucuvalas J C, Heubi J E, Specker B L, Gregg D J, Yergey A L & Vieira N E 1990 Calcium absorption in bone disease associ¬ ated with chronic cholestasis during childhood. Hepatology 12:1200-1205. Eastell R, Vieira N E, Yergey A L & Riggs B L 1989 One-day test using stable isotopes to measure true fractional calcium ab¬ sorption. Journal of Bone and Mineral Research 4:463-468. Goans R E, Abrams S A, Vieira N E, Marini J C, Perez M D & Yergey A L 1995 A three-hour measurement to evaluate bone calcium turnover. Bone 16:33-38. Hillman L S, Tack E, Coveil D G, Vieira NE & Yergey A L 1988 Measurement of true calcium absorption in premature in¬ fants using intravenous 4,’Ca and oral ^Ca. Pediatric Re¬ search 23:589-594. Kent G N, Price R J, Gutteridge D H, Rosman K J R, Smith M, Allen J P, Hichling C J & Blakeman S L 1991 The efficiency of intestinal calcium absorption is increased in late pregnancy but not established in lactation. Tissue International 48:293- 295. Lee W T, Leung S S, Fairweather-Tait S J, Leung D M, Tsang H S, Eagles J, Fox T, Wang S H, Xu Y C & Zeng W P 1994 True fractional calcium absorption in Chinese children measured with stable isotopes (42Ca and ^Ca). British Journal of Nutri¬ tion 72:883-897. Lee W T, Leung S S, Xu Y C, Wang S H, Zeng W P, Lau J & Fairweather-Tait S J 1995 Effects of double-blind controlled calcium supplementation on calcium absorption in Chinese children measured with stable isotopes (42Ca and ^Ca). Brit¬ ish Journal of Nutrition 73:11-21. Miller J Z, Smith D L, Flora L, Peacock M & Johnston C C 1989 Calcium absorption in children estimated from single and double stable calcium isotope techniques. Clinica Chemica Acta 183:107-114. Moore L J, Machlan L A, Lim M O, Yergey A L & Hansen J W 1985 Dynamics of calcium metabolism in infancy and child¬ hood. I. Methodology and quantification in the infant Pedi¬ atric Research 19:329-334. Price R I, Kent G N, Rosman K J R, Gutteridge D H, Reeve J, Allen J P, Stuckey B G A, Smith M, Guelfi Hickling C J & Blakeman S L 1990 Kinetics of intestinal calcium absorption in humans measured using stable isotopes and high-preci¬ sion thermal ionisation mass spectrometry. Biomedical and Environmental Mass Spectrometry 19:353-359. Smith D L, Atkin C & Westenfelder C 1985 Stable isotopes of calcium as tracers: methodology. Clinica Chemica Acta 146:97-101. Specker B L, Vieira N E, O'Brien K O, Ho M L, Heubi J F, Abrams S A & Yergey A L 1994 Calcium kinetics in lactating women with low and high calcium intakes. American Jour¬ nal of Clinical Nutrition 59:593-599. Yergey A L 1994 Issues in constant tracer infusion and mineral metabolism. Advances in Experimental Medicine and Biol¬ ogy 352:279-290 Yergey A L, Abrams S A, Vieira N E, Aldroubi A, Marini J & Sidbury J B 1994 Determination of fractional absorption of dietary7 calcium in humans. Journal of Nutrition 124:674-682. Yergey A L, Abrams S A, Vieira N F, Eastell R, Hilliman L S & Covell D G 1990 Recent studies of human calcium metabo¬ lism using stable isotopic tracers. Canadian Journal of Physi¬ ology and Pharmacology 68:973-976. Yergey A L, Vieira N F & Coveil D G 1987 Direct measurement of dietary fractions absorption using calcium isotopic tracers. Biomedical and Environmental Mass Spectrometry 14:603- 607. Copper Turnlund J R, Keyes W R, Anderson H L & Acord L L 1989 Copper absorption and retention in young men at three lev¬ els of dietary copper by use of the stable isotope 65Cu. American Journal of Clinical Nutrition 49:870-878. Iron Hansen C, Werner E & Kaltwasser J P 1992 Measurement of iron bioavailability by means of stable ^Fe and mass spec¬ trometry. Physics in Medicine and Biology 37:1349-1357. Turnlund J R 1983 Use of enriched stable isotopes to determine bioavailability of trace elements in humans. Science of the Total Environment 28:385-392. Stuhne-Sekalec L, Xu S X, Parkes J G, Olivieri N F & Templeton D M 1992 Speciation of tissue and cellular iron with on-line detection by inductively coupled plasma-mass spectrometry. Analytical Biochemistry 205:278-284. Lead Campbell M J & Delves H T 1989 Accurate and precise determi¬ nation of lead isotope ratios in clinical and environmental samples using inductively coupled plasma mass spectrom¬ etry. Journal of Analytical Atomic Spectrometry 4:235-236. Faccheti S 1989 Lead in petrol. The isotopic lead experiment. Acc hem Research 22:370-374. Gulson B L 1996 Tooth analyses of sources and intensity of lead exposure in children. Environmental Health Perspectives 104:306-312. Gulson B L & Calder I C 1995 Lead mobilisation during preg¬ nancy. Medical Journal of Australia 163:447. Gulson B L & Wilson D 1994 History of lead exposure in chil¬ dren revealed from isotopic analyses of teeth. Archives of Environmental Health 49:279-283. 94 Journal of the Royal Society of Western Australia, 79(1), March 1996 Gulson B L, Lee T H, Mizori K J & Eschnauer H 1992 The appli¬ cation of lead isotope ratios to determine the contribution of the tin-lead capsule to the lead content of wine. American Journal of Enology Viticulture 43:180-190. Gulson B L, Howarth D, Mizon K J, Korsch M J & Davis J J 1994a The source of lead in humans from Broken Hill mining community. Environmental Geochemistry and Health 16:19- 25. Gulson B L, Law A J, Korsch M J & Mizon K J 1994b Effect of plumbing systems on lead content of drinking water and contribution to lead body burden. Science of the Total Envi¬ ronment 144:279-294. Gulson B L, Mizon K J, Law A J, Korsch M J & Davis J J 1994c Source and pathways of lead in humans from Broken Hill mining community - an alternative use of exploration meth¬ ods. Economic Geology 89:889-908. Gulson B L, Howarth D, Davis J J, Mizon K J, Korsch M J & Law A J 1994d Lead Bioavailability in the Environment of Chil¬ dren: Blood Lead Levels in Children can be Elevated in a mining Community. Archives of Environmental Health 49:326-331. Gulson B L, Davis J J & Bawden-Smith J 1995a Paint as a source of recontamination of houses in urban environments and its role in maintaining elevated blood leads in children. Science of the Total Environment 164:221-235. Gulson B L, Davis J J, Mizon K J, Korsch M J & Bawden-Smith J 1995b Sources of lead in soil and dust and the use of dust fallout as a sampling medium. Science of the Total Environ¬ ment 166:245-262. Gulson B L, Mahaffey K R, Mizon K J, Korsch M J, Cameron M & Vimpani G 1995c Contribution of tissue lead to blood lead in adult female subjects based on stable lead isotope meth¬ ods. Journal of Laboratory and Clinical Medicine 125:703- 712. Gulson B L, Mizon K J, Korsch M J & Howarth D 1996a Non- orebody sources can be significant contributors to blood lead in some children from a major lead mining community. Sci¬ ence of the Total Environment 181:223-230. Gulson B L, Mizon K j, Korsch M J & Howarth D 1996b Impact on blood lead in children and adults following relocation from their source of exposure and contribution of skeletal tissue to blood lead. Bulletin of Environmental Contamina¬ tion and Toxicology 56:543-550. Gulson B L, Pisaniello D, McMichael A J, Mahaffey K R, Luke C, Mizon K J, Korsch M J, Ashbolt R, Vimpani G & Pederson D C 1996c Stable lead isotope profiles in smelter and general ur¬ ban communities. Comparison of biological and environ¬ mental measures. Environmental Geochemistry and Health: in press. Heusler-Bitschy S, Knutti R & Schlatter Ch 1988 Inter-individual variability of the kinetics of lead in man. In: Trace Element Analytical Chemistry in Medicine and Biology, (eds Walter de Gruyter, Berlin, 627-634. Manton W I 1985 Total contribution of airborne lead to blood lead, British Journal of Industrial Medicine 42:168-172. Manton W 1 1992 Postpartum changes in maternal blood lead concentrations. Correspondence. British Journal of Industrial Medicine 49:671-672. Rabinowitz M B, Wetherill G W & Kopple J D 1976 Kinetic analysis of lead metabolism in healthy humans. Journal of Clinical Investigation 58:260-270. Smith D R, Markowitz M E, Crick J, Rosen J F & Flegal A R 1994 The effects of Succimer on the absorption of lead in adults determined using the stable isotope 2U4Pb. Environmental Re¬ search 67:39-53. Stauber J L, Florence T M, Gulson B L & Dale L S 1994 Percuta¬ neous absorption of inorganic lead compounds. Science of the Total Environment 145:55-70. Tera O, Schwartzman D W & Watkins T R 1985 Identification of gasoline lead in children's blood using isotopic analysis. Ar¬ chives of Environmental Health 40:120-123. Yaffe Y, Flessel C P, Wesolowski J J, Rosario A D, Giurguis G N, Matias V, Degarmo T E, Coleman G C, Gramlich J W & Kelly W R 1983 Identification of lead sources in Californian chil¬ dren using the stable isotope ratio technique. Archives of Environmental Health 38:237-245. Selenium Christensen M J, Jaghorbani M, Steinke F H, Istfan N & Young V R 1983 Simultaneous determination of absorption of sele¬ nium from poultry meat and selenite in young men: applica¬ tion of a triple stable-isotope method. British Journal of Nu¬ trition 50:43-50. Ducros V, Favier A & Guigues M 1991 Selenium bioavailability as selenite (/4Se) and as a selenium drug (76Se) by stable iso¬ tope methodology. Journal of Trace Elements and Electro¬ lytes in Health and Disease 5:145-154. Janghorbani M, Kasper L J & Young V R 1984 Dynamics of selenite metabolism in young men: studies with the stable isotope tracer method. American Journal of Clinical Nutri¬ tion 40:208-218. Janghorbani M, Martin R F, Kasper L J, Sun X F & Young V R 1990 The selenite-exchangeable metabolic pool in humans: a new concept for the assessment of selenium status. Ameri¬ can Journal of Clinical Nutrition 51:670-677. Kasper L J, Young V R & Janghorbani M 1984 Short-term di¬ etary selenium restriction in young adults: quantitative stud¬ ies with the stable isotope 74SeQl(2 '. British Journal of Nutri¬ tion 52:443-455. Mangels A R, Moser Veillon P B, Patterson K Y & Veillon C 1990 Selenium utilization during human lactation by use of stable-isotope tracers. American Journal of Clinical Nutrition 52: 621-627. Martin R F, Janghorbani M & Young V R 1988 Kinetics of a single administration of 74Se-selenite by oral and intravenous routes in adult humans. Parenter Enteral Nutrition 12:351- 355. Martin R F, Young V R, Blumberg J & Janghorbani M 1989a Ascorbic acid-selenite interactions in humans studied with an oral dose of 74Se03(2'1- American Journal of Clinical Nutri¬ tion 49:862-869. Martin R F, Janghorbani M & Young V R 1989b Experimental selenium restriction in healthy adult humans: changes in se¬ lenium metabolism studied with stable-isotope methodol¬ ogy. American Journal of Clinical Nutrition 49:854-861. Moser Veillon P B, Mangels A R, Patterson K Y & Veillon C 1992 Utilization of two different chemical forms of selenium during lactation using stable isotope tracers: an example of speciation in nutrition. Analyst 117:559-562. Patterson B H & Zech L A 1992 Development of a model for selenite metabolism in humans. Journal of Nutrition 122 Suppl 3:709-714. Reamer DC & Veillon C 1983 A double isotope dilution method for using stable selenium isotopes in metabolic tracer studies: analysis by gas chromatography /mass spectrometry (GC/ MS). Journal of Nutrition 113:786-792. Sirichakwal P P, Young V R & Janghorbani M 1985 Absorption and retention of selenium from intrinsically labeled egg and selenite as determined by stable isotope studies in humans. American journal of Clinical Nutrition 41:264-269. Solomons N W, Torun B, Janghorbani M, Christensen MJ, Young VR & Steinke FH 1986 Absorption of selenium from milk protein and isolated soy protein formulas in preschool children: studies using stable isotope tracer 74Se. Journal of Pediatric Gastroenterology Nutrition 5:122-126. Swanson C A, Reamer D C, Veillon C, King JC & Levander O A 1983 Quantitative and qualitative aspects of selenium utiliza¬ tion in pregnant and nonpregnant women: an application of stable isotope methodology. American Journal of Clinical Nutrition 38:169-180. Swanson C A, Patterson B H, Levander O A, Veillon C, Taylor P R, Helzlsouer K, McAdam P A & Zech L A 1991 Human 95 Journal of the Royal Society of Western Australia, 79(1), March 1996 [74Se]selenomethionine metabolism: a kinetic model. Ameri¬ can Journal of Clinical Nutrition 54:917-926. Veillon C, Patterson K Y, Button L N & Sytkowski A J 1990 Selenium utilization in humans-a long-term, self-labeling ex¬ periment with stable isotopes. American Journal of Clinical Nutrition 52:155-158. Zinc August D, Janghorbani M & Young V R 1989 Determination of zinc and copper absorption at three dietary Zn-Cu ratios by using stable isotope methods in young adult and elderly sub¬ jects. American Journal of Clinical Nutrition 50:1457-1463. Couzy F, Kastenmaver P, Mansourian R, Guinchard S, Munoz Box R & Dirren H 1993 Zinc absorption in healthy elderly humans and the effect of diet. American Journal of Clinical Nutrition 58:690-694. Davidsson L, Kastenmayer P & Hurrell R F 1994 Sodium iron EDTA [NaFe(III)EDTA] as a food fortificant: the effect on the absorption and retention of zinc and calcium in women. American Journal of Clinical Nutrition 60:231-237. Eagles J, Fairweather Tait S J, Mellon F A, Portwood D E, Self R, Gotz A, Heumann K G & Crews H M 1989 Comparison of fast-atom bombardment, thermal ionization and inductively coupled plasma mass spectrometry for the measurement of MZn/67Zn stable isotopes in human nutrition studies. Rapid Communications in Mass Spectrometry 3:203-205. Egan C B, Smith F G, Houk R S & Serfass R E 1991 Zinc absorp¬ tion in women: comparison of intrinsic and extrinsic stable- isotope labels. American Journal of Clinical Nutrition 53:547- 53. Fairweather Tait S J, Portwood D E, Symss L L, Eagles J & Minski M J 1989 Iron and zinc absorption in human subjects from a mixed meal of extruded and nonextruded wheat bran and flour. American Journal of Clinical Nutrition 49:151-155. Fairweather Tait S J, Fox T E, Wharf S G, Eagles J & Kennedy H 1992 Zinc absorption in adult men from a chicken sandwich made with white or wholemeal bread, measured by a double-label stable-isotope technique. British Journal of Nu¬ trition 67:411-419. Fairweather Tait S J, Jackson M J, Fox T E, Wharf S G, Eagles J & Croghan P C 1993 The measurement of exchangeable pools of zinc using the stable isotope 7,,Zn. British Journal of Nutri¬ tion 70:221-234. Istfan N W, Janghorbani M & Young V R 1983 Absorption of stable 70Zn in healthy young men in relation to zinc intake. American Journal of Clinical Nutrition 38:187-194. Istfan N, Murray E, Janghorbani M & Young V R 1983 An evalu¬ ation of the nutritional value of a soy protein concentrate in young adult men using the short-term N-balance method. Journal of Nutrition 113:2516-2523. Jackson M J, Jones D A, Edwards R H, Swainbank I G & Coleman M L 1984 Zinc homeostasis in man: studies using a new stable isotope-dilution technique. British Journal of Nu¬ trition 51:199-208. Knudsen E, Jensen M, Solgaard P, Sorensen S S & Sandstrom B 1995 Zinc absorption estimated by fecal monitoring of zinc stable isotopes validated by comparison with whole-body retention of zinc radioisotopes in humans. Journal of Nutri¬ tion 125:1274-1282. Lowe N M, Green A, Rhodes J M, Lombard M, Jalan R & Jack- son M J 1993 Studies of human zinc kinetics using the stable isotope 7uZn. Clin Sci Colch 84:113-117. Mason P M, Judd P A, Fairweather Tait S J, Eagles J & Minski M J 1990 The effect of moderately increased intakes of complex carbohydrates (cereals, vegetables and fruit) for 12 weeks on iron and zinc metabolism. British Journal of Nutrition 63:597- 611 Schuette S A, Janghorbani M, Young V R & Weaver C M 1993 Dysprosium as a nonabsorbable marker for studies of min¬ eral absorption with stable isotope tracers in human subjects. Journal of the American College of Nutrition 12:307-315. Sian L, Hambridge K M, Westcott J L, Miller L V & Fennessey P V 1993 Influence of a meal and incremental doses of zinc on changes in zinc absorption. American Journal of Clinical Nu¬ trition 58:533-536. Taylor C M, Bacon J R, Aggett P J & Bremner I 1991 Intestinal absorption and losses of copper measured using 65Cu in zinc- deprived men. European Journal of Clinical Nutrition 45:187- 94. Tumlund J R, King J C, Keyes W R, Gong B & Michel M C 1984 A stable isotope study of zinc absorption in young men: ef¬ fects of phytate and alpha-cellulose. American Journal of Clinical Nutrition 40:1071-1077. Tumlund J R, Durkin N, Costa F & Margen S 1986 Stable iso¬ tope studies of zinc absorption and retention in young and elderly men. Journal of Nutrition 116:1239-1247. Tumlund J R, Keyes W R, Hudson C A, Betschart A A, Kretsch M J & Sauberlich HE 1991 A stable-isotope study of zinc, copper, and iron absorption and retention by young women fed vita¬ min B-6-deficient diets. American Journal of Clinical Nutri¬ tion 54:1059-1064. 96 Journal of the Royal Society of Western Australia, 79:97-102, 1996 Lead isotopes and pollution history K J R Rosman & W Chisholm Deptartment of Applied Physics, Curtin University of Technology, Bentley, WA 6102 Abstract Lead is one of the seven metals of antiquity and its production is highly correlated with the development of industrial civilisations. This paper reviews the awakening during the 1960s of modern humans to the environmental and health problems associated with the use of lead. The difficulty of measuring lead received little attention until the mid 1960s when it was realised by geochemist C C Patterson and colleagues at the California Institute of Technology that the world was immersed in a lead mist. This awareness prompted analysts to develop clean-air laboratories and to take special precautions when collecting and analysing samples for lead. Thermal ionisation mass spectrometry has played a vitally important role in this process by offering a high-sensitivity detection tool, and the means of uniquely identifying artifact lead contamination through its iso¬ topes. Today this tool is being used to identify the origin of lead in natural archives such as the Greenland and Antarctic ice sheets. This paper discusses the development of this field. Introduction Significant lead production began about 5000 years ago with the discovery of cupellation. Lead sulphide ores were smelted to produce lead-silver alloy which yielded metallic silver on oxidation. Smelting led to lo¬ cal pollution but also released lead to the atmosphere. However it was only during Greek and Roman times, about 2500 years ago, that atmospheric pollution from this source reached levels where evidence of the fossil remnants could be detected in Greenland ice (Hong et al. 1994). Lead has been associated with human health prob¬ lems since antiquity and may have contributed to the fall of Rome (Nriagu 1983). Its toxic effects on humans are now well understood and childhood lead poisoning is of great concern (Mushak et al. 1989; McMichael et al. 1988). Lead is a non-degradable poison and has been produced at an increasing rate during the twentieth century. Vast quantities have been used in paint and as an anti-knock agent in petrol, creating a problem of global dimensions which this and future generations must resolve (Rabin 1989; Nriagu 1990; Needleman 1991). Although lead has created tremendous problems for mankind, the formation of its isotopes by radioactive de¬ cay allowed the first accurate determination of the Earth's age (Patterson 1956). The array of isotopic abun¬ dances found in crustal rock and lead ore has been the key to the identification of anthropogenic lead in ancient ice and is the basis of a method for tracing lead and associated pollution around the globe. The two primary instruments used for these studies are the mass spec¬ trometer and the ultra-clean sample preparation labora¬ tory. The last three decades have witnessed dramatic improvements in mass spectrometer sensitivity ( >10” for some elements) through the development of new ionisation methods (Cameron et al. 1969; Rosman et al. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 1984) and vastly improved techniques for sampling and preparing small samples for analysis (Patterson & Settle 1976). These techniques have now been refined to the point where the isotopic composition of lead in remote snow and ice can be reliably measured (Rosman et al. 1993). This paper briefly reviews the motivation for con¬ structing a lead pollution history for the global atmo¬ sphere and shows how the measurement of lead isotope abundances complements concentration measurements by providing a method to identify the source of pollu¬ tion. Pollution History A variety of natural archives have been used to con¬ struct pollution histories; these include river, lake and marine sediments, peat bogs, glaciers, the polar ice caps and tree rings. In these archives, the anthropogenic spe¬ cies are trapped in layers and remain essentially isolated from each other. A pollution history can be established by measuring the species and the corresponding deposi¬ tion age. Time scales for these archives vary from years to millennia. The impetus for constructing a pollution history of lead for the Northern Hemisphere came from Clair C. Patterson of the California Institute of Technology (CIT), who, together with colleagues T Chow and M Tatsumoto, found that industrial lead was having a profound influ¬ ence on the lead content of the oceans and atmosphere. Although the measurements of lead in sea water in their earlier studies were later found to be in error by a large amount due to contamination problems, their conclu¬ sions were correct. Following the publication of a paper entitled Contaminated and Natural Environments of Man (Patterson 1965), Patterson directed his research effort to¬ wards understanding the impact that anthropogenic lead was having on the biogeochemical cycle of lead. This paper created much public interest, controversy amongst scientists, and corporate concern. Differences of opinion 97 Journal of the Royal Society of Western Australia, 79(1), March 1996 between scientists often arose from a lack of awareness of the real difficulties of analysing lead at low concentra¬ tion, either in the field or in the laboratory. Ignorance of these problems still persists today. Patterson began his 1965 paper: "A prevailing belief is that industrial and natural sources contribute more or less equal amounts of lead to the body burdens of the general popu¬ lation . He proceeded to demonstrate with geo¬ chemical arguments, using the relatively poor data of the time, that the average resident of the U.S.A. had a body burden for lead one hundred times above natural levels. Questions raised in his paper were the subject of CIT research for the following three decades. One recom¬ mendation arising from his study was the need to estab¬ lish "natural" lead levels as opposed to the "typical" lev¬ els found in modern environments, and so a lead pollu¬ tion history for the Northern Hemisphere extending back in time almost three millennia was constructed. The concentration of lead was measured in dated lay¬ ers of Greenland snow and ice extending from 1965 AD back to 800 BC (Murozumi et al. 1969). This was a re¬ markable accomplishment for its time. The time series covering the period 1753-1965 was established by analysing preserved layers of snow and ice taken from the walls of trenches and shafts at Camp Century (77° 10'N, 61°08’W, elevation 1.87 km ) and a site 80 km away in north west Greenland. The ice sample dated 800 BC was collected from Camp Tuto (76°25'N, 68°20'W). These measurements initially caused contro¬ versy among geochemists because they raised doubts about the quality of published data on polar snow, but the results were eventually confirmed and extended ( Ng & Patterson, 1981; Boutron et al. 1991; Candelone et al. 1995). Figure 1 summarises the results of these measure¬ ments. Year Figure 1. Changes in the concentration of lead in Greenland snow and ice dating from 3000 BC to the present. Adapted from Murozumi et al. (1965) - ellipses; Boutron et al. (1991) - triangles; Candelone et al. 1995 - squares, Ng & Patterson 1981 - inverted triangle. To identify the source of lead emissions to the atmo¬ sphere, it is necessary to tag each source in some way. Isotopic fingerprinting can be very effectively used for lead when there are natural variations between sources. Early workers were well aware of the potential of this technique and were successful in identifying sources of local pollution (Chow & Johnstone 1965), but only re¬ cently has it been possible to use the technique to investi¬ gate sources of atmospheric pollution on a hemispheric scale (Rosman et al. 1993). Lead Isotopes Lead has four naturally occurring stable isotopes, 204Pb which was only formed by nucleosynthesis and three others which are also end products of radioactive decay chains: 238U =* 206Pb , 235U => 207Pb and 230Th => 208Pb with halflives 4.5 109, 0.71 109 and 14 109 years respectively. Consequently, lead isotopes in rocks and minerals of dif¬ ferent chemical composition and in lead ore bodies of different ages, which are the source of industrial lead, display large abundance variations that can easily be measured in a mass spectrometer. Soddy, who introduced the term "isotope", suggested in 1913 that because lead wTas formed by the radioactive decay of uranium and thorium, then minerals rich in these elements would have lower and higher atomic weights, respectively, than common lead. This prediction was confirmed within a year by number of reports (Table 1) and provided early confirmation of the existence of isotopes. Table 1 Early evidence of natural variations in the abundance of lead isotopes (from Aston 1942). Minerals: aCeylon thorite, bCamotite, ‘Norwegian cleveite, d Norwegian thorite. atomic weight Investigators high U common Pb high Th ~207.19 Soddy & Hyman (1914) Curie (1914) 206.36b 207.6943 Richards & Lembert (1914) 206. 08c Lead Isotopes in the Environment Interest in the use of isotope abundance variations for tracing sources of industrial lead began in the early 1960s (Tatsumoto & Patterson 1963; Chow & Johnstone 1965) and evolved from earlier geochemical studies by Chov/, Patterson and Tasumoto who had all been associated with the same CIT laboratory. Since this time, lead iso¬ topes have been used as natural tracers on a number of occasions, but the number of studies have been relatively few considering the potential of this approach for solv¬ ing environmental problems. One particularly signifi¬ cant early study was that by Chow et al. (1975) who as¬ sessed lead isotopic abundance variations in aerosols, gasoline and soil in various countries by spot sampling during the period 1964 - 1974. They also monitored tem¬ poral variation in San Diego (USA) aerosols over the same period which revealed a clear trend in isotopic composition. Since this time, pollution histories that include isoto¬ pic data have been published for a range of different natural archives, including for example river sediments (Shirahata et al. 1980), lake sediments (Ritsen et al. 1994) marine sediments (Hirao et al. 1986), coral (Shen & Boyle 1987) and snow (Rosman et al. 1993). 98 Journal of the Royal Society of Western Australia, 79(1), March 1996 Isotopic Fingerprinting in Snow and Ice Pollution histories based on lead in recent snow and ice are relatively simple to interpret because the natural background contribution is generally negligible. Murozumi et al (1969) showed that the background (pre- Roman) concentration of lead in remote Greenland snow and ice was only ^lpg g'\ wThich was cleaner than the water found in most analytical chemistry laboratories. Reliable isotope abundance measurements of lead in pre-industrial ice demand that the following be avail¬ able: 1. ultra-clean procedures for the acquisition and stor¬ age of field samples; 2. ultra-clean decontamination procedures; 3. ultra-clean sample storage; 4. ultra-clean sample processing for mass spectrom¬ etry; 5. sensitive high precision mass spectrometry. Murozumi et al (1969) showed incredible skill in mea¬ suring pg g4 concentrations, of lead in Greenland ice. To minimise the contributions from artifact contamination they collected 19 kg or 50 kg size blocks of ice. Thirteen years later, Ng & Patterson (1981) required only 100-200 g samples to confirm the results of the earlier study. The corresponding laboratory blanks for these two studies were ^100 ng and 160 pg, the later amounting to ^30% of the cleanest sample analysed. The size of samples typically used today for isotopic analyses vary from 1-50 g depending on their lead con¬ centration, while the total laboratory processing blank has improved steadily to « 3 pg in 1993 (Rosman et al. 1993) and finally 0.5 pg today (author's laboratory). The improvement in the blank/sample ratio in the 1990s arose from the realisation that snow collected at remote locations is relatively free of impurities and consequently processing can be kept to a minimum, and should ideally be avoided. Whereas earlier studies (Murozumi et al 1969; Ng and Patterson 1981; Boutron & Patterson 1986) using mass spectrometers for lead concentration mea¬ surements attempted to purify the lead using chemical methods such as ion exchange or solvent extraction, cur¬ rent practice is to concentrate the sample by evaporation in an ultraclean environment (Rosman et al. 1993, 1994a,b; Chisholm et al 1995). The capability of current methods can best be illus¬ trated with measurements on the purest laboratory water. Figure 2 shows measurements of lead in three different masses. The concentration of lead in the water is calcu¬ lated from the gradient of the line through the points (0.013 pg gl) , and the procedural blank is determined from the intercept (0.47 pg). In practice, ice samples are normally measured only once or twice, procedural blanks are checked by monitoring critical reagents, and the effectiveness of the beaker cleaning procedures and the quality of the laboratory water are checked by evapo¬ rating equivalent volumes of ultra-pure water in beakers prior to using them for a sample. Lead concentrations are measured by isotope dilution mass spectrometry (1DMS) with a minor variation. A known amount of 205Pb-enriched tracer (half-life 1.5 107) is mixed with the sample prior to analysis. The quantity LABORATORY WATER Figure 2. The concentration of lead in ultra-pure laboratory water. The gradient of the line yields the lead concentration and the intercept gives the analytical blank. of natural lead can be determined from the measurement of the mixed lead mass spectrum. Because 205Pb does not occur naturally, the use of this tracer permits the isotopic composition to be determined in the same measurement. A vital step in the chain of procedures needed for a reliable measurement of lead in ice is the decontamina¬ tion process. The purpose is to mechanically remove contaminated ice, layer by layer, ideally resulting in a central core of uncontaminated ice. Each layer is analysed to confirm the effectiveness of the procedure. (Ng &l Patterson 1981; Candelone et al 1994). Figure 3 shows the concentration profile for a section of the Vostok ice core drilled in Antarctica by a team from the former Soviet Union (Chisholm et al 1995). In this case, however, the contaminant lead penetrated the central core and only an upper limit of concentration could be given. The criterion for an uncontaminated central core is that the concentration must reach a constant value. This approach was developed by Ng & Patterson (1981) and is still the most reliable method of validating the purity of the central sample. However the advent of iso¬ tope abundance measurements added another dimension to DECONTAMINATION PROFILE: VOSTOCK 500 m ICE CORE 200 160 120 o> O) a c o c 0) u c o o n Q. Figure 3. Lead concentration and isotopic ratio profile across a Vostok ice core. The sample was taken from a depth of 500 m (26.2 ka BP) using a thermally drilled hole with kerosene as the retaining fluid. 99 Journal of the Royal Society of Western Australia, 79(1), March 1996 this process by allowing different sources of contamina¬ tion to be resolved. In Figure 3, the outer layers show an initial decrease in isotopic ratio then an increase, but never reach a constant value. This pattern suggests that there are two sources of artifact lead in addition to the natural component. The lead concentration is highest on the outside and also continues to change as the centre of the core is approached, confirming that artifact lead is still present. This analysis can only yield an upper limit of concentration and a lower limit of isotopic ratio. Lead Pollution History of the Northern Hemisphere The Greenland ice cap penetrates the troposphere and rises to an altitude of ~ 3.23 km at Summit (72°35'N, 37°38'W) near the centre of the island. Aerosols are scav¬ enged by falling snow which accumulates on the ice cap and is eventually converted to ice when it reaches a depth of about 80 m. The analysis of snow and ice from various depths provides a record of the pollution history of the Northern Hemisphere. The analysis of pre-Roman ice is essential for a correct interpretation of post-indus¬ trial changes while much older ice provides useful infor¬ mation on early climatic conditions. The oldest ice collected to date (^250 ka) was re¬ trieved recently by the European GRIP (Greenland Ice Core Program) and the United States GISP (Greenland Ice Sheet Program) teams who drilled ^3 km to the base¬ ment at two sites near Summit (Dansgaard et al. 1993; Grootes et al. 1993). Shallower sampling in the same area has also provided interesting complementary informa¬ tion. For instance, an 11 m snow core taken with a poly¬ carbonate hand-operated auger (to minimise contamination) and a 70 m electromechanically drilled core were both taken during the recent European "Eurocore" program (Boutron et al 1991; Candelone et al 1995). Boyle et al. (1995) also report lead concentration measurements to a depth of 6 m at Summit. In the latter study, two series of samples were taken - one from the walls of a snow pit and the other with a solar powered electric coring rig. The analyses of these samples has allowed a lead pollution history of the Northern Hemisphere to be constructed covering one full glacial cycle (Boutron et al. 1991; Candelone et al. 1995; Hong et al. 1994), and has confirmed the earlier measurements by Murozumi et al. (1969). Figure 4 summarises the results of these measure¬ ments and shows the periods of history when lead pollu¬ tion was prominent. Only during the 1990s have reliable lead isotope abun¬ dance data in Greenland and Antarctic snow and ice be¬ come available. When Boutron et al. (1991) analysed an 11 m snow core from Summit they attributed the de¬ creasing concentration of Pb during the 1980s to the rapid reduction in the use of leaded gasoline in the United States, The analyses of lead isotopes in these samples by Rosman et al (1993, 1994a) confirmed this hypothesis. The variation of the 206Pb/207Pb isotope ratio found in these samples is shown in Figure 5 with the signatures of lead used in United States and Eurasian gasoline superimposed. From these data, Rosman et al. (1994a, b) computed the concentrations of US and Eur¬ asian lead in Greenland snow (Figure 6). The results were in excellent agreement with the known pattern of leaded petrol consumption. Year Figure 5. The 206Pb/207Pb isotopic ratios in snow from Summit, central Greenland (Rosman et al. 1993). The USA and Eurasian isotopic signatures taken from the literature are shown (Adapted from Rosman et al. 1994a ). 1000 O) D) Q. c o c CD o c o o 100 10 1 LEAD AT SUMMIT, GREENLAND Petrol Industrial * Revolution Roman/Greek *.,*••* 1000 2000 Year BP 3000 Figure 4. Recent measurements of the lead concentration in Greenland ice extending back to 3000 BP. Events causing sig¬ nificant changes in lead emissions to the atmosphere are identi¬ fied, namely the introduction of leaded petrol, the Industrial Revolution, and the Greek/ Roman civilisations. Data from Boutron et al. (1991), Candelone et al. (1995) and Hong et al. (1995). _0) D> S O c (/) _c -Q Q_ Figure 6. Comparison of USA and Eurasian lead in Summit snow with lead emissions from USA petrol. The concentrations were calculated from the isotopic ratios shown in Figure 5 and the total concentration of lead in the snow (Adapted from Rosman et al. 1994a ). 200 150 100 50 1965 1970 1975 1980 1985 1990 Year 100 Petrol Pb (Thousands tonnes/a) Journal of the Royal Society of Western Australia, 79(1), March 1996 Hong et al (1994) reported relatively high lead con¬ centrations in two thousand year old ice from the GRIP ice core and attributed this to pollution of the atmosphere by industrial emissions caused by the Greek and Roman civilisations. Preliminary measurements on these samples now show a significant decrease in the 206Pb/ 207Pb ratio which correlates with the enhanced lead con¬ centration. These isotopic measurements are highly sig¬ nificant because they provide definitive evidence that this lead had an anthropogenic origin (Rosman et al, 1995). Lead Pollution History of the Southern Hemisphere The Antarctic ice sheet is ideally located to record pol¬ lution in the Southern Hemisphere but the transfer of pollutants to Antarctica from populated regions is less direct than it is for Greenland. Although both land- masses extend to ~ 60° latitude, Greenland is flanked by continents which extend into the Arctic whereas Antarc¬ tica is isolated from the influence of other land masses by the circumpolar convergence, which is a extremely effec¬ tive barrier. Even so, Antarctica is highly polluted with anthropogenic lead although the concentrations in sur¬ face snow are one to twro orders of magnitude lower than in Greenland. Because the concentrations are very low, few reliable lead data for Antarctic snow and ice are available. Using elaborate sample collection and decontamination proce¬ dures, Boutron & Patterson (1987) measured concentra¬ tions as low’ as 2.3 pg g'1 in snow 433 km from the coast near the French base of Dumont d'Urville. This snow fell near the end of 1982 and during 1983. Geochemical arguments were used to show that 80% of this lead was anthropogenic. Recent isotopic abundance measure¬ ments on the same sample gave a 206Pb/2D7Pb ratio of 1.16 compared with 1.07 for petrol and 1.25 for natural dust which indicated that at least 50% of the lead was anthro¬ pogenic (Rosman et al. 1994b). Lead histories for Antarctica are limited to a few reli¬ able studies. Even so, only selected reports will be iden¬ tified here. Recently Wolff & Suttie (1994) produced a time series from the 1920s and observed concentrations of ** 2.5 pg g'1 for 1920-1960 increasing to a peak between 1978-1980 of - 10 pg g 1 then falling back to - 5 pg g 1 by the mid 1980s. Boutron & Patterson (1986, 1987) reported time series for ancient ice which they obtained from a thermally drilled core from Dome C (77°39 S, 124°10E and elevation 3.24 km; samples to 27 Ka BP) and an elec- tromechanically drilled core from Vostok (78°28 S, 106°48’E and elevation 3.49 km; samples to 155 ka BP). Concentrations at Dome C were as low as 0.3 pg g 1 during the Holocene (back to 13 ka BP) but they peaked at 29 pg g'1 during the coldest period of the last ice age (^ 21 ka BP). Corresponding concentrations for the earlier gla¬ cial cycle at Vostok were 2.4 pg g'1 and 19.8 pg g'1 respec¬ tively. Only three measurements of lead isotopes in ancient Antarctic ice have been reported. These include two Dome C ice core samples dated at 7.5 ka BP and 14 ka BP, which gave 206Pb/207Pb ratios of 1.25 and 1.20 respec¬ tively (Rosman et al 1994b; Chisholm et al 1995). The Future of Lead Isotopes and Pollution History Lead isotopes offer a powerful tool for identifying sources of pollution and they are a valuable complement to concentration measurements. Measurements on recent Greenland snow have al¬ ready demonstrated how leaded petrol emissions from the USA and Eurasia can be identified and their relative contributions assessed. Work is currently in progress that will extend the lead isotope record for the Northern Hemisphere back to Roman times and beyond. These data will provide new and interesting information re¬ garding the source of the lead emissions in earlier times. Measurements on deep cores both from Greenland and Antarctica will also provide additional information on the climatic conditions which existed on Earth during the recent glacial cycles. The concentration of the lead in the ice and its isotopic composition will reflect the atmo¬ spheric conditions and the geographical origin of the dust containing the lead. Very few isotope abundance measurements have been made in Antarctica but the technology is now available for the measurements of existing cores to proceed. De¬ tailed measurements at high depth resolution will be needed to identify seasonal influences and to follow the development of industry in the Southern Hemisphere. It is anticipated that high resolution cores (high deposition rates) drilled in the Law Dome by Australian teams will provide samples suitable for such studies. The origin of the dust reaching Antarctica is not known at present. Evidence from isotopic (Grousset et al 1992 ; Rosman et al. 1994) and geochemical (Delmas & Petit 1994) studies are in agreement but conflict with predictions from global circulation models (Joussaume 1993). Isotopic studies on snow and ice coupled with atmospheric monitoring of present day aerosols will help to resolve this problem. The skills required for the collection, decontamination, elemental and isotopic analysis of the cleanest snow and ice samples are not limited to one laboratory. This is an activity which continues to benefit greatly by interna¬ tional scientific collaboration. Acknoivledgements: We arc indebted to John De Laeter who established the Curtin laboratory in the late 1960s and enthusiastically promoted the application of mass spectrometry to environmental science. The Austra¬ lian Research Council has supported Curtin research on lead isotopes in snow and ice cores described in this paper. This later work is the result of a very successful collaboration with Claude Boutron of the Laboratory for Glaciology and Geophysics of the Environment, Grenoble, France. References Aston F W 1942 Mass Spectra and Isotopes. (2nd ed.). Edward Arnold London. Boutron CF & Patterson C C 1986 Lead concentration changes in Antarctic ice during the Wisconsin/Holocene transition. Nature 323:222-225. Boutron C F & Patterson C C 1987 Relative levels of natural and anthropogenic Pb in recent Antarctic snow. Journal of Geophysical Research 92:8454-8464. Boutron C F, Gorlach U, Candelone J-P, Bolshov M A & Delmas R J 1991 Decrease in anthropogenic Pb, Cd and Zn in Greenland snows since the late 1960s. Nature 353: 153-156 101 Journal of the Royal Society of Western Australia, 79(1), March 1996 Boyle E A, Sherrell R M & Bacon M P 1994 Lead variability in the western North Atlantie Ocean and central Greenland ice: Implications for the search for decadal trends in anthropo¬ genic emissions. Geochimica et Cosmochimica Acta 58:3227- 3238. Cameron A E, Smith D H & Walker R L 1969 Mass spectrom¬ etry of nanogram-size samples of lead. Analytical Chemistry 41:525-526. Candelone J-P, Hong S & Boutron C F 1994 An improved method for decontaminating polar snow or ice cores for heavy metal analysis. Analytica Chimica Acta 299:9-16. Candelone J-P, Hong S, Pellone C & Boutron C F 1995 Post- Industrial Revolution changes in large scale atmospheric pol¬ lution of the Northern Hemisphere for heavy metals as docu¬ mented in central Greenland snow and ice. Journal of Geo¬ physical Research 100, 16,605-16,616. Chisholm W, Rosman K J R, Boutron C F, Candelone J P & Hong S 1995 The determination of lead isotopic ratios in Greenland and Antarctic snow and ice with picogram per gram concentrations. Analytica Chimica Acta 311:141-151. Chow T J & Johnstone M S 1965 Lead isotopes in gasoline and aerosols of Los Angeles basin, California. Science 147:502- 503. Chow, T J, Snyder C B & Earl J L 1975 Isotope ratios of Pb as pollutant source indicators. United Nations FAO and Inter¬ national Atomic Energy Association Symposium, Vienna Austria. (IAEA-SM-191/4) Proceedings, pp. 95-108. Curie M M 1913 Sur les ecarts de poids atomiques obtenus avec le plomb provenant de divers mineraux. Comptes Rendus. Academie des Sciences. 1676-1679. Dansgaard W, Johnsen S J, Clausen H B, Dahl-Jensen D, Gundesstrup N S, Hammer C U, Hvidberg C S, Steffensen J P, Sveinbjornsdottir A E, Jouzel J & Bond G 1993 Evidence for general instability of past climate from a 250 -kyr ice core record. Nature, 364:218-220. Delmas R J & Petit J R 1994 Present Antarctic aerosol composi¬ tion: A memory of ice age atmospheric dust. Geophysical Research Letters 21:879-882. Grootes P M, Struiver M, White J W C, Johnsen S & Jouzel J 1993 Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366:552-554. Grousset F E, Biscaye P E, Revel M, Petit R-J, Pye K, Joussaume S & Jousel J 1992 Antarctic (Dome C) ice-core dust at 18 k.y. B.P.: Isotopic constraints on origins. Earth and Planetary Science Letters 111:175-182. Hirao Y, Mabuchi H, Fukuda E, Tanaka H, Imamura T, Todoroki H, Kimura K & Matsumoto E 1986 Lead isotope ratios in Tokyo Bay sediments and their implications in the lead consumption of Japanese industries. Gechemical Jour¬ nal 20:1-15. Hong S, Candelone J P & Boutron C F (1994) Greenland ice evidence of hemispheric scale pollution for lead two millen¬ nia ago by Greek and Roman civilizations. Science 265:1841- 1843. Joussaume S 1993 Paleoclimatic tracers: An investigation using an atmospheric general circulation model under ice age con¬ ditions 1. Desert dust. Journal of Geophysical Research 98: 2767-2805. McMichael A J, Baghurst P A, Wigg N R, Vimpani G V, Robertson E & Roberts R J 1988 Port Pirie cohort study: Environmental exposure to lead and children's abilities at the age of four years. New England Journal of Medicine 319:468-475. Murozumi M, Chow T J & Patterson C C 1969 Chemical con¬ centrations of pollutant aerosols, terrestrial dusts and sea salts in Greenland and Antarctic snow strata, Geochimica et Cosmochimica Acta 33:1247-1294. Mushak P, Davis J M, Crocetti A F & Grant L D 1989 Prenatal and postnatal effects of low-level lead exposure: Integrated summary of a report to the U.S. Congress on childhood lead poisoning. Environmental Research 50:11-36. Needleman H L 1991 Childhood lead poisoning: A disease for the historv texts. American Journal of Public Health 81:685- 687. Ng A & Patterson C C 1981 Natural concentrations of lead in ancient Arctic and Antarctic ice. Geochimica et Cosmochimica Acta 45:2109-2121. Nriagu J O 1983 Lead and Lead Poisoning in Antiquity. Envi¬ ronmental Science and Technology, John Wiley, New York. Nriagu J O 1990 The rise and fall of leaded gasoline. Science of the Total Environment 92:13 - 28. Patterson C C 1956 Age of meteorites and the earth. Geochimica et Cosmochimica Acta 10:230-237. Patterson C C 1965 Contaminated natural environments of man. Archives for Environmental Health 11:344-360. Patterson C C & Settle D M 1976 The reduction of orders of magnitude errors in lead analysis of biological materials and natural waters by controlling the extent and sources of in¬ dustrial lead contamination introduced during sample col¬ lecting, handling and analysis. Proceedings of the Seventh Materials Research Symposium. US Government Printing Of¬ fice, Washington DC. National Bureau of Standards Special Publication 422:321-351. Rabin R 1989 Warnings unheeded: A history of child lead poi¬ soning. American Journal of Public Health 79:1668-1674. Ritsen P I, Esser B K, Niemeyer S & Flegal A R 1994 Lead isotopic determination of historical sources of lead to Lake Erie, North America. Geochimica et Cosmochimica Acta 58:3297-3305. Richards T W & Lembert M E 1914 The atomic weight of lead of radioactive origin. Journal of the American Chemical Soci¬ ety 36:1329-1344. Rosman K J R, Chisholm W, Boutron C F, Candelone J P & Gorlach U 1993 Isotopic evidence for the source of lead in Greenland snow’s since the late 1960s. Nature 362:333-335. Rosman K J R, Chisholm W, Boutron C F, Candelone J P & Hong S 1994a Isotopic evidence to account for changes in the concentration of lead in Greenland snowf between 1960 and 1988. Geochimica et Cosmochim. Acta 58:3265-3269. Rosman K J R, Chisholm W, Boutron C F, Candelone J P & Patterson C C 1994b Anthropogenic lead isotopes in Antarc¬ tica. Geophysical Research Letters 21:2669-2672. Rosman K J R, Chisholm W, Hong S, Boutron C F & Candelone J P 1995 Lead isotope record in ancient Greenland ice. Pro¬ ceedings of 10"1 International Conference on Heavy Metals in the Environment, Hamburg. 1:34-36. Rosman K J R, Loss R D & de Laeter J R 1984 The isotopic composition of tin. International Journal of Mass Spectrom¬ etry and Ion Physics 56:218-291. Shen G T & Boyle E A 1987 Lead in Corals: reconstruction of historical industrial fluxes to the surface ocean. Earth and Planetary Science Letters 82:289-304. Shirahato H, Elias R W & Patterson C C 1980 Chronological variations in concentrations and isotopic compositions of an¬ thropogenic atmospheric lead in sediments of a remote sub- alpine pond. Geochimica et Cosmochimica Acta 44:149-162. Soddy F 1913 Radioactivity. Annual Report of the Chemical Society 105:1402-1408. Tatsumoto M & Patterson C C 1963 Concentrations of common lead in some Atlantic and Mediterranean waters and in snow. Nature 199:350-352. Wolff E W & Suttie E D 1994 Antarctic snow record of southern hemisphere lead pollution. Geophysical Research Letters 21:781-784. 102 Journal of the Royal Society of Western Australia, 79:103-107, 1996 Surface ionization sources and applications J E Delmore, A D Appelhans, J E Olson, T Huett, G S Groenewold, J C Ingram & D A Dahl Idaho National Engineering Laboratory, Box 1625, Idaho Falls, Idaho USA 83415-2208 Abstract A series of new instruments have been developed for the study of ion emission from hot inorganic deposits, these being; a tube ion source, an ion source imaging instrument, and an ion/ neutral mass spectrometer. Techniques have been developed for applying these instruments to the development of new ion emitters as well as in studying the properties of previously developed emitters. This line of study has led to a new type of anion emitter that has found use as the primary gun for the static SIMS analysis of insulators. Applications of the new primary anion gun to the measurement of environmental contaminants are discussed. Introduction Surface ionization (SI) was recognized at least 70 years ago as a scientific phenomenon, and has been employed as an analytical tool for almost 60 years. SI has been investigated in a variety of contexts, but in general this is an area that has been understudied, in spite of the wide spread applications in mass spectrometry for the mea¬ surement of isotope ratios. Of the studies that have been conducted on the fundamentals of SI, most have involved the study of atoms striking a pure, hot metal surface. Few have investigated mechanisms of the sublimation of ions from complex deposits. We have been interested for many years in the chemistry of deposits that directly sub¬ lime (or evaporate in the case of liquid deposits) ions from the surface. This is a complex field, and after many years we are evolving a series of tools which are helping us to better understand the inorganic and physical chem¬ istry which determines which types of emitters, when heated, will efficiently produce ions. This is not a ma¬ ture field, however, and could easily accommodate addi¬ tional investigators. In the early 1960s the Idaho National Engineering Laboratory (INEL) began a program for the measurement of fission yields for the major fissioning species in a vari¬ ety of types of nuclear reactors. This led development of methods for performing isotope ratio measurements on a wide range of fission product elements. The culmination of these studies were the Oklo Natural Fission Reactor studies were performed jointly between Los Alamos Na¬ tional Laboratory, Curtin University and the INEL (Loss et al 1988; Curtis et al 1989; Loss et al 1989). The devel¬ opment of new methods for measuring isotope ratios led to an interest in the chemistry of the deposits on the hot filaments which emitted ions. These activities have in turn been developed into a new line of ion guns for use as the primary ion beam in static secondary ion mass spectrometry (static SIMS) for the measurement of envi¬ ronmental pollutants. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science * Curtin University of Technology, Perth, 1996 If one studies a variety of ion emitters, over time there will be certain systems which produce ions with unex¬ pectedly high intensity ("boomers"). Examples of this type include (but are certainly not limited to); 1) the alkali metal cations from aluminosilicates (typi¬ cally, but not always, zeolites) (Beck et al 1989; Satoh et al 1987; Pargellis & Seidl 1978); 2) borate (B02 ) from rare earth oxides; 3) perrhenate (ReO;) from barium perrhenate in eu¬ ropium and ytterbium oxide matrices (Delmore et al 1995); 4) halides (CL, Br, T) from lanthanum hexaboride (Rachidi et al 1976; Delmore 1982). When one attempts to intentionally produce an emit¬ ter which produces ions as efficiently as those which are produced from contaminants, the results can sometimes be disappointing. An example is iodine from lanthanum hexaboride (LaB^). Iodine present as a contaminant in LaBh is ionized with much higher efficiency than any intentionally added for analysis, and similar issues arise with all of the above examples. These are examples of the importance of the chemical form of the element or molecule to be ionized; ionization efficiency can drop dramatically unless the appropriate chemical species is identified and added in that form. We have chosen to investigate these high intensity "boomers" because they are easier to study, and hence we would be more likely to discover basic features that would lead to fundamental understandings. The two emitters which we have studied the most are alkali metal zeolites and perrhenate /rare earth oxides, and both ap¬ pear to be "scalable." That is, the intensity of the emitted ions can be increased by increasing the size of the de¬ posit, by increasing the loading of the species to be ion¬ ized in the matrix, and by increasing the temperature. This is in contrast to the molten glass ion emitters (silica gel type matrices) which are much more difficult (per¬ haps impossible) to scale (Huett et al 1995). There will be three main themes developed in this paper. First, we will concentrate on the two ion emitters that have found use in this laboratory, alkali metal cat- 103 Journal of the Royal Society of Western Australia, 79(1), March 1996 ions from zeolite, and perrhenate (Re04 ) anions from a europium oxide/barium perrhenate ceramic emitter. Second, we will briefly describe two instruments built specifically for the study of these emitters, and methods developed to study these emitters. Third, we describe how these two ion emitters have been incorporated into ion guns used for SIMS instruments, and show that the Re04 anion gun has proven to have great utility for the SIMS analysis of partial monolayers of adsorbates on in¬ sulating surfaces. Experimental procedures The literature has many references to the emission of alkali metal cations from aluminosilicates, and in prac¬ tice almost any zeolite will upon heating emit large cur¬ rents of alkali metal cations. These zeolites can be soaked in aqueous solutions of an alkali metal chloride or nitrate at room temperature and when the material is heated to 700 to 800°C (Beck et al. 1989) in vacuum, will emit that particular alkali metal cation. These emitters (for all of the alkali metals) are widely used in ion guns for a vari¬ ety of applications, and are even sold commercially by several companies. The perrhenate anion emitter was developed at the INEL and has been used as the active element in several ion guns used for static SIMS. These SIMS instruments have been used for a wide range of analyses, which have opened many new analytical possibilities. The perrhenate anion emitter is a ceramic material, the composition of which is being published elsewhere (Delmore et al. 1995), and operates in the range of 800 to 900°C. It can deliver up to 10 nA into a 2mm spot size, but typically is oper¬ ated at about 100 pA. Hardware called the "tube ion source" has been used in most of our studies. It entails pressing the powdered emitter material into a tube, spot welding this tube to rhenium ribbons, and then spot welding the ribbons to the posts of a filament support (Delmore et al. 1994). The zeolite-based materials were pressed into 304 stainless steel tubes which were pinched off at the back. The perrhenate emitters were pressed into rhenium tubes with a tantalum plug in the back. Two instruments have been developed for testing the properties of these emitters. These have proven to be very useful in helping to understand the ion formation mechanisms. These instruments are an ion source imag¬ ing instrument (Delmore et al. 1994) and an ion/neutral mass spectrometer. The ion source imaging instrument utilizes a tube ion source mounted in an imaging lens, and the emitted ions are projected onto a screen where the image can be recorded. The ion/neutral mass spectrometer is still very early in the development cycle. It employs a source which combines both a surface ionization capability and an elec¬ tron impact capability for ionizing the neutral gases which are desorbed from the emitter material. A com¬ puter system, which is now working for most of the de¬ sired functions, is being developed to control the mass scanning, the data collection, correlation and display, and ion source focusing. The focusing voltages for the two ionization modes are very different, and the com¬ puter program can step between these modes, so that a clear definition can be made between ions originating from the two ionization modes. The value of this instru¬ ment is that it allows both the ion and the neutral com¬ ponents to be individually mass-analyzed so that the relative proportions of ions to neutral components, and the change in the ratio of these components with time, can be measured. This in turn gives insights into ion formation mechanisms, and permits optimization of the emitter being studied. An attempt to quantify the ratios of ions to neutrals has not yet been made. The perrhenate anion emitter has been incorporated into an ion gun, which in turn has proven to be very useful in static SIMS instruments used for the analysis of insulating materials. There are five static SIMS instru¬ ments in our laboratory which use this type of ion gun, two using single quadrupoles (SQSIMS), one using a triple quadrupole (TQSIMS), and two using ion traps (ITSIMS). The SQSIMS have been described elsewhere (Appelhans et al. 1987; Appelhans & Delmore 1989; Appelhans et al. 1990; Appelhans et al. 1994). Results Ion Emitters The ion source imaging instrument has been used quite extensively to image three types of emitters, the two being discussed here (Delmore et al. 1994), and molten glass emitters made using the silica gel method (Huett et al. 1995). Briefly, the ions were shown to come from the face of the deposit, and not from interfacial regions. This might be obvious in hindsight, but before these imaging measurements were made there was the distinct possibil¬ ity that the ions could originate from the metal surfaces that supported and surrounded the deposits. The imag¬ ing studies demonstrated conclusively that the ions were coming from the face of the deposits, and by inference that the chemical and physical properties of the deposit determined the emitter properties of the system. The substrate (the tube which holds the deposit) served the purpose of supporting the deposit, hopefully, but not al¬ ways, inertly. The ion/neutral mass spectrometer has been used to perform a preliminary study of potassium cation emis¬ sion from zeolites pressed into stainless steel tubes. The alkali metal zeolites are known to be ionic conductors, which means that they conduct electricity with cations which are mobile in the lattice, as opposed to conduction with electrons. These ion emitters are stable and intense emitters of potassium cations. At no time could any neu¬ tral species be observed which contained potassium. The only potassium species observable was the atomic cation. This demonstrates that the mobile cations in the solid phase are being desorbed as cations into the gas phase, and that this is a straight-forward example of preformed ions passing from the condensed phase into the gas phase. We have conducted extensive studies with various perrhenate salts blended with various rare earth oxides (Delmore et al. 1995). We have found that the alkaline earth perrhenates are reasonably refractory compounds, and are the only perrhenate salts that give good ion emis¬ sion when blended with a rare earth oxide. Ba perrhenate is the salt of choice. The best ion emitters by 104 Journal of the Royal Society of Western Australia, 79(1), March 1996 far are composed of Ba perrhenate in a Eu203 or Yb,03 matrix. These rare earth oxides are in the +3 oxidation state, but stable +2 oxidation states are available. It is hypothesized that the +3 oxidation states of Eu and Yb can function as mild oxidizing agents since they can be reduced to the +2 state. This in turn helps maintain rhe¬ nium as perrhenate, in which rhenium has the maximum oxidation state of +7. An alternate explanation is that the accessibility of the +2 oxidation state allows for a differ¬ ent crystal structure which might enhance ion emission. The Eu oxide matrix gives an order of magnitude more intensity than rare earth oxides like Nd early in the life cycle of the emitter, largely due to the fact that it can be heated to higher temperatures without depleting perrhenate. At the higher temperatures and emission currents, the lifetime of the Eu based emitter is from one to two orders of magnitude longer. This supports the hypothesis that perrhenate is being lost in the Nd host matrix, while it is preserved in the Eu based emitter. If it were related to factors such as work function, or mobility of perrhenate in the host matrix, factors which might be expected to be related to crystal structure, then it would be expected that the intensity at any given time would be different but not that the lifetimes would be so different. We feel that the evidence supports a model where pre¬ formed ions are sublimed from the surface (Delmore et al. 1995). SIMS Applications It has been recognized since the early days of SIMS that when an anion strikes an insulating surface, that electrostatic charging is much less of a problem than when a cation strikes the same surface (Anderson et al. 1969). This is because there are more secondary elec¬ trons sputtered from the surface than ions, and an elec¬ tron leaving the surface causes the sample to charge posi¬ tive, the same as an arriving positive ion. An arriving negative ion tends to balance the lost secondary elec¬ trons, greatly reducing the charging rate. The limitation in implementing this concept was the lack of anion guns, with the O' gun (Anderson et al. 1969) the only practical one available. The mass of this ion is so light that its use has been limited. The halide guns have the threat of corroding the vacuum system, since they generally em¬ ploy elemental halogen (Rachidi et al. 1976). Hence most primary ion guns for SIMS have employed cations, with Ar, Xe, Cs and Ga most commonly used. When insula¬ tors are to be analyzed, electron flood guns are used to balance surface charging. The balance point between the flux from the cation gun and the flux from the electron flood gun can be a problem with many instruments. This problem is much less severe with TOF-SIMS, however, since the sample can be flooded with electrons when the cation beam is turned off. The perrhenate anion beam provides great versatility as a primary SIMS gun for the analysis of insulators, especially when combined with the technique of pulsed extraction (Appelhans et al. 1990). Also, the heavy mass (250 daltons) provides excellent secondary ion yields, al¬ though one-on-one comparisons of secondary ion yields are just now being conducted, and the results of the com¬ parison are not yet published. The extraction voltage applied to the sample, which causes secondary ions to be focused into the mass spectrometer, is reversed about 20 time per second, alternately focusing positive and then negative secondary ions into the mass spectrometer. The duty cycle between positive and negative ion extraction is adjusted to find a charge neutrality point, with ion collection times tied to this duty cycle. This has proven to be a highly efficient method for maintaining charge neutrality on insulating samples. We have built five static SIMS instruments which em¬ ploy the perrhenate anion source used with pulsed ex¬ traction, and have had remarkable success with the analysis of insulators (Delmore & Appelhans 1991; Groenewold et al. 1995a, b,c,d, 1996; Ingram et al 1995). The large perrhenate molecular anion has proven to be particularly effective at lifting adsorbate molecules from surfaces. We have analyzed molecules such as the pesti¬ cide malathion (Delmore & Appelhans 1991) directly on the surfaces of plant leaves, the extractant tri-n-butyl phosphate adsorbed on soil and rock (Groenewold et al. 1995a,b), the indoor air contaminant cyclohexyl amine on dust particles and miscellaneous laboratory objects (Groenewold et al. 1996), the degradation products of chemical warfare agents on a variety of surfaces (Ingram et al. 1995; Groenewold et al. 1995c), and the detection of inorganic complexes on carbon (Groenewold et al. 1995d). These analyses are but a few of the many types which these instruments have been able to accomplish. The analysis of 'salt cake' is one example of the effi¬ cacy of the perrhenate primary ion for producing rich inorganic species information. Salt cake is a material which has formed in radioactive waste storage tanks, and there exists an important need for characterization. A synthetic salt cake was prepared minus radio nuclides to simulate actual material, to evaluate different character¬ ization techniques (Bajic et al. 1995). The synthetic sample consisted primarily of sodium nitrate, with a sig¬ nificant quantity of nitrite and sodium nickel ferrocya- nide. The composition of the sample is described in terms of the relative number of moles added (Table 1). Table 1 Relative molar composition of synthetic salt cake, normalized to NaNQ3. Comonent mole fraction NaN03 1.000 NaN03 0.100 NaN03S04H20 0.010 Na2NiFe(CN)6 0.220 The anion spectrum was particularly informative re¬ garding inorganic species present in the sample. The most abundant ion in the anion spectrum was CN , and relatively abundant Na, Fe and Ni cyanide adduct ions were also readily observable at m/z 75', 108', 110', 112', 134', 136', and 138' (Figure 1; note that the Ni adducts have two major isotopes). Fe and Ni both have two cya¬ nide adduct ions, one each for the +2 oxidation states, and one each for the +3 oxidation states. In addition to the cyanide bearing ions, nitrate/nitrite bearing species were observed at m/z 115', 131" and 147'. These ions cor¬ respond to Na(N02)2', Na(N02)(N 03)', and Na(N03)2'. What is significant about the observation of these ions is that they were either low abundance, or could not be 105 Journal of the Royal Society of Western Australia, 79(1), March 1996 30.00% 25.00% -- S 20.00% c ra ~a c | 15.00% + G) > ■3 10.00% 5.00% 0.00% J 26 CN' 1 00% Fe(CN)2- P03‘ NCO' N03 Ni(CN)2 r Fe(CN)3’ Ni(CN)3‘ N02 ; Na(CN)2% ■ ’ 1 Figure 1. Anion SIMS spectrum from salt cake (described in Table 1). Note the prevalence of complex molecular species in relation to the more simple species. observed when the sample was analyzed using a SIMS spectrometer equipped with an atomic primary ion gun (Ga+). Conclusions We have studied the chemical and physical character¬ istics of ion emitting deposits, and have developed unique instrumentation and methods for investigating the nature of these emitters. We feel that the real high intensity "boomers," which give very high ion yields, are probably systems which form ions in the condensed phase prior to emission. While we have gained new insights into these emitters, much remains to be learned. For example, probably many other high intensity ion sources could be developed but have not yet been dis¬ covered. There is also the question as to whether europic oxide is truly an oxidizing matrix, or if the ability of this matrix to enhance perrhenate anion emission is due to some other factor. Many such research topics can be envisioned, and they only await study. The original intent of these studies was to gain new insights into ion emitters so as to allow better isotope ratio analyses for nuclear and geologic studies. This has remained an interest, but in the process emitter systems have been discovered which emit heavy and unusual ions (in particular the perrhenate anion), that have proven to be valuable for the static SIMS analysis of heavily insulating surfaces. The perrhenate anion has proven to be particularly valuable for the analysis of par¬ tial monolayers of adsorbed contaminants on insulating surfaces. This work also demonstrates the extent to which a scientific project can lead to new lines of study, some of which can have practical applications that could never have been predicted. The fission yield studies of 30 years ago have led us through a sequence of programs and studies which today give us the capability to measure environmental pollutants. The common thread that al¬ lowed this to occur has been a better understanding of ion emission from hot surfaces. References Anderson C A, Roden H J & Robinson C F 1969 Negative ion bombardment of insulators to alleviate surface charge up. Journal of Applied Physics 40:3419-3420. Appelhans A D & Delmore J E 1989 Comparison of polyatomic and atomic primary beams for secondary ion mass spectrom¬ etry of organics. Analytical Chemistry 61:1087-1093. Appelhans A D, Delmore J E & Dahl D A 1987 Focussed, rasterable, high-energy neutral molecular beam probe for secondary ion mass spectrometry. Analytical Chemistry 59:1685-1691. Appelhans A D, Dahl D A & Delmore J E 1990 Neutralization of sample charging in secondary ion mass spectrometry via a pulsed extraction field. Analytical Chemistry 62:1679-1686. Appelhans A D, Dahl D A, Delmore J E, Groenewold G S, & Ingram ] C 1994 Book of Abstracts, PITTCON 94; Chicago, Illinois. Abstract 1229. Bajic S J, Luo K S, Jones R W & McClelland J F 1995 Analysis of underground storage tank waste simulants by fourier trans¬ form infrared photoacoustic spectroscopy. Applied Spectros¬ copy 49:1000-1005. Beck S T, Warner D W, Garland B A & Wells F W 1989 A simple alkali-metal ion gun. Review of Scientific Instruments 60:2653-2656. Curtis D B, Benjamin T M, Gancarz A J, Loss R D, Rosman K J R, DeLaeter J R, Delmore J E & Maeck W J 1989 Fission product retention in the Oklo natural fission reactors. Ap¬ plied Geochemistry 4:49-62. Delmore J E 1982 Isotopic analysis of iodine using negative surface ionization. International Journal of Mass Spectrom¬ etry Ion Processes 43:273-281. Delmore J E & Appelhans A D 1991 Detection of agricultural chemical residues on plant leaf surfaces with secondary ion mass spectrometry. Biological of Mass Spectrometry 20:237- 246. 106 Journal of the Royal Society of Western Australia, 79(1), March 1996 Delmore J E, Appelhans A D & Olson J E 1994 Self imaging of surface ionization ion sources - where do the ions come from? International Journal of Mass Spectrometry and Ion Processes 140:111-122. Delmore J E, Appelhans A D & Peterson E S 1995 A rare earth oxide matrix for emitting perrhenate anions. International Journal of Mass Spectrometry and Ion Processes 146:15-20. Groenewold G S, Ingram J C, Delmore J E & Appelhans A D 1995a Static secondary ionization mass spectrometry analy¬ sis of tributyl phosphate on mineral surfaces: effect of Fe(II)- Journal of the American Society for Mass Spectrometry 6:165- 174. Groenewold G S, Ingram J C, Delmore J E, Appelhans A D & Dahl D A 1995b Rapid Detection of tri-n-butyl phosphate on environmental surfaces using static SIMS. Journal of Hazard¬ ous Materials 41:359-370. Groenewold G S, Ingram J C, Appelhans A D, Delmore J E & Dahl D A 1995c Detection of 2-chloroethyl ethyl sulfide end sulfonium ion degradation products on environmental sur¬ faces doing static SIMS. Environmental. Science and Tech¬ nology 29:2107-2111. Groenewold G S, Ingram J C, Appelhans A D, Delmore J E & Pesic B 1995d Static SIMS detection of gold and gold cynaide complexes on carbon using crown ether enhancement. Ana¬ lytical Chemistry 67:1987-1991. Groenewold G S, Ingram J C, Gianotto A K, Appelhans A D & Delmore J E 1996 Static secondary ionization mass spec¬ trometry detection of cyclohexaamine on soil surfaces ex¬ posed to laboratory air. Journal of the American Society for Mass Spectrometry 7:168-172. Huett T, Ingram J C & Delmore JE 1995 Ion-emitting molten glasses-silica gel revisited. International Journal of Mass Spectrometry and Ion Processes. 146:5-14. Ingram J C, Groenewold G S, Appelhans A D, Delmore J E & Dahl D A 1995 Static SIMS detection of gold and gold cya¬ nide complexes on carbon using crown ether enhancement. Analytical Chemistry 67:187-195. Loss R D, DeLaeter J R, Rosman K J R, Benjamin T M, Curtis D B, Gancarz A J, Delmore J E & Maeck W J 1988 The Oklo natural reactors; cumulative fission yields and nuclear char¬ acteristics of Reactor Zone 9. Earth and Planetary Science Letters 89:193-206. Loss R D, Rosman K J R, DeLaeter J R, Curtis D B, Benjamin T M, Gancarz A J, Maeck W J & Delmore J E 1989 Fission- product retentivity in peripheral rocks at the Oklo natural fission reactors, Gabon. Chemical Geology 76:71-84. Pargellis A N & Seidl M 1978 Thermionic emission of alkali ions from zeolites. Journal of Applied Physics 49:4933-4938. Rachidi I, Monte J, Pelletier J, Pompt C & Rinchet F 1976 Surface ionization negative ion source. Applied Physics Letters 28:292-294. Satoh Y, Takebe M & Iinuma K 1987 Emission characteristics of zeolite a-ion source. Review of Scientific Instruments 58:138- 140. 107 Journal of the Royal Society of Western Australia, 79:109-117, 1996 SHRIMP: Origins, impact and continuing evolution W Compston Research School of Earth Sciences, Australian National University, Canberra ACT 0200 Abstract The ion microprobe SHRIMP was engendered by the need for in situ micro isotopic analysis of minerals relevant to research in isotope geology at RSES. It was home-built and made large to avoid the loss of transmission that accompanies small ion probes of traditional design. The presence of high quality mechanical and electronic workshops at the RSES, the experience of S W J Clement in ion optical design, the advent of new mass-analyser designs by H Matsuda, and the funding regime then enjoyed at ANU that allowed long-term and high risk projects, were each an essential component in the decision to proceed. Some details are given of the history and construction of SHRIMP I, a description of an alternative model now under construction termed SHRIMP-RG, and an account of the initial applications to isotope geoscience that have been more widely developed since. Comparison is made between the performance of SHRIMP with other methods for in situ isotopic analysis now available or under development. Introduction This is an account of one particular analytical instru¬ ment, the Sensitive High Resolution Jon MicroProbe (SHRIMP), rather than of the development of ion micro¬ probes in general. Such a restriction is appropriate to the present symposium because the first commercial SHRIMP was manufactured for the Perth Consortium, and John de Laeter was the driving force behind it. I wish to emphasize that the development of SHRIMP from its conception to realization has been very much a team effort. Finally, history has its own uncertainties, equivalent to experimental error in analytical sciences, the main one being the recollection of exactly what did happen rather than how it was rationalized later. Development of SHRIMP I The need for in situ micro-isotopic analysis Since about 1960, use of the electron microprobe with its in situ capability for elemental analysis has trans¬ formed our understanding of the geochemical relation¬ ships between coexisting minerals. However, its perfor¬ mance for trace elements is limited by the high x-ray background continuum, and for the light elements by the very low energy of their characteristic x-radiation. In addition, the electron microprobe cannot measure the isotopic compositions of elements such as Pb, so that it cannot give direct in situ age determinations. In the late 1960s, physical-chemist C Andersen of the Applied Research Laboratories (ARE) used a small ion microprobe designed by H Liebl to develop procedures for successful in situ determination of concentrations for many elements. Andersen & Hinthorne (1972) used the same instrument to measure the age of fine-grained U- © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 rich minerals in lunar rocks. At that time, their results were disparaged by some isotope geochemists, but the possible use of in situ instrumental analysis was a very appealing prospect for others struggling to develop clean laboratories and to miniaturise their chemical separation procedures for analysis of the very small amounts of re¬ turned lunar samples. No chemistry or mineral separa¬ tion would be required, and the ion probe itself would pre-clean the area selected for analysis. The idea of ob¬ taining an ion microprobe at the Australian National University for the purposes of geochemistry and within- grain age determinations thus arose directly from Andersen's demonstrations. To buy or to build? Mainly for its promise of sensitive and rapid elemen¬ tal analysis, S R Taylor proposed in early 1971 that the ANU should purchase an ARL instrument. The Bureau of Mineral Resources in Canberra indicated its willing¬ ness to help with the purchase, but the proposal was abandoned shortly afterwards due to a prohibitive rise in the cost. This might be viewed in retrospect as a fortu¬ nate escape, as the extent of secondary molecular frag¬ ments that would interfere with the required elemental ions was not realized until later. Discussion began later that year between S W J Clement and myself on the pos¬ sibility of building our own high resolution instrument. The types and abundances of secondary molecular fragments produced by sputtering depend, among other factors, on the chemistry of the target. Andersen was able to identify the main such interferences for his lunar U-rich minerals, and to correct for them by subtracting a calculable fraction of the ion beams at other masses where the particular interference dominated (the 'peak¬ stripping' procedure). The procedure worked well with the old and U-rich lunar minerals for which the radio¬ genic Pb isotopes constituted a large fraction of each Pb mass. However, when the interferences dominate the latter, as for example for young and/or U-poor zircon, a 109 Journal of the Royal Society of Western Australia, 79(1), March 1996 serious magnification of experimental errors must result. In addition, how could we know that all possible isobars from any given mineral had been identified? When these complications had been fully realised, our attention turned to the possibility of obtaining an ion microprobe that could operate routinely at high mass-resolution, at least 5000 in the case of zircon, enough to separate nearly all the unwanted isobars. In March 1973, in response to continued discussion, Anton Hales (first Director of the newly-established Re¬ search School of Earth Science) requested the advice of a formally-constituted RSES Ion Probe Committee compris¬ ing Taylor, S J B Reed and myself. One alternative was to buy one of the first commercially-available high reso¬ lution ion probes, which had been developed by Associ¬ ated Electronic Industries by combining the highly-suc- cessful spark-source mass spectrograph, the MS7, with the primary probe-forming column developed by I W Drummond and J W P Long of Cambridge University. One of these instruments already had been ordered by the University of Chicago (J V Smith) and the University of Cambridge (J W P Long) for geological applications. The other alternative was to design and build our own ion probe in-house, including the employment of S W J Clement for the ion optical and engineering design as an essential ingredient. Clement and I had become concerned that the sensi¬ tivity of the AEI and other small commercial machines then under development (Micromass, Cameca) would be too low at high resolution for effective use in U-Pb age determination. There was no doubt that the MS7 could achieve 5000 R, but we realized that like all other small mass-analyzers, it did so by the use of very narrow slits both at the source and the collector. The narrow source slit reduced the transmitted secondary ion beam and hence reduced the sensitivity for trace-element detection. The narrow' collector slit removed the 'flat-top' from the mass spectral peaks that experience with thermal ioniza¬ tion mass spectrometry showed to be essential for accu¬ rate quantitative measurements. There was a fundamen¬ tal trade-off here, sensitivity for high resolution. The field of beam transport theory, which Clement had used in his PhD project, was obviously relevant for an optimal design. It was also clear that the use of a wide source slit for higher sensitivity at high mass-reso¬ lution would demand a much larger magnet than any available from mass spectrometer manufacturers. The latter (very sensibly) wished to make their ion probes as an assembly of mass-spectrometer components that had been developed already (and debugged) for other pur¬ poses, typified by the adaptation of the MS7 spark-source mass spectrograph. Manufacturers also aimed for small instruments to lower the price and reduce floor space. We felt that a custom-designed ion probe had to be supe¬ rior. Design, Personnel and Construction Following support by Faculty Board, The RSES Direc¬ tor (A Hales) approved the proposal to build the ion probe in-house in late 1973. The decision was influenced by the favourable infrastructure for instrument construc¬ tion that was the School's heritage from its former pres¬ ence within the Research School of Physical Sciences. RSES had well-equipped and expert machine and elec¬ tronics shops, accustomed to the design and manufacture of new apparatus and much larger than found in con¬ temporary geology departments elsewhere. The high- risk nature of the project was recognized from the outset. We hoped to progress with capital costs spread over a five-year period. The costs of establishment salaries were not included. The initial working group comprised S Clement, F Burden (mechanical), N Schram (electronics), D Millar (technical officer) and myself, with Professor G Newstead from the Department of Engineering Physics joining in early 1974 for magnet design and D Kerr of the BMR for computer-control in 1975. This group met each month from 1975 to ensure a full exchange of information, with reports to Faculty Board. The Matsuda mass-analyser. Clement arrived in early 1974 to make the detailed design. The key part of the instrument was the mass-analyzer, which needed to give simultaneous high resolution and high sensitivity. Clem¬ ent had completed an original analysis for a double-fo¬ cussing mass analyzer by mid-1974, but then H Matsuda of Osaka University published his most relevant paper on sector mass-analyzer design (Matsuda 1974). We needed a design that minimized the complex image-ab¬ errations inescapably present in ion optical systems, as well as one that seemed to lie within our practical capa¬ bilities such as employing a cylindrical rather than toroi¬ dal electrostatic analyzer (ESA). Matsuda (1974) ad¬ dressed exactly those conditions. He considered the ion optics of sector mass analyzers as an integrated whole, as distinct from aberration-free individual components. Just as the energy-dispersion of the ESA, which as energy- spread represents a first-order aberration of the image, is adjusted to be equal but opposite to that in the magnet to obtain a double-focussing system, Matsuda arranged the various ion optical components so that their second-or¬ der aberration coefficients also cancelled each other. An essential part of this was the introduction of an electro¬ static quadrupole lens (the 'Matsuda' lens) between the ESA and magnet. There was also curvature in the z- direction of the entrance edges to the ESA plates. For the particular Matsuda design that we favoured, the theo¬ retical image-aberrations would sum to ^ 6 microns for the magnet turning radius of 1 metre (which we consid¬ ered to be the largest that we could manage) and for angular divergences and energy-window of ± 0.005. Clement confirmed Matsuda's calculations using an in¬ dependent computer program (from C Stevens of the Argonne Laboratory at the University of Chicago) based on phase-space transformations. We concluded that there could be no better design for our purpose, and adopted it henceforth. Tribulations of fabrication. The 6 tonne magnet was powered without pole-pieces in late 1976. Each pole- piece was intended to be an assembly of five insulated segments of silicon steel but this material was abandoned after very slow delivery followed by external handling mistakes. They were made instead of soft-iron segments machined and annealed in-house with great care. In 1979, segments laminated from transformer shim were tried to permit faster field switching. This indeed was the case but the refocussed image at the collector ap¬ peared to be variable in quality, which may or may not 110 Journal of the Royal Society of Western Australia, 79(1), March 1996 have been due to the laminations (see below), and we returned to the solid segments. Practice can never be quite as good as theory. Matsuda (1974) envisioned perfectly uniform magnetic and electric fields and perfect geometry, but it is hard to achieve perfection with real materials and real engineer¬ ing. A further example concerned with the magnet will illustrate this. No magnetic field terminates exactly at the edges of the pole-pieces; rather, it falls off sharply over a finite distance beyond them so that it is necessary to determine the position of an effective field boundary. In practice, the magnet for small sector instruments is traversed on rails at right angles to the line joining the source and collector slits until the narrowest image is found empirically. The shape and weight of our ion probe magnet did not allow it to be moved, so instead the collector had to be traversed along the beam-line to find the exact image point. However, the image quality was found to depend strongly on the particular mass collected, and also on the preceding values for the mag¬ netic field. These observations produced the lowest points of mo¬ rale in the project. What was wrong with the magnet? No such effects wrere previously described, although we had noticed with thermal ionization instruments that Sr and Pb ions needed slightly different positions of the magnet for the best refocussing. The discovery that the image point moved in reproducibly with decreasing field when the magnetic field was kept in a fixed cycle of steps saved the situation. The collector could then be driven under computer control to a preset optimum posi¬ tion for each required mass. This remains our solution to the problem. We learned later that at least one manufac¬ turer corrected the effect when making wide-range scans at high mass-resolution by an in-built computer-driven einzel lens. This was not publicised as if the need for such a correction was an admission of deficiency. We have since found that the size of the effect (several mm per 50 dal tons) can vary with different SHRIMP magnet designs and may even be reversed. The effect might be due to very slight change in the pole-gap width as the field is changed, or to slight changes in field distribution due to changes in flux-density in the yoke, or both. The fabrication of the ESA vacuum housing and 1.3 m radius electrodes was subcontracted to the Bendigo Ord¬ nance Factory in late 1975, and delivered in early 1977. J Coles was appointed to assist in final assembly and test¬ ing of the ion probe in late 1977, and to develop its appli¬ cation to light element stable isotope ratios. The first detailed report on the testing of the ion probe mass- analyser using a thermal ionization source appeared in the Annual Report of the Research School of Earth Sciences (Clement et al. 1978). The first secondary ions generated by sputtering were produced a year later (Clement et al. 1979). First geological applications S and Pb isotopes in sulfides. Using an Arf primary beam, the first geological tests of the instrument were for Pb+ isotopic ratios from Broken Hill galena by lon-count- ing (Coles et al 1980). The beam was very stable and the results precise. Tests were extended to the isotopically variable Mississippi Valley sulfides and to the use o large PbS~ and PbS, beams measured via the Faraday cup, with peak-stripping of the sulfur isotopes. Signifi¬ cant variations both in Pb and S isotopic ratios were found, but the work was never submitted to any refereed scientific journals. S isotopic ratios in galena, measured both as S* by ion-counting and as S by the Faraday cup, were also demonstrated. Ti isotopic composition of terrestrial minerals. With the discovery of small nucleogemc ^Ti isotopic anoma¬ lies in Ca-Al-rich chondrites using thermal ionization mass-spectrometry, one of the first ion probe applications to be examined was the accurate and precise measurement of Ti isotopes in several terrestrial minerals (Compston et al 1981). The crucial need for accurately-known and constant values for the dead-time of the detector and counting system was encountered first-hand, and also the pres¬ ence of systematic differences in mass-discrimination in the sputtering of different mineral species. U-Pb dating. The potential of within-grain analysis for determining the crystallization history of zircons had been appreciated since the early 1970s, following discov¬ eries using conventional dating techniques that old grains had preserved their ages through Alpine-age granulite-facies metamorphism, and that xenocrysts in some granites had survived fusion temperatures. I S Williams, who had analysed zircons and other U-rich minerals by the conventional method, was appointed at RSES in 1981 to participate in the hoped-for use of SHRIMP for this purpose. We sought to find an analyti¬ cal method that would allow us to generate realistic val¬ ues for the radiogenic 207Pb/235U and 206Pb/238U per analysed spot, that could be plotted on the Concordia diagram. The particular problem for zircon was the need to de¬ termine 206 Pb /mU reliably, as it was quickly established that its 207Pb/20ftPb as measured using sputtered second¬ ary ions agreed to within ion counting statistics with that measured by thermal ionization. In great contrast, the observed PbVU* from sputtering and mass-analysis was about three times larger than the 20*Pb/238U in the target for zircons of known age, with repeat analyses from grains having the same age varying over a 10% range. The previous work by Andersen & Hinthorne (1973) on transforming their measured ion probe data into chemi¬ cal concentrations, combined with our practice from ther¬ mal ionization of normalizing to a fixed ratio to correct for variable discrimination, led us to an empirical but successful procedure. Reasonable linear correlations were found between the observed Pb+/U+ and the ob¬ served UO/tb in zircons that have the identical age (Williams et al 1981). This allowed us to correct Pb+/U+ as measured for variable discrimination by normalizing to a fixed value for U07U% then to use the corrected Pb+ /U* of a concurrently measured zircon of known age as a comparison-standard. With continued development, the procedure now allows us to measure the 206Pb/238U ages of very young zircons with sufficient accuracy and precision both to compete with the conventional method and to detect inherited zircons as whole grains and in¬ clusions that are not otherwise detected (Compston & Williams 1992). Impact of Shrimp in Isotope Geoscience There is no doubt that zircon U-Th-Pb measurement has been the dominant use of SHRIMP so far. Of a minimum 111 Journal of the Royal Society of Western Australia, 79(1), March 1996 of ca. 150 articles that have appeared in refereed research journals to the end of 1995, only ca. 15 % deal with iso¬ tope ratios of the 'light' elements (S, Ti, Mg). The remain¬ der have used zircon geochronology for geological re¬ connaissance or to test various geological hypotheses. However, the 15 % light element applications demon¬ strate that SHRIMP can be used over the whole nuclidic range, and the dominance of zircon studies simply re¬ flects our current perceptions of what has been the most rewarding scientifically. This balance may change in the future. We regard the following projects as originating a number of important types of applications of SHRIMP to the earth sciences that have been used widely and devel¬ oped further. Accurate age measurements on terrestrial zircons. Antarctic zircons, previously dated using the original (small) ARL ion probe and its peak-stripping methods for isobaric interferences, were reanalysed by SHRIMP. The ages found were wholly consistent with geological constraints, and previously contentious results were shown to be an artefact of uncorrected isobaric interfer¬ ences in the ARL data (Williams et al 1982; Williams et al 1983) . Ages of zircon overgrowths in polymetamorphic epi¬ sodes. The main reason for the development of SHRIMP was in situ microanalysis, with the prospect of detecting age-differences between different growth-zones within a single mineral. This capability was first illustrated for zircons from the Archaean Napier Complex in Antarctica and in zircon xenocrysts from Palaeozoic granites in Eastern Australia (Williams et al. 1981, 1982). Unsup¬ ported radiogenic Pb within a single zircon was observed in high-grade gneiss at Mt Sones (Williams et al. 1983, 1984) , and the growth at high metamorphic grade of characteristically low-Th/U zircon rims around Protero¬ zoic grains was documented first in the Scandinavian Caledonides (Williams & Claesson, 1987). In situ age measurements on lunar zircons and ini¬ tial Pb in lunar feldspars. Ion probe geochronology started and continued for 10 years on lunar zircons that have been discovered by diligent search of many lunar thin-sections by Lunar Sample Curator Charles Meyer (Compston et al. 1983). Serial lunar magmatism since 4.35 Ga has been established. Meteoritic zircons have been dated also (Ireland & Wlotzka 1992). More recently, the high sensitivity and within-grain targeting capability of SHRIMP showed that the extremely radiogenic com¬ position of Ba-rich feldspars in lunar felsites is an origi¬ nal magmatic characteristic rather than a product of the 'terminal lunar cataclysm' (Compston et al. 1988; Compston et al. 1991). Oldest-known terrestrial zircons. This discovery arose from our use of the age-spectra of detrital zircons in Ar¬ chaean sediments to explore the ages and nature of their source rocks (Froude et al. 1983). It generated studies of detrital zircon ages in other Archaean metasediments, and of trace-elements, especially REE, in inclusions within the old zircons (Maas et al. 1992) that bear on the nature of the primary zircon sources. The dating of young zircons. Palaeozoic zircons were at first considered to be too young for ion probe dating because the amount of 207Pb is usually too low to obtain useful 207Pb/206Pb ages. Consequently, the dating of young grains by ion probe must rest mainly on their 206Pb/23HU ages, which in turn are limited by the ability to control the variable discrimination of 206Pb relative to 238U during sputtering and secondary ion extraction. Mainly for this reason, we obtained and analysed zircons from the collection made by R Ross and others from bentonites in the British Ordovician stratotvpes (Compston et al. 1982). The project has developed into continuing appli¬ cations of SHRIMP to the definition of the geological time-scale (Compston & Williams 1992; Compston et al. 1992), with the realization that some of the zircon samples used for conventional time-scale dating were probably composite in age due to inheritance, a point hotly disputed and now in the process of resolution. The capability of SHRIMP to determine extremely young zir¬ con ages was first illustrated by analyses of 2 Ma-old zircons from the western Himalaya (Zeitler et al. 1989). Terrane mapping in Archaean gneisses. The high block¬ ing-temperature of zircon means that intense deforma¬ tion and repeated metamorphism does not erase earlier zircon ages in gneisses that contain multiple age compo¬ nents, which in many instances allows their detailed his¬ tory to be unravelled in a single sample. This was evi¬ dent for the Napier Complex (Williams et al. 1982), and was used to decipher the history of the Narryer Gneiss terrane (Kinny et al. 1990; Nutman et al. 1991) during the search for intact 4.2 Ga rocks. Our study of the Narryer region has shown clearly that correlations between iso¬ lated exposures of gneiss cannot be made reliably with¬ out zircon geochronology. Currently, the oldest terres¬ trial rock known is the 4.0 Ga Acasta gneiss in Canada, first indicated as very old by thermal ionization mass- spectrometry then determined accurately as 4.0 Ga by SHRIMP (Bowring et al 1989). SHRIMP dating of other U-bearing minerals. The first lunar paper (Compston et al. 1984) also included the es¬ sential analytical and data-reduction procedures for zircon ion-probe work that, with various refinements, we con¬ tinue to use today. The same correction method for variable discrimination in zircon analysis has been applied suc¬ cessfully to a number of other U-bearing minerals; perovskite (Compston et al. 1985), and baddeleyite, monazite, titanite and rutile (Camacho et al. 1993). In situ Hf isotopic analysis. The potential for hafnium as a radiogenic isotope fingerprint is well-known from thermal ionization mass-spectrometry, but Hf is a diffi¬ cult element to separate chemically and to mass-analyze with the necessary precision. Because of the high Hf contents of zircons, an early reconnaissance study of the potential of SHRIMP for Hf isotopic analysis was made, with encouraging results (Compston et al. 1982; Kinny et al. 1991). Mathematical procedures for peak-stripping were developed to correct for hydride, oxide, Lu and Yb isobars, which cannot be mass-resolved by SHRIMP II without losing sensitivity. Further analyses of Hf have been deferred pending the development of the SHRIMP- RG and/or the availability of the multicollector for SHRIMP II. Isotopic mapping of trace-element distribution within minerals. With the advent of the computer-controlled sample-stage in SHRIMP II, it has been possible to construct 112 Journal of the Royal Society of Western Australia, 79(1), March 1996 detailed maps of the spatial distribution of any nuclidic or molecular secondary ion at high mass-resolution (Clem¬ ent et al. 1992). This facility has been combined with imag¬ ing by back-scattered electrons and cathodoluminescence us¬ ing the electron microscope to obtain very powerful targetting and interpretive information (Williams et al. 1995). Depth-profiling using SHRIMP. The lens configuration of the primary-beam focusing column was designed originally for two-stage demagnification of the duoplasmatron extraction-aperture to produce the ca. 20 pm probe itself. However, the probe was plagued by a low-intensitv but wide halo that caused a slow release of surface-related common Pb during analysis. Initially, as an experiment, it was modified to implement 'Kohler il¬ lumination' (Compston et al. 1983). This was successful; the probe now had a sharply-defined edge plus a highly- uniform ion-density within the sputtered area. It was then realized that 'Kohler illumination' had an advan¬ tage over spot-rastering for depth-profiling; it gave 'par¬ allel' information over the full width of the spot rather than 'serial' as in rastering, which means a much higher sensitivity per second. In addition, it removed the need for the very small probe diameter required for depth¬ profiling by rastering. Depth-profiling in this way was first employed in ex¬ periments on the diffusion of radiogenic Pb from the (high-U) SL3 zircon standard (Williams et al. 1988). Dif¬ fusion was induced, holding the samples at constant high temperature under hydrothermal conditions, and profiles for the escape of Pb, U and Th were determined. Results for diffusion coefficients were not then reported because it was realized that the process was not pure volume- diffusion. It was found that annealing of the samples prior to the experiments failed to remove zones of struc¬ tural breakdown due to radiation damage. The work has been repeated recently using the low-U zircon SL13 and with improved control and documentation (Lee et al. 1995). Ti and Mg isotopic anomalies in meteoritic minerals. Small isotopic anomalies in ^Ti observed using thermal ionization in bulk refractory inclusions in carbonaceous chondrites were transformed into huge anomalies, from 2.0 % to -4.0 %, when the carrier mineral hibonite was selectively probed (Ireland et al. 1985). Mg isotopic analysis of the same hibonites revealed the presence of excess 26Mg, attributable to in situ decay of 26A1 in some grains, ranging up to more than 1000 % in one instance (Ireland & Compston, 1987). In situ S isotope analysis as an isotopic fingerprint. De¬ tailed SHRIMP analyses of different S-bearing mineral species confirmed that isotopic discrimination during sputtering is controlled mainly by the crystal bond- strengths, and that accurate differences in S isotopic com¬ position can be measured relative to standards of the same mineral-type (Eldridge et al. 1985; Eldridge et al. 1987). The established techniques were first tested by applications to fine-grained sulfide ores from Mt Isa (Australia) and the classical ore deposits at Rammelsberg (W Germany), demonstrating that different generations of sulfides of the same species may be precipitated and co-exist without any isotopic equilibration between younger and older despite superposed metamorphism. Future Developments The continued development of SHRIMP is subject to the possibility that it might be made redundant over¬ night by the arrival of some new instrumental technique that is both better and much cheaper. I am aware of two micro-techniques that have been predicted as substitutes for SHRIMP. These are accelerator mass spectrometry (AMS) and laser-sampling, inductively-coupled plasma ionization mass-spectrometry (ICPMS). Accelerator mass spectrometry AMS for trace-element and isotope ratios is currently being developed in Australia at CSIRO, North Ryde, un¬ der the project name AUSTRALIS (Sie & Suter 1994). It involves the use of sputtering to take a continuous mi¬ cro-sample just as SHRIMP does, but it uses a quite dif¬ ferent method of dealing with isobaric interferences. Negative secondary ions of the one nominal mass are selected, then accelerated by several MeV in a tandem van der Graaf accelerator. At the positive high-voltage terminal, they are then passed through a column of Ar gas at low pressure, which does two things; it strips up to several electrons from the negative atomic ions to form multiply-charged positive atomic ions, and it ruptures molecular ions into their constituent atoms in addition to stripping off electrons. As a result, all molecular ions are destroyed so the problem of molecular isobars vanishes. The positive ion beam is then mass-analysed at low reso¬ lution. AUSTRALIS is larger, more complex and more ex¬ pensive than SHRIMP. In addition, it is one of several projects at North Ryde that will share use of the tandem accelerator, so its capacity for geological analysis will be constrained. There will be the need to correct for vari¬ able elemental discrimination during sputtering, because the collected Pb+/U^ cannot be normalized to the same UOVU‘ in the absence of molecular ions. Only negative secondary ions can be used. Previous AMS experience has concentrated greatly on low counting-rates, and the use of sophisticated propor¬ tional counters that allow each ion to be identified from its charge-state and momentum which gives AMS its unique power for ultra-low trace-elements. In contrast with this, the U-Pb, Rb-Sr, Nd-Sm and oxygen isotope applications for geology demand the use of very high count-rates, to realise enough precision in a reasonable analytical time. AMS will thus have the same high count-rate linearity challenge in the detector as SHRIMP, and rather than detecting ions of 10 keV energy, AMS will need to handle ion energy of several MeV. Laser-sampling, inductively-coupled plasma ionization mass-spectrometry At the recent meeting of the ANZSMS in Sydney, G Hieftje from the Department of Chemistry, University of Indiana assessed the present and future status of 'Atomic Mass Spectrometry' (Burgoyne et al. 1995). He addressed the question of just how good ICPMS is, as presently available commercially, for making quantitative fast de¬ terminations of all the elements and isotopes. For in situ micro-analysis, a succession of pulsed laser shots is fired to ablate the target. This ablated material is swept away in a stream of gas, usually He, then the gas 113 Journal of the Royal Society of Western Australia, 79(1), March 1996 plus suspended particulate sample introduced to the in¬ ductively-coupled plasma to be ionized. Ionization is done with great efficiency hence high sensitivity could be available. Then the ions are extracted and mass-ana¬ lyzed either on a fast quadrupole analyzer for wide mass-range and speed, or on a double-focussing sector analyzer with multicollector if a relatively small mass range is sufficient. Hieftje emphasized several practical problems. The laser gives a non-continuous supply of the sample. The intensities of the sample ions rapidly rise with the start of the pulse then decline, and all the isotope and elemen¬ tal ratios must be measured while this is happening. Be¬ cause quadrupole mass-analysers measure masses in se¬ quence, the influence of a variable sample signal is diffi¬ cult to eliminate, and therefore use a time-of-flight mass- analyser which analyzes ions of all masses formed dur¬ ing the same short time-interval was necessary. Cur¬ rently, the TOF analyser has an efficiency-problem; the heavy ions must travel to the detector before taking an¬ other laser pulse. To improve this even to 10% efficient duty-cycle, complex electronic timing and electric field sweeps are necessary. Irrespective of the mass-analyser, isobar ic interferences remain even using plasma-ioniza¬ tion. Oxides are present for many elements. The exact species must be known and peak-stripping used if the mass-analyser operates at low resolution, or a high reso¬ lution mass-analyser should be used. There are also ma¬ trix-effects attributed to space-charge dispersion in high density beam-waists in the plasma; the light ions are spread more than the heavy ions. Overall, Hieftje made the point that present ICPMS instruments have serious deficiencies and that it might be 10 years before the nec¬ essary control on TOF analysers has been perfected. The above problems are greatly reduced if instrumen¬ tal needs can be restricted to isotopic ratios of single ele¬ ments within the limited mass-range of multi-collector sector analyzers. However, there is one ultimate com¬ parison that should be made between the SHRIMP ion microprobe and all ICPMS instruments, laser-sampling or otherwise. Which method of analysis has the greater intrinsic sensitivity? The fundamental definition of sen¬ sitivity for a given element is its useful yield, the ratio of ions collected for analysis per second to the number of atoms of that element consumed per second by the sam¬ pling process. Here, SHRIMP remains well ahead of ICPMS. For SHRIMP I at ^ 5000R, the useful yield for Pb in zircon was measured approximately as 0.7% (Compston & Williams, unpublished), and more accu¬ rately for SHRIMP II at ^ 1.5% (J W K Lee & I S Wil¬ liams, pers. comm.). The only published ICPMS value known to me is 0.22 % for Pb in a standard glass (Walder et al. 1993). Even lower values are reported verbally for Pb in zircon, 0.14% (P. Turner, pers. comm.). It is there¬ fore quite clear that the total sensitivity found for trace elements using laser-sampling ICPMS has been obtained by the use of much greater amounts of sample than SHRIMP. This is consistent with the multiple laser shots used per analysis as well as the hemispherical shape of the laser pits ( e.g . 20 pm diameter and 20 pm deep), as compared with single 20 x 25 pm elliptical spots using SHRIMP that are typically only 1 pm deep. Recent theoretical advances in sector mass analyser design Advances in ion optical theory have been made by H Matsuda, whose 1974 design we employed for SHRIMP I and II. Matsuda (1990) has now shown that it is possible to design families of double-focussing sector mass-analysers that have strong demagnification of the final image without an equivalent reduction in the mass- dispersion. This means that for a mass-analyser of a given size ('size7 means turning radius of the magnet), the mass-resolution at a given object-slit width (which sets the sensitivity) can be increased by a large factor. The particular design selected for the SHRIMP-RG will operate routinely at the same sensitivity as SHRIMP II, but with 4x greater mass-resolution. That in turn means that a number of new and important types of geochemi¬ cal analyses, not practical for SHRIMP II, could be under¬ taken using an ion probe based on the post-1990 design. There are some constraints that accompany the new design. One is that multiple collection can never be used, because the beam must be sent first through the mass- dispersing magnet rather than first through the energy- analyser. This arrangement is known as 'reverse-geom¬ etry7. Only one selected mass at a time can be transmit¬ ted through the energy analyser to the collector slit, thereby limiting the collection of data to the intrinsic in¬ efficiency of a single collector. On the other hand, the latter has the virtues of equal gain for all masses, sim¬ plicity and cheapness. In addition, reverse-geometry is well-known to give much higher abundance-sensitivity than forward-geometry i.e. the background of scattered ions will be very small so that low abundance peaks can be measured more accurately. The new design must be well-corrected for image-ab¬ errations to give highly demagnified images of good quality. The present SHRIMP IIs have low values for all 2nd order aberration coefficients, but the new design re¬ quires that several 3rd order coefficients must be low also. As in SHRIMP II, the design itself incorporates the means of nulling the necessary image aberrations through its particular selection of ion optical parameters. We there¬ fore consider that the attainment of good focussing in the new design will be no more difficult than in SHRIMP II. Practical design problems for a new model SHRIMP A constraint in the new design is the need for a much wider pole-face area than that used in the SHRIMP II magnet. At its widest, the secondary ion beam will di¬ verge to ± 120 mm within the sector magnet compared with ± 30 mm in SHRIMP II. This places a greater de¬ mand on the design and engineering of the magnet, es¬ pecially for the laminated pole-pieces that are necessary for fast field-switching. The question arises of whether the theoretical perfor¬ mance of the new designs can be made in practice, bear¬ ing in mind the imperfections of fabrication, variable magnetic properties of iron, and small misalignment of components. The same concerns were held for SHRIMP I but proved groundless. In addition, the manufacturer JEOL has successfully produced more than 100 small double-focussing mass spectrometers based on one of Matsuda's new designs, which gives confidence that the theoretical performance can be achieved. We are satisfied 114 Journal of the Royal Society of Western Australia, 79(1), March 1996 therefore that the necessary low image-aberrations will be achieved at the same level of accuracy in machining and alignment of ion optical components as used in the manufacture of SHRIMP II. Nevertheless we will take the precaution of including at least one ion optical stigmator to compensate empirically for image imperfec¬ tions. Advantages of the SHR1MP-RG The principal benefit of the SHRIMP-RG would be the removal of isobars, such as hydrides of the heavier ele¬ ments, that trouble or disallow a number of important geochemical applications. Examples are listed below but there will be many more. Discoveries in research cannot be predicted. In general, mass-resolution should be viewed as a commodity that cannot be overdone. In ad¬ dition, SHRIMP-RG could be operated with a very wide source slit for other applications, such as stable isotope analysis, that do not demand high mass resolution but are especially susceptible to variable mass discrimina¬ tion. Discrimination in the secondary ion mass-analyser cannot occur in the absence of beam truncation. In SHRIMP II, truncation occurs both at the source slit (at least 10 % of the beam at resolution 5500) and at other apertures during the transfer of ions from the analysed spot to the source slit. SHRIMP-RG offers the opportu¬ nity of improved phase-space matching between the mass-analyser and the analysed spot. If the present trun¬ cations can be avoided totally, only the sputtering pro¬ cess itself would remain to cause variable discrimination. Some geological applications for SHRIMP-RG Interference between Sr isotopes and Ca isotope dimers, such as wCa43Ca with H7Sr (m/Am of 16000) dis¬ courage serious ^Sr/^Sr fingerprint studies for the above and other Ca rich minerals. Sr isotope 'chemostratigraphy' is plagued by ambiguity caused by diagenetic alteration. Using whole rock samples, it is rarely clear whether a genuine shift in seawater isotope composition has been discovered in a stratigraphic sequence or a zone of later fluid alteration. Ion probe analysis will reveal diagenetic changes in many cases through their variable effects on a 10 micron spatial scale, and preserved areas should be identified. 40K - 40Ca dating of K-rich minerals There is a moderate prospect that radiogenic ^Ca will be retained more strongly than 40 Ar during cooling after igneous emplacement or metamorphism, so that ^K-^Ca ages of mica and K-feldspar would contribute to cooling history studies. The problem will be to find minerals in which the radiogenic ^Ca is not diluted by common Ca. There are good prospects that in situ 10 pm scale analy¬ ses will be better in this respect than bulk mineral sepa¬ rates. The very high abundance-sensitivity of SHRIMP- RG would minimize scattering from the adjacent intense 39K peak. In addition, it would be possible to resolve from ^Ca (m/Am of 28000) during analysis so that the ratio ^Ca / uCa would be a direct measure of the radio¬ genic enrichment without peak-stripping of K isotopes. 147Sm - 143Nd dating and 143Nd isotope labelling of REE- rich minerals. There are at least two problems with ion probe analysis for the above; the need for maximum sen¬ sitivity to amass enough counts for worthwhile precision. and the need for minimum peak-stripping of isobars to keep the analysis process simple. Some minerals such as perovskite have no obvious isobaric interferences other than hydrides, which SHRIMP-RG would remove at resolution 25000. More complex interferences must be expected for silicates and phosphates. Removal of 206Pb1H from 207Pb in the dating of hydrous minerals The variable water content of pitchblende or any other U-rich hydrous mineral inhibits us from attempting to date such minerals at present. Peak-stripping is unreli¬ able because mass 209 might be occupied by 20QBi as well as ^Pb'H. The ratio of ^Zr’H / wZr could be used for Zr-bearing minerals as shown by Long & Hinton (1984), but this requires very large mass jumps during data-col- lection. Although large jumps might be practical on SHRIMP II when the laminated magnet poles are fitted, it would be far better to remove the hydrides at resolu¬ tion 29000 which should be possible for SHRIMP-RG with some loss of sensitivity. Removal of hydrides in Hf isotope ratio measure¬ ment At present, peak-stripping must be used to remove 180, Yb, Lu and hydrides. Moreover, the hydrides of Yb and Lu differ in proportion to that of Hf. There is only one independent Hf isotope ratio left after the stripping to judge the veracity of the process, which is dangerous. The absence of the need to strip hydrides would be an enormous advantage, and it would be routinely possible to operate SHRIMP-RG at the necessary 22000 mass-reso¬ lution. Easy separation of light REE oxides from heavier REE metal ions Our REE measurement procedure is presently limited by the need to select the best mass-channels to avoid oxide overlaps. The routine operating resolution for SHRIMP-RG would remove these and other isobaric in¬ terferences that will arise when different minerals and REE levels are attempted. When trace-elements are present at the ppb level, many new possibilities exist for isobaric interferences that can be neglected at the ppm level. Examples are low abundance combinations of very light nuclides (e.g. boron) which have nuclidic mass 'sur¬ pluses' with oxides of heavier nuclides which have mass deficiencies. If geochemistry is to examine trace-elements in minerals at such levels, that problem must be faced. Meteoritic isotope anomalies SHRIMP-RG could resolve isobars in the Fe-group of elements, such as 58Ni and 58Fe, which opens the way for specialised cosmogenic isotope studies in addition to high-quality terrestrial geochemical /isotope analyses. Concluding Remarks A new Australian manufacturing company now exists to build SHRIMPs, Australian Scientific Instruments (ASI), in response to several requests made to the ANU for commercial sales. I wish to acknowledge here the very important role played by John de Laeter in precipitat¬ ing our decision to participate in this commercial venture. 115 Journal of the Royal Society of Western Australia, 79(1), March 1996 De Laeter formed the view that more than one SHRIMP was needed in Australia; one should also be located in Perth, for geological research at two Universities and to aid geological mapping by the State Geological Survey and by the mining industry. He also believed, and stated persuasively, that he could raise the necessary money for its purchase. At the same time, the demand for access to the ANU SHRIMP had become excessive, both from in¬ ternal ANU projects, from visiting scientists and from the Commonwealth Bureau of Mineral Resources. It be¬ came clear that a second and hopefully improved ANU SHRIMP could be built as a commercial prototype, fol¬ lowed by manufacture by ASI from this prototype and with the Perth Consortium led by John de Laeter as the first customer. Happily, this desirable course of events came to pass. Acknowledgements: I thank S W J Clement, T R Ireland, I S Williams, and S Sie for comments on various parts of the manuscript. References Andersen C A & Hinthome J R 1972 U, Th, Pb and REE abun¬ dances and 207Pb/206Pb ages of individual minerals in re¬ turned lunar material by ion microprobe mass analysis. Earth & Planetary Science Letters 14:195-200. Andersen C A & Hinthome J R 1973 Thermodynamic approach to the quantitative interpretation of sputtered ions in mass spectra. Analytical Chemistry 45:1421-1438 Bowring S A, Williams I S & Compston W 1989 3.97 Ga gneisses from the Slave province, Northwest Territories, Canada. Ge¬ ology 17:971-975. Burgoyne T W, Chambers D M, Heintz M A, Hieftje G M, Li G, Mahoney P P, Myers D P, Ray S J & Ross-Buckley B S 1995 Atomic mass spectrometry: the next generation. Australian and New Zealand Society for Mass Spectrometry Fifteenth Conference, Sydney NSW, WeO-Ol. Camacho A, Compston D, Compston W, Fanning C M, Sircombe K N & Williams I S 1993 U-Th-Pb dating of baddelevite, rutile and monazite using SHRIMP II. Research School of Earth Sciences Annual Report, 93-94. Clement S W J, Coles J, Compston W, Kerr D & Newstead G 1978 High resolution ion microprobe. Research School of Earth Sciences Annual Report, 172-174 Clement S W J, Coles J, Compston W, Schram N, Foster J J & Newstead G 1979 High resolution ion microprobe. Research School of Earth Sciences Annual Report, 159-161. Clement S W J, Compston W, Dabrowski R, Foster J J, Reinfrank R F & Williams I S 1992 SHRIMP II performance, isotopic mapping with SHRIMP II. Research School of Earth Sciences Annual Report, 70-72. Coles J, Compston W & Foster J J 1980 Microanalysis of Pb and S isotopes in galena. Research School of Earth Sciences An¬ nual Reportl75-179. Compston W, Foster J J, Williams I S, Coles ] N & Heydegger H R 1981 Ti isotopic composition using the ion microprobe. Research School of Earth Sciences Annual Report, 212-217. Compston W, Kinny P, Williams I S & Foster J J 1982 Hafnium isotopes in zircons. Research School of Earth Sciences An¬ nual Report, 249-252. Compston W, Williams I S, Bristow J & ireland T R 1985 Mea¬ surement of U-Pb ages of kimberlitic perovskite by ion mi¬ croprobe. Research School of Earth Sciences Annual Report, 82-83. Compston W, Williams I S, Froude D & Foster J J 1982 U-Pb dating of zircons from lower Palaeozoic tuffs using the ion microprobe. Research School of Earth Sciences Annual Re¬ port, 244-246. Compston W, Williams I S, Froude D O, Ireland T R, Kinny P D & Foster J ] 1982 Ion microprobe (SHRIMP) research and development. Research School of Earth Sciences Annual Re¬ port, 110-116. Compston W, Williams I S & Meyer C 1983 Ages determined on lunar zircons in thin section. Research School of Earth Sci¬ ences Annual Report, 116-119. Compston W, Williams I S & Meyer C 1984 U-Pb geochronol¬ ogy of zircon from lunar breccia 73217 using a sensitive high mass-resolution ion microprobe. Journal of Geophysical Re¬ search 89: Supplement B525-B534. Compston W, W'illiams I S & Meyer C 1988 A new insight into the early history of the moon. Research School of Earth Sci¬ ences Annual Report, 73-74. Compston W, Williams I S & Meyer C 1991 Initial Pb isotopic compositions of lunar granites as determined by ion micro¬ probe. Stable Isotope Geochemistry: A Tribute to Samuel Epstein (eds H P Taylor, J R O'Neil & I R Kaplan). The Geochemistry Society Special Publication 3:473-486. Compston W & Williams I S 1992 Ion probe ages for the British Ordovician and Silurian stratotypes. In: Global Perspectives on Ordovician Geology (eds B D Webby & J R Laurie). Pro¬ ceedings of the 6th International Symposium on the Ordovi¬ cian System, Sydney, 59-67. Compston W, Williams I S, Kirschvink J L, Zhang Zichao & Ma Guogan 1992. Zircon U-Pb ages for the Early Cambrian time-scale. Journal of the Geological Society of London 149:171-184. Eldridge C S, Compston W & Williams I S 1985 Determination of sulphur isotopic composition by ion microprobe. Research School of Earth Sciences Annual Report, 83-84. Eldridge C S, Compston W, Williams I S & Walshe J L 1987 In situ microanalysis for MS/32S ratios using the ion microprobe SHRIMP. International Journal of Mass Spectrometry and Ion Processes 76:65-83. Froude D O, Ireland T R, Kinny P D, Williams I S, Compston W, Williams I R & Meyers J S 1983 Ion microprobe identifica¬ tion of 4,100-4,200 Myr-old terrestrial zircons. Nature 304:616-618. Ireland T R & Compston W 1987 Large heterogeneous 2(1Mg excesses in a hibonite from the Murchison meteorite. Nature 327:689-692. Ireland T R, Compston W & Heydegger H R 1985 Titanium isotopic anomalies in hibonites from the Murchison carbon¬ aceous chondrite. Geochimica et Cosmochimica Acta 49:1989-1993. Ireland T R & Wlotzka F 1992. The oldest zircons in the solar system. Earth and Planetary Science Letters 109:1-10. Kinny P D, Wijbrans J R, Froude D O, Williams I S & Compston W 1990 Age constraints on the geological evolution of the Narryer Gneiss Complex, Western Australia. Australian Journal of Earth Sciences 37:51-69. Kinny P D, Compston W, & Williams I S 1991 A reconnaissance ion-probe study of hafnium isotopes in zircons. Geochimica et Cosmochimica Acta 55:849-859. Lee ] K W, Williams I S & Ellis D J 1995 Pb, U and Th diffusion in zircon. Research School of Earth Sciences Annual Report, 105-106. Long J V P & Hinton R W 1984 The intensity of metal hydride peaks in secondary positive-ion spectra from silicates. Inter¬ national Journal of Mass Specrometry and Ion Processes 55:307-318. Maas R, Kinny P D, Williams I S, Froude D O & Compston W 1992 The Earth's oldest crust: A geochronological and geo¬ chemical study of 3900D4200 Ma old zircons from Mt Narryer and Jack Hills, Western Australia. Geochimica et Cosmochimica Acta 54:2535-2547. Matsuda H 1974 Double focussing mass spectrometers of sec¬ ond order. International Journal of Mass Spectrometry and Ion Physics 14:219-233. 116 Journal of the Royal Society of Western Australia, 79(1), March 1996 Matsuda H 1990 High performance mass spectrometers of mag¬ netic section type. International Journal of Mass Spectrom¬ etry and Ion Processes 100:31-39. Nutman A P, Kinny P D, Compston W & Williams I S 1991 SHRIMP U-Pb zircon geochronology of the Narryer Gneiss Complex, Western Australia. Precambrian Research 52: 275- 300. Sie S H & Suter G F 1994 A microbeam AMS systems for miner- alogical applications. Nuclear Instruments and Methods in Physics Research B 92:221-226. Walder A J, Abell l D, Platzner A I & Freeman P A 1993 Lead isotope ratio measurement of NIST 610 glass by laser abla¬ tion inductively coupled plasma mass spectrometry. Spectrochimica Acta 48B:397-402. Williams I S & Claesson S 1987 Isotopic evidence for the Pre¬ cambrian provenance and Caledonian metamorphism of high grade paragneisses from the Seve Nappes, Scandina¬ vian Caledonides. II. Ion microprobe zircon U-Th-Pb. Con¬ tributions to Mineral Petrology 97:205-217. Williams I S, Compston W, Black L P, Ireland T R & Foster J 1984 Unsupported radiogenic Pb in zircon: a cause of anomalously high Pb-Pb, U-Pb and Th-Pb ages. Contribu¬ tions to Mineral Petrology 88:322-327. Williams I S, Compston W, Collerson K D, Arriens P A & Lovering J F 1983 A reassessment of the age of the Windmill Metamorphics, Casey Area. In: Antarctic Earth Science (eds R L Oliver, P R James & J B Jago). Australian Academy of Science, Canberra, 73-76. Williams I S, Shah J S & Stowe S 1996 Elemental and isotopic microanalvsis of zircons and backscattered electron contrast: Surface and Interface Analysis: in press. Williams I S, Black L P & Compston W 1983 Excess radiogenic Pb in zircons from Mt Stones, Antarctica. Research School of Earth Sciences Annual Report, 1390131. Williams I S, Black L P, Foster J J. Coles J N, Kinny P & Compston W 1982 Ancient zircon from the Fyfe Hills, Enderby Land, Antarctica. Research School of Earth Sciences Annual Report, 205-208. Williams I S, Compston W & Coles J N 1981 Zircon U-Pb geo¬ chronology using the ion microprobe. Research School of Earth Sciences Annual Report, 217-221. Williams I S, Harrison T M & Compston W 1988 Depth profiling by ion microprobe: a study of the diffusion of Pb, U and Th in zircon. Research School of Earth Sciences Annual Report, 72. Williams I S, Stowe S & Shah J S 1995 Microbeam imaging of zircon growth structure. Research School of Earth Sciences Annual Report, 106-107. Zeitler PK, Sutter J F, Williams I S, Zartman R & Tahirkheli R A K 1989. Geo-chronology and temperature history of the Nanga Parbat-Haramosh massif, Pakistan. Geological Soci¬ ety of America, Special Paper 23:1-22. 117 Journal of the Royal Society of Western Australia, 79:119-122, 1996 Zircons: What we need to know R T Pidgeon School of Applied Geology, Curtin University of Technology, Perth, WA 6001 Abstract The SHRIMP ion microprobe provides a new and exciting capability for determining U-Pb ages on parts of complex zircons. This opens up new opportunities for solving some equally complex geological problems, but this will depend on an understanding of the geological significance of the internal structures of the zircons, which is the key to the interpretation of the SHRIMP ages. CL imagery, HF etching, SEM imagery, X-ray and vibrational techniques for resolving zircon internal structures and crystallinity on a micro-scale are discussed. Introduction It is said that major advances in technology introduce as many new problems as they solve. Such may yet be the case for the new breed of high- resolution high-sensi¬ tivity micro-isotopic analytical instruments of which SHRIMP is the forerunner. SHRIMP has a broad analyti¬ cal capability but was conceived primarily for the pur¬ pose of analysing U-Th-Pb isotopic systems of micro-ar¬ eas of common uranium-bearing minerals of which zir¬ con is the prime example. SHRIMP was made at a time of rapid improvement in conventional geochronology, particularly in the techniques for measuring zircons on a scale of single crystals or parts of crystals involving high precision mass spectrometric analysis of tens of pro¬ grams of Pb with blanks expressed as numbers of atoms. At this time, in the late 1970s, a wealth of knowledge was available for the mineral zircon. For example the breakdown of the zircon structure under radiation, the relationship between isotopic stability and crystallinity, and the ability of crystallised zircon to resist dissolution under magmatic conditions and to retain an isotopic memory of its pre-magmatic age, were well known. It is with this last aspect that the first major advantage of SHRIMP micro-analysis became apparent. This is its ability to measure directly the age and chemical compo¬ sition of cores of ancient zircon (e.g. Williams 1992) en¬ cased in a mantle of younger magmatic zircon and then to determine the age of the zircon mantle. Such a feat is impractical using conventional technology and this alone signals that the scale of isotopic analysis and age deter¬ minations of zircons and other uranium-bearing miner¬ als has changed forever. From now on, the frontier of zircon geochronology is within individual zircon grains. Where zircons are uniformly the same age, for example in a rapidly crystallising felsic magma, there may be no advantage in making SHRIMP analyses on small areas of zircon crystals. However, it is well known that zircons can have complex structures which are presently incom¬ pletely understood, but which are thought to reflect events in their geological history. SHRIMP analytical spots can be located on these structures once they are © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 identified. An example of such a grain is contained in Figure 1. The grain pictured has a complex structure consisting of a crystalline core surrounded by an oscilla¬ tory zoned rim. What we need to know now is the geo¬ logical significance of these structures, so we can inter¬ pret SHRIMP U-Pb ages and the U-Th-Zr chemistry of cores and rims of such grains in terms of geological pro¬ cesses. We also need to have further information about the stability of zircon in a variety of geological environ¬ ments. Zircon microstructures Chemical and structural inhomogeneities in zircons, such as oscillatory zoning, have been known for many years. What is happening at present is that techniques are being developed and improved for the purpose of highlighting the internal structures of zircons so they can be studied in detail prior to analysis by the SHRIMP. The capabilities of some of the variety of techniques used for this purpose are as follows. Cathodoluminescence (CL) imagery is possibly the most popular technique for examining the internal structures of zircons. Cathodoluminescence is the light emitted from semi conductor or insulator material when an elec¬ tron beam creates electron hole pairs which then recom¬ bine to emit light in the wavelength range from the ultra¬ violet to the near infrared (200-2000 nm). Although the controls of the CL spectra are not well known, it has been proposed that concentrations of some trace constitu¬ ents are the determining factors. In zircons, Dy3+ is con¬ sidered to be the principal spectral factor though other constituents such as SnW, Eu2+, Tb3+, and Y3+ may also be CL emitters (c.g Hanchar & Miller 1993). The application of cathodoluminescence in highlighting micro-structures of magmatic zircons has been demonstrated by Vavra (1994). Spectral cathodoluminescence (the analysis of the CL light spectrum) also has interesting possibilities for in¬ vestigating the chemistry and structure of zircons on a micro-scale (e.g. Koschek and Lork, 1992) but up till now has been little used. Back scattered electron imaging (BSE) reveals contrasts in average atomic number of regions of a phase; the 119 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 1. A cathodoluminescence (CL) image and a photomicrograph of the same polished zircon grain that has been subjected to HF etching. A CL of this grain is weak, although faint euhedral zoning can be seen in the rim and the rounded outline of a central core can be distinguished. An irregularly-zoned rectangular area defined by relatively strong luminescence is present in the core and may be an earlier zircon fragment. B The HF etched zircon also shows the outline of the central zircon core. The core itself is highly crystalline and unetched except for a weakly expressed irregular grey zone within the core at a position approximately marginal to the rectangular centre identified by strong CL. This marginal zone does not appear to be recorded by the CL image. On the other hand the etched zircon shows no sign of the rectangular area in the core clearly defined by CL. HF etching shows the fine euhedral zoning of the zircon margin much more clearly than the CL image. higher the number, the more electrons an area will " reflect" and the brighter it will appear (Hanchar & Miller 1993). Hence the images can be termed atomic number (Z) contrast images. BSE imaging is now used widely in a variety of geological studies and is recognised as a powerful tool for studying zonation in accessory miner¬ als. (e.g. Paterson et al. 1992; Miller et al. 1992; Wayne et al. 1992). Hanchar & Miller (1993) consider that Hf is primarily responsible for the variability in BSE intensity, with U having a secondary effect. Both elements have a much higher atomic number than the principal constitu¬ ents of zircon (Zr, O, Si), so their substitution results in increased brightness. BSE and cathodoluminescence appear to be similar except that in general dark areas in CL are bright in BSE, and vice versa. However, in the light of explanations for these effects it is not evident that cathodoluminescence will always provide an identical image to BSE. This has yet to be investigated. HF etching is another technique used to highlight the internal structures of zircons. This measures the reactiv¬ ity of the zircon to HF vapour. Radiation damage disor¬ dering of the lattice appears to enhance reactivity to HF vapour but high concentrations of contaminant elements may also be a factor in increasing reactivity to HF. There is some suggestion that quartz can form through unmixing in some areas in a zircon lattice (Sommerauer 1976; McLaren et al. 1994) and this would certainly enhance reactivity with HF. Highly crystalline zircon is very resistant to HF etching. We have observed differences in the zircon images determined by CL and HF etching The example in Figure 1 shows HF etching and CL imagery' of a single zircon crystal from an Archaean granite from the Darling Range. The HF etching identifies an unetched central area surrounded by an inner rim of oscillatroy zoned zircon and a thin outer rim of granular etched zircon. The CL image shows features in the central part of the zircon not identified by HF etching and also the margin of the central zircon core, but does not resolve the zoned structure in the outer zircon rims. These differences may reflect the sensitivity of the two techniques to differences in zircon crystallinity. So far the possibility that differ¬ ences in the degree of metamictness of a zircon may in¬ fluence the CL emission has not been investigated. To investigate this and other questions involving the effects of recrystallisation on SI 1R1MP results, it is necessary to be able to measure quantitatively the crystallinity of an area on a zircon crystal the size of a SHRIMP analysis spot. Qualitative determination of the crystallinity of zircon on the scale of a SHRIMP spot X-ray powder diffractometry is commonly used to measure the crystallinity or the degree of 120 Journal of the Royal Society of Western Australia, 79(1), March 1996 metamictization in zircons. However, the amount of sample necessary for this method of analysis is much larger than the average mass of a single zircon grain. Thus, in most cases X-ray diffractometry gives only sum¬ mary information about entire grains or populations of grains. Important advances in understanding the nature of the metamict state and the relevance of this to the inter¬ pretation of zircon U-Pb ages have been made by McLaren et al (1994) using high-resolution transmission electron microscopy (HRTEM) studies of zircon on the scale of the zircon lattice. HRTEM gives local informa¬ tion about the degree of crystallinity, but at present this technique needs specially prepared samples and cannot be used for the routine determination of the crystallinity of micro-areas on the scale of SHRIMP analytical spots. Vibrational spectral analysis. Vibrational spectra tech¬ niques of infrared and Raman spectroscopy appear to have the greatest promise for investigating zircon struc¬ ture quantitatively on a microscale. Woodhead et al. (1991) report that the Infrared spectra of zircon vary as a function of metamictization. This is shown by increasing band widths and decreasing inten¬ sities with increasing U-Th content. This technique has also been used to investigate the presence of OH in the zircon lattice (Woodhead et al. 1991). These authors have applied the technique to single grains, but although McLaren et al. (1994) refer to measurements with this technique on a micro-scale, the author is not aware of applications on the scale of a SHRIMP analytical spot. On the other hand the ability of Raman spectrometry to quantitatively measure zircon crystallinity on this scale has been convincingly demonstrated (Nasdala et ah 1996; van Bronswijk & Pidgeon 1994). Modern Raman micro¬ probes can determine the crystallinity of zircon with a lateral resolution of about one micron and a volume reso¬ lution of less than 5 pm'. This opens up new opportuni¬ ties for correlating structural, chemical and isotopic data on the same micro area within a zircon grain. With de¬ creasing crystallinity, the Raman bands become less in¬ tense, are increasingly broadened and show lowered fre¬ quencies. In particular the increasing width of the inter¬ nal nu3(Si04) vibration (asymmetric stretching of Si04 tetrahedra) is sensitive enough to give a precise measure of the increasing degree of metamictization. Some questions Numerous question have been raised by studying the complex internal structures of zircons with these tech¬ niques combined with SHRIMP measurements. Some of these are as follows. Radiation damage ages versus SHRIMP ages. An in¬ vestigation of the relationship between zircon crystallin¬ ity and structures revealed by cathodoluminescence has not been made, although Raman spectroscopy has con¬ firmed the relative crystallinity of zircon areas shown by HF etching. With careful calibration, involving standard zircons, crystallinity measurements by Raman spectros¬ copy can be combined with U-Pb concentrations deter¬ mined by SHRIMP, to calculate a radiation damage age for a SHRIMP analytical spot. In this way it would be possible to construct a radiation damage age map of the zircon and investigate the recrystallisation history of the zircons and the relationship of recrystallisation to the dis¬ tribution of U,Th,Hf and U-Pb ages. Diffusion versus reaction in zircon crystals. Zircon is considered the most refractory of the geochronologically useful uranium bearing minerals and is known to sur¬ vive anatexis, granitic magma emplacement and mag¬ matic crystallisation without loss of radiogenic Pb. This, together with the sharpness of zonal boundaries in zir¬ cons as old as 3000Ma, suggests that diffusion in zircon is extremely slow. However there is now evidence from HF etching studies and Raman spectroscopy that suggests some igneous zircons have undergone recrystallisation, with accompanying loss of Pb and U, soon after crystallisation. This does not appear to be related to ra¬ diation damage but is thought to involve expulsion of contaminant elements during lattice recrystallisation. The significance of low U,Th crystalline cores. The cores of many zircons consist of low U-Th crystalline zir¬ con, very resistant to HF etching, but sometimes show structure revealed by CL (Figure 1). Cores can represent xenocrysts of much older source material and have older SHRIMP ages. In some cases apparent cores have ages close to the magmatic age. These may represent older source material that has been updated or reflect processes affecting early formed zircon in the magma. Cathodoluminescence combined with HF etching and Raman spectroscopy will play an important part in re¬ solving this question. Structural complexities in other minerals. Catho¬ doluminescence and etching techniques appear to be lim¬ ited in investigating minerals other than zircon. How¬ ever, back scattered electron imagery has successfully re¬ vealed complex structures in titanite (Paterson & Stephens 1992) and monazite (De Wolf et al. 1993) and it is likely that other minerals will also have complex struc¬ tures which will be important in interpreting their U-Pb ages. References De Wolf C P, Belshaw N & O'Nions R K 1993 A metamorphic history from micro-scale 207Pb/206Pb chronometry of Ar¬ chaean monazite. Earth and Planetary Science Letters 120 : 207-220. Hancher J M & Miller C F 1993 Zircon zonation patterns as revealed by cathodoluminescence and backscattered electron images: Implications for interpretation of complex crustal histones. Chemical Geology 110:1-13. Koschek G & Lork A 1992 Material analysis with catho¬ doluminescence standard spectra. Scanning 14:100-103. McLaren A C, Fitzgerald J D & Williams I S 1994 The micro¬ structure of zircon and its influence on the age determina¬ tion from Pb/U isotopic ratios measured by ion microprobe. Geochimica et Cosmochimica Acta 58:993-1005. Miller C F, Hanchar J M, Wooden J L, Bennett V C, Flarrison T M, Wark D & A Foster D A 1992 Source region of a granite batholith: evidence from lower crustal xenoliths and inher¬ ited accessory' minerals. Transactions of the Royal Society of Edinburgh Earth Sciences 83:49-62. Nasdala L, Pidgeon R T & Wolf D 1996 Heterogeneous metamictization of zircon on a micro scale. Geochimica et Cosmochimica Acta 60:1091-1097. Paterson B A & Stephens W E 1992 Kinetically induced compo¬ sitional zoning in titanite: implications for accessory-phase/ 121 Journal of the Royal Society of Western Australia, 79(1), March 1996 melt partitioning of trace elements. Contributions to Miner¬ alogy and Petrology 109:373-385. Paterson B A, Stephens W E, Rogers G, Williams I S, Hinton R W & Herd D A 1992 The nature of zircon inheritance in two granite plutons. Transactions of the Royal Society of Edinburgh Earth Sciences 83:459-471. Sommerauer J 1976 Die Chemisch-physikalische stabilitat natiirlicher zirkone und ihr U-(Th)-Pb system. Ph D Thesis, ETH, Zurich. van Bronswijk W & Pidgeon R T 1994 The determination of crystallinity in naturally occuring zircons by Raman Spec¬ troscopy. XIVth International Conference on Raman Spec¬ troscopy. Extra Booklet E13-E14. Vavra G 1994 Systematics of internal zircon morphology in major Variscan granitoid types. Contributions to Mineral¬ ogy. and Petrology 117:331-344. Wayne D M, Sinha A K & Hewitt D A 1992 Differential re¬ sponse of zircon U-Pb isotopic systematics to metamorphism across a lithological boundary: an example from the Hope Valley Shear Zone southeastern Massachusetts. USA Contri¬ butions to Mineralogy and Petrology 109:408-420. Williams I S 1992 Some observations on the use of zircon U-Pb geochronology in the study of granitic rocks. Transactions of the Royal Society of Edinburgh Earth Sciences 83:447-458. Woodhead J A, Rossman G R & Silver L T 1991 The metamictization of zircon: radiation dose-dependent struc¬ tural characteristics. American Mineralogist 76:74-82. 122 Journal of the Royal Society of Western Australia, 79:123-129, 1996 A review of Pb-isotope constraints on the genesis of lode-gold deposits in the Yilgarn Craton, Western Australia N J McNaughton & D I Groves Key Centre for Strategic Mineral Deposits, Department of Geology and Geophysics, The University of Western Australia, Nedlands WA 6907 Abstract Archaean lode-gold deposits within the Yilgarn Craton of Western Australia have made a major contribution to world gold production over the last century. However, models for the genesis of this class of deposits, particularly aspects of fluid and solute sources, remain controversial and require specific tests of their veracity. A systematic study of the initial Pb-isotopic compositions of Yilgarn lode-gold deposits and potential source rocks for ore components places constraints on possible source regions, and hence genetic models of ore formation. The inferred initial isotopic composition of Pb in Yilgarn ore deposits correlates with that of the crustal Pb in the rocks of hosting greenstone belts, but is independent of the hostrock to the deposit. Accordingly, the ore-deposit initial Pb-isotope compositions show regional variations with crustal rocks. These are inferred to vary with the age, composition and degree of metamorphic- magmatic depletion of the lower to middle crust, which is considered to be the source of Pb in the deposits. With the available geological constraints, the heterogeneous source can be modelled as follows; derivation from early Archaean (>3.7 Ga) crust which is metamorphically depleted and/ or partially melted at ca. 3.3 Ga; further reworking from ca. 2.7 Ga; and then sampling of the Pb in this heterogeneous crust by auriferous fluids at ca. 2.63 Ga. The Pb isotope modelling of source regions is not unique, but requires that the ore deposit Pb has an older crustal component which varies systematically on a regional scale independent of the composition, metamorphic grade and age of crustal rocks. This constrains the source of Pb in lode-gold deposits to include a component from the high-grade mid- to lower-crust via either metamorphic or magmatic processes. Introduction Lode-gold deposits in the Yilgarn Craton account for much of the annual gold production in Australia (i.e. 175 tonnes Au in 1995). Understanding the mode of forma¬ tion of such deposits is pivotal to discovering new unex¬ posed deposits, and hence much effort has been directed towards the establishment of robust deposit models which are predictive. The isotopic composition of Pb in ore deposits may be used as a tracer of Pb sources, and hence help constrain mineralisation processes. To use Pb isotopes as a source tracer, it is important to determine; (i) the initial Pb-isotopic composition and age of the ore system, and (ii) the coeval Pb in potential sources in, or below, the terrain hosting the gold deposit. Over the last decade, colleagues and students of the authors at the Key Centre for Strategic Mineral Deposits have utilised the joint mass spectrometer facilities at Curtin University as part of multidisciplinary research into lode-gold deposits. This contribution describes the role of Pb-isotope tracer studies in developing the cur¬ rent understanding of Archaean lode-gold deposit source-regions within the Yilgarn Craton of Western Australia. Geological background The Yilgarn Craton is composed of; (i) greenstone belts consisting of volcanic, sedimentary and intrusive © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 rocks, which are typically metamorphosed and variably deformed, and (ii) complex granitoid-gneiss terrains which separate the greenstone belts (Gee ct al 1981). Al¬ most all of the Yilgarn lode-gold production comes from areas within greenstone belts. Mineralisation may occur in various styles, from brec¬ cia lodes, vein sets, laminated veins or shear zones and as disseminated lodes. However, there are a number of common characteristics of lode-gold deposits (summarised by Groves 1993; Groves el al 1995); an epigenetic timing of mineralisation, varying broadly from syn-metamor- phic in deposits hosted by mid-amphibolite to granulite facies rocks, to post-metamorphic in deposits hosted by greenschist to low-amphibolite facies rocks, a ubiquitous structural control on mineralisation with deposits commonly forming adjacent to secondary or higher order splays off major faults or shear zones, a gold-only element associa¬ tion where base metal enrichments are generally negli¬ gible and K, LILE, CO?, Na or Ca are added to proximal wallrocks, an H^O ± C02 ± CH4, low-salinity ore fluid, and a range of crustal depths, from near surface (very low metamorphic grade) environments to granulite facies conditions, for gold mineralisation. Age of lode-gold mineralisation For Pb isotope modelling purposes, it is important to constrain the age of lode-gold mineralisation. As pre¬ dicted by the crustal continuum model, Archaean lode- gold deposits of the Yilgarn Craton are broadly coeval at 123 Journal of the Royal Society of Western Australia, 79(1), March 1996 2.64-2.63 Ga (Groves 1993). Although the available reli¬ able data are few, Pb-Pb and Sm-Nd isochrons, mineral U/Pb ages and Ar-Ar plateau ages for hydrothermal or alteration minerals associated with mineralisation are broadly synchronous (Clark et al 1989; Bamicoat et al. 1991; Wang et al. 1993; Kent 1994; Kent & McDougall 1995), although at least one major deposit (Mt Charlotte) may be 30 m.y. younger than the nearby Golden Mile mineralisation (Kent & McDougall 1995). In addition, age constraints provided by the youngest granitoids cut by lode-gold deposits and dykes which cut mineralisation are in agreement with the direct dating of mineralisation (e.g. Kent 1994; Bloem et al 1995; Kent & McDougall 1995). Hence, for the Pb isotope modelling presented be¬ low, an age of 2.63 Ga is taken for mineralisation. In addition to the dating of the lode-gold deposits, relative timing criteria and age estimates for metamorphic and deformational events (McNaughton et al. 1990a) are also compatible with this age. Magmatic activity broadly syn¬ chronous with mineralisation includes suites of late granitoids in the Murchison Province (Wiedenbeck & Watkins 1993; Yeats 1996), and a suite of late granites within the granitoid-gneiss terrains of the southwestern Southern Cross Province (Qiu et al. 1995). Lead reservoirs coeval with lode-gold mineralisation Unlike ore systems where the initial Pb-isotope com¬ positions can often be determined from ore sulphides, it is more difficult to determine the coeval Pb-isotope com¬ position of the rocks which may represent the reservoirs for Pb contributing to the ore fluid. The only common mineral which may allow initial Pb-isotope ratios to be estimated is K-feldspar, which is generally restricted to granitoids. Other minerals and rocks have U/Pb>0 and hence their Pb-isotope ratios have changed since forma¬ tion. However, these materials may be analysed, and knowing the age of mineralisation, allow constraints to be placed on their Pb-isotope ratios at the time of mineralisation. Granitoids Granitoids of the Norseman-Wiluna belt of the East¬ ern Goldfields Province of Western Australia have been analysed and data are shown in Figure 1. The estimated composition of crustal rocks within the belt at 2.63 Ga is based on the Pb in a synvolcanic VHMS deposit at Teu¬ tonic Bore, and granitoids at Stennet Rocks, both compo¬ sitions recalculated from their formation age to 2.63 Ga (McNaughton et al. 1990a, b). The initial Pb-isotope compositions of the granitoids, based on K-feldspar data, are divided into three groups (Fig V). The main group shows all data constrained to within two 2.63 Ga isochrons shown as parallel lines, and with K-feldspar data concentrated towards the inter¬ section of the crosscutting crustal line. This group in¬ cludes all available data on granitoids which are intru¬ sive into the greenstone belts north of the Victory mine, Kambalda. Towards the west and south of Kambalda, the granitoids fall above these parallel lines, whereas, towards the northeast and east in the Eastern Goldfields Province they fall below these lines (both open circles; Fig 1). For comparison, granitoid K-feldspar data from 206Pb/204Pb Figure 1. Common Pb-isotope plot showing K-feldspar data compared to the Stacey & Kramers (1975) lithospheric growth curve and the estimated composition of the crustal rocks of the Norseman-Wiluna belt (NWB) at the time of gold mineralisation ( i.e . from Stennet Rocks to Teutonic Bore at 2.63 Ga; McNaughton et al. 1990b). Error bars are 2a. Symbols; closed circles = NWB granitoids intrusive into greenstones north of Kambalda with 2.63 Ga isochrons bracketing these data; open circles (below parallel lines) = intrusive granitoids of the North¬ eastern Goldfields; open circles (above parallel lines) and Stennet Rocks = intrusive granitoids of the Norseman area and south of Coolgardie; open squares = Murchison and Southern Cross data. References; Oversby (1975). McNaughton & Bickle (1987), Perring & McNaughton (1992), Knight (1994). Wang et al. (1993), Ojala (1995), and unpublished data. the Murchison and Southern Cross Provinces are also shown (open squares; Fig 1), and broadly overlap the granitoid trend in the Norseman-Wiluna belt. The granitoids of the northeastern Eastern Goldfields Province are interpreted to have a different source region to all other areas. The inferred boundary of this isotopi- cally distinctive terrane approximately corresponds to the position of the Keith-Kilkinny lineament. Towards the south and west of Kambalda, older crustal rocks in the granitoid source regions have been inferred on the basis of isotopic data (Oversby 1975; Perring & McNaughton 1992). Greenstones The Pb-isotope data for lithologies within greenstone belts in the Norseman-Wiluna belt are shown in Figure 2. Relative to the estimated crust at 2.63 Ga and the main granitoid field (shown as parallel lines; from Fig 1), the rocks of the Norseman region plot largely above the granitoid data, whereas those of the Kambalda- Kalgoorlie-Mt Pleasant region plot below. These data are modern Pb-isotopic compositions. The Pb in each rock at the time of mineralisation is not known, but will plot along a 2.63 Ga isochron fitted to each data point but at a less radiogenic composition. Scatter in the data may be induced by open system behaviour of Pb-U between vol- canism and modern times, but the general adherence of data to consistent fields suggests this effect is minor. 124 207Pb/204Pb Journal of the Royal Society of Western Australia, 79(1), March 1996 Therefore, the initial Pb-isotope ratios for the greenstones could lie on the 2.63 Ga crustal mixing line, slightly above and below the granitoid data. 206Pb/204Pb Figure 2. Common Pb-isotope plot showing the Norseman- Wiluna belt crustal mixing line and 2.63 Ga isochrons bracket¬ ing the granitoid Pb field (from Fig 1) compared to all green¬ stone data for the Kambalda-Kalgoorlie-Mt Pleasant region (tri¬ angles), and the Norseman region (squares). Data from; open squares (Perring & McNaughton 1992; McCuaig & McNaughton, unpublished data); closed triangles (Roddick 1983; Chauvel et al. 1985; McNaughton et al. 1988; Gebre- Mariam et al. 1993). As with the granitoid data, the distinctly more radio¬ genic Pb in greenstones from the Norseman area, com¬ pared to the Kambalda-Kalgoorlie-Mt Pleasant area, rep¬ resents a fundamental and provincial difference in the source of the Pb in both the greenstones and intrusive granitoids. The (gradational?) boundary between these different regions lies somewhere south of Kambalda and north of Norseman. Importantly, the southern part of the Norseman-Wiluna belt shows the most radiogenic Pb in both the granitoid and greenstone lithologies. To date, there are no published Pb-isotope data on greenstone lithologies from outside the Norseman-Wiluna belt. Initial lead isotope composition of ore systems The initial Pb of lode-gold ore systems is best deter¬ mined from Pb-rich minerals genetically associated with the ore. Lead sulphide (galena) or tellurides are used, if they are present, but Fe-sulphides (pyrite, pyrrhotite) are more common and reliably preserve the initial Pb-iso- tope compositions in some cases ( e.g . Ho et al. 1994). In a few cases, K-feldspar, sphalerite and scheelite associated with the ore have yielded reliable initial Pb-isotope ratios (Barnicoat et al. 1991; unpublished data). Galena, with U/Pb = 0, is considered to be the best sampling medium to establish the initial Pb-isotope com¬ position of lode-gold ore system. Galena occurs as a trace mineral in a proportion of lode-gold deposits. However, previous studies have shown that it does not always pre¬ serve the initial Pb of the ore system (e.g. Browning et al. 1983; Perring & McNaughton 1990; McNaughton et al. in press). Similarly, Pb tellurides may be used (Browning et al. 1983), but are rarer than galena. Iron sulphides are common ore-related minerals in the majority of deposits, and, with proper sampling methodologies (e.g. Ho et al. 1994), can yield the initial Pb-isotope ratios. For most deposits, a number of ore sulphide samples have been analysed and an assessment of the data must be made to determine if the initial Pb ratios are preserved. In depos¬ its where a range of Pb-isotope compositions occur, the least radiogenic composition is taken as the best estimate of the initial Pb (e.£. Ho et al. 1994). More radiogenic compositions are attributed to in situ U-decay, from U either within the sulphide or within inclusions in the sulphide, or from more radiogenic Pb mobilised into the sulphide from an external source after sulphide forma¬ tion. In the following discussion, only initial Pb-isotope compositions, which are inferred using the criteria dis¬ cussed above and by McNaughton et al. (1990b) and Ho et al. (1994), are considered. Figure 3 presents the initial uranogenic Pb-isotope ra¬ tios of deposits within the Norseman-Wiluna belt. The data form a linear array, and show provinciality i.e. a systematic change in Pb-isotope composition between geographic regions (Groves et al. 1995). The least radio¬ genic end of the array is represented by deposits to the north of the belt, whereas the most radiogenic data are from around Norseman at the southern end of the belt. These differences correlate with both the granitoid (Fig 1) and greenstone data (Fig 2) from these regions. Fur¬ ther, within some areas along the regional strike of the greenstone belts, the initial Pb-isotope ratios from depos¬ its are invariant at the scale of >100kms (e.g. Mt Pleasant area to Kalgoorlie to Victory mine, Kambalda; McNaughton et al. 1993), whereas other areas show sys¬ tematic changes on this scale (e.g. from Victory mine, Kambalda to Norseman; Fig 3). There is a weak correlation 206Pb/204Pb Figure 3. Common Pb-isotope plot showing the initial Pb for gold deposits from the Norseman-Wiluna belt, Stacey & Kramers (1975) growth curve and the crustal mixing line at 2.63 Ga from Fig 1. The gold deposit data fall into groups which correlate with geographical regions. Data from; McNaughton et al. (1990b, 1993) and Groves et al. (1995). 125 207Pb/204Pb Journal of the Royal Society of Western Australia, 79(1), March 1996 between initial Pb-isotope compositions of greenstone- hosted deposits and location across regional strike to¬ wards the granitoid-gneiss terrains (McNaughton et al 1993). However, the initial Pb-isotope compositions for the deposits appears to be independent of hostrock (McNaughton et al 1993; Ho et al 1994). The data for lode-gold deposits elsewhere in the Yilgam craton are shown in Figure 4. There is a signifi¬ cant overlap with the Norseman-Wiluna belt data, al¬ though some deposits in the Murchison Province show significantly lower 206Pb/204Pb at similar 207Pb/204Pb (Fig 4). Griffin's Find in the Southern Cross Province is spe¬ cial in that the deposit occurs in the highest metamorphic grade setting of any deposit so far recognised (discussed below), and it is noteworthy that it has a similar initial Pb-isotope composition to other deposits (Fig 4). 144 - * - j - * - 1 - * - ■ - * - ' - ■ - ■ - ■ - 1 - 4 - 13.0 13.2 13.4 13.6 13.8 14.0 14.2 14.4 206Pb/204Pb Figure 4. Common Pb-isotope plot showing the initial Pb for gold deposits from the Eastern Goldfields (closed squares), Murchison and Southern Cross Provinces (open circles, with Griffin's Find indicated), and the Norseman-Wiluna belt crustal mixing line at 2.63 Ga from Figure 1. Data from; McNaughton et al (1990, 1993), Vielreicher et al (1994) and Ojala (1995). Discussion Archaean Yilgam Craton: Source of Pb in deposits of the Norseman-Wiluna belt The initial Pb-isotope data from the lode-gold depos¬ its of the Yilgam craton (Fig 4) are compared with the estimated crustal Pb at the time of mineralisation in Fig¬ ure 5. The ore deposit data for the Norseman-Wiluna belt correlate well with the 2.63 Ga crustal trend (Fig 3). As the crustal trend at 2.63 Ga is based on estimates of the Pb-isotope compositions of exposed upper crustal rocks, it could be inferred that the source of Pb in the deposits is also from the same upper crustal reservoir. However, the radiogenic endmember of the 2.63 Ga crustal trend is represented by a granitoid which undoubtedly attained its initial Pb from its lower to middle crustal source. Simi¬ larly, the lower granulite facies deposit at Griffin's Find in the Southern Cross Province (Barnicoat et al 1991) also has an initial Pb-isotope composition close to this crustal trend (Fig 4), suggesting that both the upper and mid- lower crust in the craton were broadly similar in their Pb-isotope characteristics at 2.63 Ga. Similarly, the over¬ lap of Pb-isotope data from deposits from the Murchison, Southern Cross and Eastern Goldfields provinces (Fig 4) suggests that this heterogenous upper-lower crustal source of Pb found in the gold deposits occurs on a re¬ gional scale, and that mineralisation was broadly coeval on this scale. 206Pb/204Pb Figure 5. Common Pb-isotope plot showing the ore-deposit ini¬ tial Pb-isotope compositions from Figure 4, and the inferred crustal endmembers (modified from McNaughton et al 1990b). The radiogenic nature of Pb from deposits, green¬ stones and granitoids of the Norseman region, compared to other regions, implies that all owe their Pb characteris¬ tics to a common source. Given that zircon xenocrysts and cores to ca. 2.7-2.63 Ga magmatic zircon grains within granitoids from this region and other parts of the Yilgam have ages up to 3 5-3.3 Ga (Hill et al 1989; Qiu et al 1995), the preferred explanation of the Norseman ra¬ diogenic signatures is that they are due to older crustal rocks underlying this region (cf Oversby 1975; Campbell & Hill, 1988). This radiogenic Pb signature can be mod¬ elled (see below) and must include a significant period of evolution, prior to 2.63 Ga, with enhanced U/Pb, typical of upper crustal rocks. The least radiogenic endmember of the 2.63 Ga crustal trend, and overlapping data for lode-gold deposits (Fig 5), fall close to the global lithospheric growth curve of Stacey & Kramers (1975), but at model ages significantly older than the age of mineralisation (McNaughton et al 1990a). Assuming that the lithospheric growth curve is valid, this implies that the Yilgam craton has not evolved like typical Archaean crust (e.g. Richards 1983), and in¬ cludes a stage of retarded Pb-isotope growth prior to 2.63 Ga. Modelling of potential sources to achieve the ob¬ served results yield an infinite number of solutions, and other geological constraints must be applied. The sim¬ plest model requires older felsic crust undergoing high grade metamorphism and depletion of U with respect to Pb, at least a few hundred million years prior to mineralisation. This process would produce the Teutonic Bore endmember Pb (of the crustal trend) at 2.63 Ga with 126 Journal of the Royal Society of Western Australia, 79(1), March 1996 an older model Pb age (Fig 3). As the radiogenic endmember of the Norseman-Wiluna Belt 2.63 Ga crustal trend also requires an older crustal source, it follows that all the Pb from the 2.63 Ga crustal trend in Figure 5 re¬ quires an older crustal component at 2.63 Ga. The greenstone data (Fig 2) do not show the large variation recorded in either the crustal trend or the ore deposit data. Assuming the outcropping greenstones are representative of greenstones at deeper levels, it follows that the greenstones cannot be the dominant source of Pb in the deposits. A deep source of Pb (i.e. below the green¬ stones) for the lode-gold ore deposits from underlying crust isotopically similar with the granitoids must be in¬ ferred at 2.63 Ga. A deep source of Pb in the deposits is also compatible with the observation that the initial Pb is independent of hostrock composition and crustal depth of mineralisation (i.e. depths corresponding to lower granulite to subgreenschist facies metamorphic condi¬ tions), as well as the epigenetic timing and ubiquitous relationship of deposits to splays of major faults and shear zones, which could allow ore fluids from deeper levels to access the entire crustal section. Archaean Yilgam Craton: Source of Pb in all deposits The ore-deposit initial Pb data from the entire Yilgam Craton are also shown in Figure 5 with estimated com¬ positions of crustal reservoirs. The Pb mixing trend at 2.63 Ga for the crustal rocks of the Norseman-Wiluna belt in Figure 5 is interpreted to represent mid-lower to upper crustal Pb from older crust. The radiogenic endmember reflects a growth stage with enhanced U/Pb (i.e. enriched or upper crustal Pb), whereas the other end is from depleted or mid-lower crust with low U/Pb and retarded Pb-isotope growth. Data falling significantly to the left of this trend in Figure 5 come from the northern Yilgarn Craton (i.e. north of Leonora in the Norseman-Wiluna Belt and the northern Murchison Province), and require an additional Pb component which has experienced retarded growth in 206Pb/2a4Pb with respect to 207Pb/2WPb, compared to the other data. However, the timing of the inferred U-deple- tion event cannot be the same as that producing the Norseman-Wiluna belt crustal and deposit trend at 2.63 Ga. At the time of gold mineralisation, Figure 5 shows the trend for crust which is 3.3 Ga old and was variably depleted-enriched (i.e. lower and higher U/Pb, respec¬ tively) at ca. 3.3 Ga. This crustal array has been fitted to the upper end of the Norseman-Wiluna belt crustal array at 2.63 Ga to essentially define the second side of a tri¬ angle of Pb-isotope compositions which enclose all initial Pb-isotope compositions for the ore deposits in Figure 5. The slope of the upper side of the triangle is defined by the ages for t, and t2 (i.e. 3.3 and 2.63 Ga, respectively; see McNaughton 1987). To obtain crustal Pb-isotope compositions significantly above the lithospheric growth curve requires crust older than 3.3 Ga which evolves up to 3.3 Ga with an enhanced U/Pb (i.e. as upper crust). Modelling granitoid source regions suggests that this crust is at least 3.7 Ga old (Cassidy & McNaughton, unpublished observations). This is approximately the age of the oldest crustal segment known in the Yilgarn Craton (Kinny et al. 1990), and 3.7 Ga is used in the modelling (shown in Fig 6) as the age of the protocrust underlying the Yilgarn craton. Figure 6. Common Pb-isotope plot showing the evolution of Pb from 3.7 Ga crust to produce the trianglar field of crustal Pb- isotope compositions at 2.63 Ga, which encompasses the Yilgarn ore-deposit initial Pb-isotope compositions shown in Fig 5. The triangular field for Pb-isotope compositions shown in Figure 5 can be modelled by forming a mafic- intermediate crust at ca. 3.7 Ga. The intermediate compo¬ nent is interpreted to have evolved with an average U/Pb higher than the Stacey & Kramers (1975) growth curve, whereas the mafic components remained near the lithos¬ pheric growth curves (i.e. U/Pb near the average lithos¬ phere) until about 3.3 Ga, at which time the 3.7 Ga crust was variably depleted (i.e. metamorphosed-partially melted) in U with respect to Pb. On average, the mafic components are interpreted to have experienced an over¬ all lowering of U/Pb, leading to retarded Pb-isotope growth from 3.3 Ga until 2.63 Ga, and produced com¬ positions near, but slightly below, the growth curve, and to the left of the Stacey & Kramers lithosphere at 2.63 Ga (Fig 6; represented by the Teutonic Bore data in Fig 3). The 3.7 Ga old crustal components of intermediate composition must have evolved well above the Stacey & Kramers growth curve by 3.3 Ga (Fig 6), and were then variably depleted during the 3.3 Ga event, producing the upper side of the inferred trianglar field in Figures 5 and 6 by 2.63 Ga. The average U/Pb of the intermediate crust must have been lowered to less than that of the mafic crust during the 3.3 Ga event, probably due to the preferential partial melting of the intermediate crust. This is interpreted to have produced very retarded Pb-isotope growth of the intermediate crust between 3.3. and 2.63 Ga. The modelling described above allows for timing con¬ straints of known and inferred formation ages of Yilgarn rocks and events. However, in detail, the modelled U/Pb for the post-3.3 Ga crust is unusually low, and the U/Pb of the 3.7 Ga intermediate crust prior to 3.3 Ga is unusu¬ ally high. Both these anomalies would be lessened if the initial crust was older than 3.7 Ga, and/or the major depletion event was older than 3.3 Ga. The inferred ca. 4.2 Ga felsic crustal rocks of the northwestern Yilgarn (Kinny et al. 1990), and the presence of 3. 3-3. 5 Ga zircon xenocrysts in granitoids, indicate that such alternative models are feasible. 127 Journal of the Royal Society of Western Australia, 79(1), March 1996 Conclusions Lead isotope data from Yilgarn lode-gold deposits and presently exposed rocks may place important constraints on the crustal evolution and metallogenesis of the craton. Modelling of Pb-isotope compositions of rocks and ores within the known and partly inferred temporal frame¬ work of the craton can yield an infinite number of solu¬ tions. The Pb-isotope modelling suggested in this review (Fig 6) suggests that the Yilgarn craton is underlain by crustal rocks which may be >37 Ga old. This old crust comprised both mafic and intermediate components which evolved to form heterogeneous crust by the time of a major depletion event (metamorphism-partial melt¬ ing) at ca. 3.3 Ga or older. Thereafter, the depleted crust evolved to produce Pb-isotope compositions at 2.63 Ga, the time of broadly coeval gold mineralisation across the craton, which lie within a trianglar field of Pb-isotope compositions on a 207Pb/204Pb vs 206Pb/204Pb diagram, de¬ fined by crustal components with different geochemical compositions and histories. The initial Pb-isotope compo¬ sitions of the gold deposits closely adhere to this triangu¬ lar field, and, together with geological constraints, are best interpreted to indicate that all deposits obtained their Pb mostly from the older depleted crust at mid to lower crustal levels. This conclusion is compatible with a crustal continuum of deposits (Groves 1993) from granu- lite to subgreenschist facies settings, a ubiquitous spatial association with major crustal-scale shear or fault zones, the occurrence of deposits in all rock types and a similar¬ ity of geochemical and ore fluid characteristics on a cra¬ ton scale (Ho et al. 1992; McNaughton et al. 1992, 1993). The deep source of Pb constrains the fluid source to high metamorphic grade regions, and implies that mineralisation resulted from either high grade metamor¬ phic or deep magmatic processes, Acknoivledgements: This summary reflects a decade of research by many past and former members of the Key Centre for Strategic Mineral Depos¬ its and it precursors; all are thanked. Special thanks go to M Dahl for technical assistance, the Curtin University mass spectrometrists, particu¬ larly W Chisolm, I Fletcher and K Rosman, continued funding from UWA, Curtin and ARC, and finally, and most importantly, to John de Laeter whose foresight and energies were instrumental in establishing the world-class mass spectrometry facilities in Western Australia. References Bamicoat A C, Fare R J, Groves D I & McNaughton N J 1991 Syn-metamorphic lode-gold deposits in high-grade Archaean settings. Geology 19:921-924. Bloem E J M, McNaughton N J, Groves D I & Ridley J R 1995 An indirect age determination of gold mineralisation at Corinthia Mine, Yilgarn Block, Western Australia. Australian Journal of Earth Sciences 42:447-451. Browning P, Groves D l, Blockley J G & Rosman K J R 1983 Lead isotope constraints on the age and source of gold min¬ eralization in the Archean Yilgarn Block, Western Australia. Economic Geology 82:971-986. Campbell I H & Hill R 1 1988 A two-stage model for the forma¬ tion of the granite-greenstone terrains of the Kalgoorlie- Norseman area, Western Australia. Earth and Planetary Sci¬ ence Letters 90:11-25. Chauvel C, Dupre B & Jenner G A 1985 The Sm-Nd age of Kambalda is 500 Ma too old! Earth and Planetary Science Letters 74:315-324. Clark M E, Carmichael D M, Hodgson C J & Fu M 1989 Wallrock alteration, Victory gold mine, Kambalda, Western Australia: processes and P-T-XC02 conditions of metasoma¬ tism. Economic Geology Monograph 6:445-459. Gebre-Mariam M, Groves D 1, McNaughton N J, Mikucki E J & Veamcombe J R 1993 Archaean Au-Ag mineralization at Racetrack, near Kalgoorlie, Western Australia: a high crustal- level expression of Archaean composite lode-gold system. Mineralium Deposita 28:375-387. Gee R D, Baxter J L, Wilde S A & Williams 1 R 1981 Crustal development in the Archaean Yilgarn Block, Western Aus¬ tralia. Geological Society of Australia, Special Publication 7:43-56. Groves D I 1993 The crustal continuum model for late- Archaean lode-gold deposits of the Yilgarn Block, Western Australia. Mineralium Deposita 28:366-374. Groves D I, Ridley J R, Bloem E, Gebre-Mariam M, Hagermann S G, Knight J T, McNaughton N J & Ojala J 1995 Lode-gold deposits of the Yilgarn Block: products of late-Archaean crustal-scale overpressure hydrothermal systems. In: Early Precambrian Processes (eds M P Coward and A C Reis) Spe¬ cial Publications of the Geological Society (London) 95:155- 172. Hill R I, Campbell I H & Compston W 1989 Age and origin of granitic rocks in the Kalgoorlie-Norseman region of Western Australia: implications for the origin of Archaean crust. Geochimica et Cosmochimica Acta 53:1259-1275. Ho S E, McNaughton N J & Groves D I 1994 Criteria for deter¬ mining initial lead isotopic compositions of pyrite in Ar¬ chaean lode-gold deposits: a case study at Victory, Kambalda, Western Australia. Chemical Geology 111:57-84. Ho S E, Groves D I, McNaughton N J & Mikucki E J 1992 The source of ore fluids and solutes in Archaean lode-gold de¬ posits of Western Australia. Journal of Volcanology and Geo¬ thermal Research 50:173-196. Kent A J R 1994 Geochronological constraints on the timing of Archaean gold mineralisation in the Yilgarn Craton, Western Australia. Ph.D. Thesis, The Australian National University. Kent A J R & McDougall I 1995 "Ar-^Ar and U-Pb constraints on the timing of gold mineralization in the Kalgoorlie Gold Field, Western Australia. Economic Geology 90:845-859. Kinny P D, Wijbrans J R, Froude D O, Williams I S & Compston W 1990 Age constraints on the geological evolution of the Narryer Gneiss Complex, Western Australia. Australian Jour¬ nal of Earth Sciences 37:51-70. Knight J T 1994 The geology and genesis of Archaean amphibo¬ lite-facies lode-gold deposits in the Coolgardie Goldfield, Western Australia, with special emphasis on the role of granitoids. Ph.D. Thesis, The University of Western Austra¬ lia. McNaughton N J 1987 Lead isotope systematics for Archaean sulphide studies. In: Recent Advances in Understanding Pre¬ cambrian Gold Deposits (eds S E Ho & D I Groves). Geology Department and University Extension, University of Western Australia, Perth, Publication 11:181-188. McNaughton N J & Bickle M J 1987 K-feldspar Pb-Pb isotope systematics of Archaean post-kinematic granitoid intrusions of the Diemals area, central Yilgarn Block. Chemical Geology 66:193-208. 67 McNaughton N J, Cassidy K F, Dahl N, de Laeter J R, Golding S D, Groves D I, Ho S E, Mueller A G, Perring C S, Sang J H & Turner J V 1992 The source of ore components in lode-gold deposits of the Yilgarn Block, Western Australia. In: The Ar¬ chaean: Terrains, Crustal Processes and Metallogeny (eds J E Glover and S E Ho). Geology Department and University Extension, University of Western Australia, Perth, Publica¬ tion 22:351-363. McNaughton N J, Cassidy K F, Dahl N, Groves D I, Perring C S & Sang J H 1990b Constraints on genesis of primary gold deposits: Lead isotope studies. In: Gold Deposits of the Ar¬ chaean Yilgarn Block, Western Australia: Nature, Genesis and Exploration Guides (eds S E Ho D I Groves D I and J M Bennett). Geology Department and University Extension, 128 Journal of the Royal Society of Western Australia, 79(1), March 1996 University of Western Australia, Perth, Publication 20:226- 236. McNaughton N J, Cassidy K F, Groves D I & Perring C S 1990a Constraints on genesis of primary gold deposits: Timing of mineralization. In: Gold Deposits of the Archaean Yilgam Block, Western Australia: Nature, Genesis and Exploration Guides (eds S E Ho D I Groves D I and J M Bennett).Geology Department and University Extension, University of Western Australia, Perth, Publication 20:221-225. McNaughton N J, Frost K M & Groves D I 1988 Ground melting and ocellar komatiites: a lead isotopic study at Kambalda, Western Australia. Geological Magazine 125:285-295. McNaughton N J, Groves D I, Jackson S, Newton P & Sang J H in press Lead isotopic compositions of sulphides in Archaean lode-gold deposits of the Yilgam Block, Western Australia: monitors of cryptic Proterozoic events. Mineralium Deposita, in press. McNaughton N J, Groves D I & Witt W K 1993 The source of lead in Archaean lode-gold deposits of the Menzies- Kalgoorlie-Kambalda region, Yilgam Block, Western Austra¬ lia. Mineralium Deposita 28:495-501, Ojala V J 1995 Structural and depositional controls on gold mineralisation at the Granny Smith mine. Western Australia. Ph.D. Thesis, The University of Western Australia. Oversby V M 1975 Lead isotopic systematics and ages of Ar¬ chaean acid intrusives in the Kalgoorlie-Norseman area, Western Australia. Geochimica et Cosmochimica Acta 39:1107-1125. Perring C S & McNaughton N J 1990 Proterozoic remobilization of ore metals within Archaean gold deposits: lead isotope evidence from Norseman, Western Australia. Australian Journal of Earth Sciences 37:369-372. Perring C S & McNaughton N J 1992 The relationship between Archaean gold mineralization and spatially associated minor intrusives at the Kambalda and Norseman gold camps, West¬ ern Australia: lead isotope evidence. Mineralium Deposita 27:10-22. Qiu Y M, McNaughton N J & Groves D I 1995 Lead isotope and SHRIMP zircon age constraints on the timing and sources of late- to post-tectonic granitoids from the central southern Yilgam Craton, Western Australia. The 3rd Australian Con¬ ference on Geochronology and Isotope Geoscience, Perth, 28. Richards J R 1983 Lead isotopes as indicators of old stable craton in Western Australia. Geochemical Journal 17:247-255. Roddick J C M 1984 Emplacement and metamorphism of Ar¬ chaean mafic volcanics at Kambalda, Western Australia: geo¬ chemical and isotopic constraints. Geochimica Cosmochimica Acta 48:1305-1318. Stacey J S & Kramers J D 1975 Approximation of terrestrial lead isotope evolution by a two-stage model. Earth and Planetary Sciences Letters 26:207-221. Vielreicher R M, Groves D I, Ridley J R & McNaughton N J 1994 A replacement origin for the BIF-hosted gold deposit at Mt Morgans, Yilgarn Block, Western Australia. Ore Geology Re¬ views 9:325-347. Wang L G, McNaughton N J & Groves D 1 1993 An overview of the relationship between granitoid intrusions and gold min¬ eralization in the Archaean Murchison Province, Western Australia. Mineralium Deposita 28:482-494. Wiedenbeck M & Watkins K P 1993 A time scale for granitoid emplacement in the Archaean Murchison Province, Western Australia, by single zircon geochronology. Precambrian Re¬ search 61:1-26. Yeats C J 1996 Overprinting of a VHMS-style event by late Ar¬ chaean lode-gold mineralisation at the base-metal rich Mount Gibson Gold Deposits, Western Australia. PhD Thesis, Uni¬ versity of Western Australia. 129 Journal of the Royal Society of Western Australia, 79:131-139, 1996 Isotopic constraints on the age and early differentation of the Earth M T McCulloch Research School of Earth Sciences, Australian National University, Canberra 0200 Abstract The Earth's age and early differentiation history are re-evaluated using updated isotopic con¬ straints. From the most primitive terrestrial Pb isotopic compositions found at Isua Greenland, and the Pilbara of Western Australia, combined with precise geochronology of these localities, an age 4.49 ± 0.02 Ga is obtained. This is interpreted as the mean age of core formation as U/Pb is fractionated due to sequestering of Pb into the Earth's core. The long-lived Rb-Sr isotopic system provides constraints on the time interval for the accretion of the Earth as Rb underwent significant depletion by volatile loss during accretion of the Earth or its precursor planetesimals. A primitive measured ^Sr/^Sr initial ratio of 0.700502 ± 10 has been obtained for an early Archean (3.46 Ga) barite from the Pilbara Block of Western Australia. Using conservative models for the evolution of Rb/Sr in the early Archean mantle allows an estimate to be placed on the Earth's initial Sr ratio at ^4.50 Ga, of 0.69940 ± 10. This is significantly higher than that measured for the Moon (0.69900 ± 2) or in the achondrite, Angra dos Reis (0.69894 ± 2) and for a Rb/Sr ratio of **1/2 of chondrites corresponds to a mean age for accretion of the Earth of 4.48 ± 0.04 Ga. The now extinct ,4*Sm-142Nd (T1/2146=103 lCPyrs) combined with the long-lived 147Sm-143Nd isoto¬ pic systematics can also be used to provide limits on the time of early differentiation of the Earth. High precision analyses of the oldest (3. 8-3.9 Ga) Archean gneisses from Greenland (Amitsoq and Akilia gneisses), and Canada (Acasta gneiss) do not show measurable (> ± lOppm) variations of 142Nd, in contrast to the 33 ppm l42Nd excess reported for an Archean sample. The general lack of 142Nd variations, combined with the presence of highly positive c values (+4.0) at 3.9 Ga, indi¬ cates that the record of large-scale Sm/Nd fractionation events was not preserved in the early- Earth from 4.56 Ga to ^4.3 Ga. This is consistent with large-scale planetary re-homogenisation during ongoing accretion of the Earth. The lack of isotopic anomalies in short-lived decay systems, together with the Pb and Sr isotopic constraints is thus consistent with core formation and accre¬ tion of the Earth occurring over an ^100 Ma interval following the formation of meteorites at 4.56 Ga. Introduction A fundamental question concerning the origin of the Earth is its age. Did the Earth accrete shortly after the formation of primitive meteorites at 4.56 Ga or was there a substantial interval before final accretion and early dif¬ ferentiation of the Earth? Conclusive answers to this question remain elusive, but a number of lines of evi¬ dence that are considered here now appear to indicate that the accretion of the Earth was not completed until at least 100 Ma after the formation of chondrites. The diffi¬ culty m establishing a precise age for the Earth arises from the lack of a geological record of the first ^500 Ma of Earth history. The only known remnants from this period are 4.27 Ga detrital zircons found in an Archean sediment from the Jack Hills region of Western Australia (Compston & Pidgeon 1986). This provides a firm mini¬ mum age for the Earth, but the question of what hap¬ pened between 4.27 Ga and 4.56 Ga remains. Insights into the Earth's earliest history can however be obtained from both the long-lived Rb-Sr, Sm-Nd and U-Pb sys¬ tems as well as the now extinct t46Sm-142Nd decay series. © Royal Society of Western Austialia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 The formation of the Earth by accretion of planetismals was undoubtedly a complex process involv¬ ing large-scale melting, internal differentiation of a me¬ tallic core and formation of a crust. These processes were characterised by distinctive patterns of geochemical frac¬ tionation, and it is this time, or mean age of parent/ daughter fractionation which is registered by isotope sys¬ tems. in the case of the U-Pb system, Pb has as a strongly chalcophile character and as a result was probably strongly partitioned into the Earth's core. Thus the sub¬ stantially higher (**100x) U/Pb ratio for the Earth's mantle relative to carbonaceous chondrites is attributed to Pb being sequestered into the Earth's core along with sulphur. For this reason, the U-Pb system provides con¬ straints on the average time of formation of the Earth's core. For the Rb-Sr system, the major cause of fraction¬ ation is the volatile nature of Rb compared to Sr. Thus the **10x depletion of Rb in the Earth compared to chon¬ drites is attributed to volatile loss processes during accre¬ tion of the Earth. In contrast to these systems, Sm and Nd are refractory elements, and are strongly fractionated only during magmatic differentiation processes. The question of the 'age of the Earth' is therefore be consid¬ ered from a number of different aspects; the time-scales for accretion, core formation and early mantle differen¬ tiation. 131 Journal of the Royal Society of Western Australia, 79(1), March 1996 Pb isotopic constraints on core formation The application of long-lived radiometric chronom¬ eters to the question of the Earth's age was pioneered by Rutherford (1929). Based on the present-day abundances of the long-lived isotopes of uranium 235U and 238U, he deduced an approximate estimate for the timing of stel¬ lar nucleosynthesis and hence a maximum age of the Earth's of **4 billion years. The first precise estimate of the Earth's age was made by Patterson (1956) who showed that the Pb isotopic evolution of the Earth and meteorites approximately corresponds to a single-stage evolution from about 4.55 Ga to the present. This con¬ clusion was based on the close proximity of the average modern terrestrial Pb to a 4.55 Ga isochron defined by meteorites. The value of modern terrestrial Pb used by Patterson was derived from the average Pb isotopic com¬ position of marine sediments in the Pacific ocean. Since this pioneering work, it has now become apparent that modern rocks have a much more diverse range of Pb isotopic compositions indicative of complex, multi-stage evolutionary histories. For example, Allegre et al (1995) has shown that modern mid-ocean ridge basalts (MORB's) have a wide range of generally younger single- stage Pb-Pb ages of from 4.3 Ga to 4.6 Ga. In order to critically address the question of whether the Earth is significantly younger than its meteorite par¬ ent bodies, it is useful to consider the initial Pb isotopic composition of the oldest terrestrial rocks. Primitive ter¬ restrial Pb compositions provide the least equivocal record of U-Th-Pb fractionation following the Earth's for¬ mation and hence the most reliable isotopic constraints on the Earth's age and early differentiation history. For a single-stage model of Pb evolution, the relationship be¬ tween the parent 238U and 235U and daughter 206Pb and 207Pb isotopes is given by; 207pb/2°4pb(t) - 207Pb/204pb(I) 235U(e X Tcore . eX t) 206Pb/204Pb(t) - 2(*Pb/204Pb(I) 238U(e x W - e* [) where 207Pb/2(HPb(I) and 206Pb/204Pb(I) are the initial Pb compositions at the time of core formation T . 207Pb/ 204Pb(t) and 206Pb/204Pb(t) are the initial Pb compositions in early Archean rocks of age t (Table 1). The decay constants for 238U and 235U are X = 0.155125 JE'* 1 and X1 = 0.98485 4E'1 respectively, with 235U/238U = 1/137.88. This equation enables the age of the Earth's core to be calcu¬ lated using estimates of the Earth's initial Pb composi¬ tion derived from meteorites, together with the most primitive Pb compositions preserved in ancient terres¬ trial rocks. Listed in Table 1 are samples with primitive Pb isotopic compositions; ie 206Pb/204Pb <12. An equally important constraint in the application of this equation is uncertainties in age of the early Archean initial Pb's (e.g. Gancarz & Wasserburg 1977). SHRIMP U-Pb dating of zircons has recently been undertaken at both Isua (Nutman et al 1995) and in the Pilbara, for the Duffer dacite (McNaughton et al 1993), the host of the Big Stubby galena. This improved chronology is particularly helpful in the interpretation of initial Pb's from the Isua supracrustal belt (Appel et al 1978, Richards & Appel 1987). Nutman et al (1995) have shown that this com¬ plex includes at least two felsic volcanic complexes with ages of 3.708 ± 0.003 Ga and 3.807 ± 0.002 Ga, with the galenas listed in Table 1 being associated with the older volcanics of the western section. Following the approach of Gancarz & Wasserburg (1977), the initial Pb's listed Table 1, are presented, using plots of 207Pb/20bPb versus 204Pb/206Pb (Fig 1). This dia¬ gram emphasises the evolution of Pb in the early Archean. It is apparent from Figure lb that plausible ranges for 238U/204Pb (p) of from 8.5 to 9.5, are required to be consistent with the evolution of Isua, Big Stubby as well as younger conformable Pb's (Stacey & Kramers 1974). With this range of p the Pb-Pb age of the Earth's core, Tcore is estimated to be from 4.52 Ga to 4.46 Ga (Table 1). This age is clearly younger than that of chon- dritic meteorites (4.563 Ga), and hence the value of initial Pb given by Canyon Diablo (Tatsumoto et al 1973) must be adjusted to take into account the Pb evolution from the time of meteorite formation, at 4.563 Ga to «M.50 Ga. The average p for carbonaceous chondrites is low (< 0.2) and based on the correlation of Pb/U versus K/U, a p « 1 is inferred (Allegre et al 1995) for the proto-Earth prior to core formation. Thus the corrected initial Pb for the Earth at **4.50 Ga is given by; 206pb/2CUpb(I) = 20.pb/2Wpb(CD) + ^(^4,56 . eXTcore) £) Table 1 Most primitive terrestrial Pb isotopic compositions and the age of the Earth's core Locality Age (Ma) 206Pb/2,,4Pb 207Pb/2,,4Pb 208Pb/2(,4Pb Tcore(Ga) ISUA 261619“ 3807 ± 2 11.176 263654“ 3807 ± 2 11.311 2966 b 3808 ± 2 11.146 Amitsoqc 3590 ± 50 11.468 PILBARA Big Stubbyde 3465 ± 3 11.912 Canyon Diablo 4563 f 9.307 INITIAL Earth Pb 4490« 9.330 13.047 31.130 4.492 9.11 13.147 31.267 4.419 11.0 13.025 31.148 4.510 8.72 13.203 31.365 4.462 8.46 13.707 31.838 4.495 8.71 10.294 29.476 10.339 29.495 F. = 1.0 “ Richards & Appel (1987); b Appel et al. (1978); c Gancarz & Wasserburg (1977); d Richards (1986); * Pidgeon (1978V 1 Tatsumoto et al. (1973); * this study. 132 Journal of the Royal Society of Western Australia, 79(1), March 1996 n CL CD O .Q Q- n- o CM 204pb/206pb -Q CL o CM 0.080 0.084 0.088 0.092 0.096 204Pb/206pb Figure 1A. Pb growth curve plotted using 2,,hPb/2(,7Pb versus 2(^Pb/20hPb. Primitive terrestrial Pb's are shown with solid symbols and open circles are galena Pb's tabulated by Stacey & Kramers (1974). Earth initial Pb from Canyon Diablo (Tatsumoto et al. 1973). B. Plot of the most primitive terrestrial Pb (20hPb/2,l4Pb < 12) for galena from Isua, West Greenland (Appell et al. 1978; Richards & Appel 1987) and Big Stubby, Pilbara (Pidgeon 1978; Richards 1986). Amitsoq feldspar from Gancarz & Wasserburg (1977). Open circle shows disturbed younger sample from Isua (Richards & Appel 1987). The curves show the Pb evolution for p = 8.5, 9.0 and 9.5. The best fit to both the Isua and Big Stubby Pb's is given for an age of the Earth's core of TC()re = 4.49 Ga and u2 = 9.0. Initial Earth Pb is evolved Canyon Diablo using M, = 1 from 4.56 Ga to 4.49 Ga. 133 Journal of the Royal Society of Western Australia, 79(1), March 1996 0 Q- E 03 C / ) 7 6 5 4 3 2 1 0 □ Stacey Kramers : Isua Amitsoq Big Stubby Isua 4.49+0.02 Ga 4.35 4.40 4.45 4.50 Pb-Pb Age of the Earth's Core (Ga) 4.55 Figure 2. Histogram of single-stage Pb-Pb ages for the Earth's core calculated using data points shown in Fig 1. Arrow shows the best estimate for the mean age of formation of the Earth's core based on Pb- Pb ages from Isua and Big Stubby. and an equivalent equation for 207Pb/204Pb(I). With this minor correction to the Earth's initial Pb and with p, = 1, (the p from 4.56 Ga to 4.50 Ga) the best estimate of the age of the Earth's core using equation 1, is T. re = 4.49 ± 0.02 Ga. This latter estimate is based on both the Isua as well as Big Stubby initial Pb's and is consistent with p2 = 9 (the p from 4.49 Ga to the present). It is noted that the initial Pb from the Amitsoq feldspar is not compatible with this data set, which may be due to the effects of metamorphic disturbance as discussed by Gancarz & Wasserburg (1977) and/or uncertainties in the crystallisation age of the Amitsoq feldspar sample. Using the same approach, T.ore ages have also been calculated for the set of younger conformable galena Pb's listed by Stacey & Kramers (1974). A histogram of Tcore ages is shown in Figure 2 with ages of from 4.40 Ga to 4.50 Ga; these ages are somewhat younger than those from Isua and Big Stubby. Taken at face value, they are indicative of a longer interval (100-200 Ma) for core for¬ mation, however the younger galena have significantly more evolved Pb compositions and therefore their inter¬ pretation is more problematic. Sr isotopic constraints on timescales for the accretion of the Earth Based on the same reasoning as outlined for Pb, the most rigorous constraints on the Earth's initial Sr are those derived from the most primitive terrestrial Sr isoto¬ pic compositions. With this in mind, McCulloch (1994) reported a measured 87Sr/wSr initial ratio of 0.700502 ± 9 for the 3.46 Ga barite from the North Pole dome of West¬ ern Australia. This result, together with a less precise initial Sr ratio for a basaltic komatiite from the Onverwacht Group (Jahn & Shih 1974), indicates an initial Sr isotopic composition for the early Archean (3.46 Ga) mantle of 0.70050 ± 2. This determination currently represents the most reliable value for the early Archean mantle and has important ramifications for the origin of the Earth-Moon system as well as the evolution of the Archean mantle. The key question is how accurate and how representative is this value for the early Archean mantle? Barite is a useful recorder of ^Sr/^Sr ratios, but due to its hydrothermal origin may provide only an up¬ per limit to the Sr composition of the Earth's mantle, although in the early Archean the difference in Sr isoto¬ pic composition between seawater and the upper mantle is probably insignificant (McCulloch 1994). The consis¬ tency of the Pilbara barite result with the few available reliable determinations of initial Sr in the Archean mantle (Jahn & Shih 1974; Machado et al. 1986) together with the presence of primitive Pb in the nearby Big Stubby galena (Pidgeon 1978; Richards 1986) indicates that this is a plausible result. The Earth's initial Sr ratio (BEBI = Bulk Earth Best Initial) can be calculated from the early Archean barite using the following single-stage model: BEBI = (S7Sr/84Sr)tarite - (»7Rb/*Sr)EARTH (e»V - e") (3) where Tacc = 4.50 Ga is the mean age for of accretion of the Earth, t = 3.46 Ga, the age of the Pilbara barite (McNaughton et al 1993), \ = Q.0142^*1 and (87Rb/*Sr)EARTH the ratio from Tacc to t. Using a value for (S/Rb/^Sr)EARTH of 0.07 ± 0.007 (McCulloch 1994), which reflects the partially depleted character of the early Archean mantle, equation 3 gives a well defined value of BEBI = 0.69940 ± 10. This estimate is significantly greater than that previously assumed using the initial Sr compo¬ sition of achondrites {e.g. Angra dos Reis (ADOR) with ^Sr/^Sr = 0.69893 ± 2). Moreover, BEBI is significantly greater than the initial Sr ratio for the Moon (Nyquist et al 134 Journal of the Royal Society of Western Australia, 79(1), March 1996 1973; Alibert et al 1994) of LUNI = 0.69900 ± 2 and thus allows chronological constraints to be placed on the first 100 Ma of evolution of the Earth-Moon system. If we consider ADOR as a zero reference time-line for the accretion of planetary bodies, then the difference in the initial Sr compositions between ADOR & BEBI can be translated into a formation interval for planetary accre¬ tion relative to ADOR using; ISr(BEBI) - ISr(ADOR) = X AT (87Rb/86Sr)p (4) where AT is the formation interval and (^Rb/^Sr) the parent/daughter ratio in the planetary precursor bodies. By far the most significant uncertainty in applying the AT-ISr methodology to planetary accretion is in estimat¬ ing the (^Rb/^Sr) ratio of the terrestrial planetisimals. The use of a chonaritic Rb/Sr ratio cannot be justified as the volatile/refractory ratio of the asteroid belt, as sampled by meteorites, is probably not representative of the solar nebular from which the terrestrial planetesimals accumulated (Taylor & Norman 1990). The ratio of vola¬ tile/involatile elements would be expected to increase ra¬ dially outwards from the Sun following the T-Tauri stage, implying a lower (^Rb/^Sr)^ for proto-Earth mate¬ rials than for chondrites. Unfortunately, the volatile/ involatile ratio of the terrestrial planetesimals cannot be directly ascertained as these bodies where presumably swept up and accreted into the Earth-Moon system. The other factor that needs to be considered is the extent of volatile loss during planetary formation, par¬ ticularly if this occurs via accretion of only a relatively small number of more massive bodies with large gravita¬ tional potential's and consequently high accretional ener¬ gies. Rb is volatile under reducing conditions at tem¬ peratures <1000 0C whereas it is likely that the Earth's outer surface reached temperatures of >1500 °C (Stevenson 1987). With reasonably efficient mixing of the upper portion of the Earth, concomitant with impact-ac¬ cretion, it is highly likely that a significant fraction of the proto-Earth's volatile budget including Rb (and Pb) would not have been accreted. The actual magnitude of the Rb/Sr (i.e. volatile/refractory) fractionation that oc¬ curred during accretion of the Earth although obviously significant, is difficult to quantify. An upper limit to the extent of fractionation is given by the difference between the bulk Earth and chondritic 87Rb/^Sr ratios of 0.082 and ^0.8 respectively. This indicates a maximum pos¬ sible fractionation of Rb/Sr resulting from a combination of both planetary accretion and Solar nebular processes of ^10x. An intermediate estimate is given by (H7Rb/ ^Srjp = 0.4 which implies a depletion of Rb/Sr in the terrestrial region of the solar nebular of ^xl compared to chondrites. For BEBI = 0.69940, this corresponds to AT of ^80 Ma relative to ADOR (Fig 3). Substantially greater depletion in the volatile content of the terrestrial plan- 4.25 4.30 4.35 4.40 4.45 4.50 4.55 4.60 TIME (Ga) Figure 3. Plot of initial Sr composition versus time for the Earth = BEBI (McCulloch 1994), Moon shown with solid circles (Alibert et al. 1994) and ordinary chondrites with open symbols (Brannon et al. 1987), relative to the primitive differentiated achondrite Angra dos Reis = ADOR (Lugmair & Galer 1992; McCulloch 1994). Bj = Bjurbole, G= Guarena, PR = Peace River, SB = Soko Banja, (Brannon et al. 1987). The mean age for accretion of the Earth ranges from ^=80 Ma to 100 Ma for a 87Rb/86Sr ratio of ^0.4, which assumes an ^50% loss of volatiles during accretion. 135 Journal of the Royal Society of Western Australia, 79(1), March 1996 3.7 3.8 3.9 4 4.1 4.2 4.3 4.4 4.5 TIME (Ga) TIME (Ga) Figure 4A. Plot of b versus time for an initial 14nSm/l'HSm - 0.007 (Prinzhofer pf nl ioq?t Tf t r j?cc a . j ,, ,, „ , ,n the first 200-300 Ma of Earth history then within the analytical would be \alues. The general absence of terrestrial K]J excesses (McCulloch & Bennett 1993 therefore suecesh that frarMnn^n J? rf-4i reservoirs were not preserved until after M.3 Ga, after which was extinct nniv,™,i-, ^ fractionated terrestrial so far been reported by Harper & Jacobsen (1992). Lunar data ,s from Nvquist et J/ (1995) ^'"'piot of/ vers^ ¥ * 4ppm'hds 136 Journal of the Royal Society of Western Australia, 79(1), March 1996 etesimals is probably unlikely, as this would give AT > 100 Ma, which would correspond to an age for the Earth that is younger than that obtained for terrestrial core for¬ mation (=4.49 Ga) using Pb-Pb isotope systematics out¬ lined previously. Sm-Nd constraints on the Earth's age and early differentiation In contrast to the U-Pb and Rb-Sr systems, the rare earth elements Sm and Nd are refractory elements that are strongly fractionated only during magmatic differen¬ tiation processes. The Sm-Nd isotopic system is however unique in having both extinct and "live" parent-daugh¬ ter decay systems; the now extinct l46Sm-142Nd with a half-life of 1.03 104 yrs, and the commonly used long- lived 147Sm-143Nd system with a half-life of 1.06 10n yrs. Owing to its short half-life, 14*Sm effectively became ex¬ tinct at =4.30 Ga, at which time the 142Nd/144Nd ratio became essentially invariant. Positive or negative devia¬ tions in the observed ,42Nd/144Nd ratio must therefore represent the effects of differentiation processes opera¬ tive prior to 4.30 Ga. Thus ,42Nd isotopic abundances have the potential to reveal events occurring in the first 250 million years of the Earth history. The combined 142Nd-143Nd isotopic systematics can po¬ tentially provide constraints on the magnitude and the longevity of chemical differentiation in the early Earth due, for example, to the presence of an early magma ocean. Positive or negative deviations of el42 values from the bulk Earth value are possible if a REE fractionated reservoir was formed within the first 100 to 200 Ma of Earth history and if this reservoir survived intact as a closed system until it could be sampled via magmatic processes and preserved in the oldest terrestrial rocks (3.70 - 3.96 Ga). This scenario is shown in Figure 4A where e142 values are plotted versus the time of reservoir fractionation and isolation. In these calculations, it is assumed that the Solar System had l4*Sm/144Sm = 0.007 at 4.56 Ga, the same Sm isotopic composition as that deter¬ mined from meteorites (Prinzhofer et at. 1992; Nyquist et al 1995). The magnitude of any el42 anomaly is directly dependent on the fractionation (f) where f = [(Sm/Nd)/ (Sm/Nd)rHUR - 1], as well as the time T when the reser¬ voir fractionated and became isolated. Following Harper & Jacobsen (1992), the £]42 effects can be calculated from the following relationship; e142(t) = f Q142[146Sm / ,44Sm] [e ^i^(T0-T)-e'xi46(T0-t)] (5) where Q]42 = 354 Ga1, T0= 4.56 Ga, T = time of mantle differentiation, and t = crystallisation age of the rock. Apollo 17 lunar basalts (Nyquist et al. 1995) show small £]42 effects of =25 ppm (Fig 4) together with very positive £143 values of >+7 at 3.9 Ga (Fig 4B). This is consistent with crystallisation of the lunar magma ocean at =4.3 Ga with f values of from 0.25 to 0.4 (Fig 4B). The terrestrial samples analysed by McCulloch & Bennett (1993) are consistent with the lunar results of Nyquist et al. (1995), having e143 values = 1/2 that of the Moon and no measurable £142 effects >10 ppm. Harper & Jacobsen (1992) have however reported an £]42 of 33 ppm value for a metasedimentary sample from Isua, with £143 = 3.5, the latter being significantly smaller value than that found in Apollo 17 basalts. Thus at this time, the Harper & Jacobsen (1992) result must be considered anomalous with respect to both the Earth and Moon and not repre¬ sentative of the terrestrial early Archean mantle. The strongly positive e]43 values shown in Figure 4B for terrestrial samples together with the lack of £142 effects (McCulloch & Bennett 1993) provides a firm upper limit to the time of differentiation of the terrestrial mantl. These constraints require that highly fractionated LREE depleted reservoirs were formed and only preserved sometime after 4.3 Ga, by which time decay of ,46Sm was essentially complete. In the interval from =4.5 Ga, to 4.3 Ga the Earth's mantle was probably well mixed and con¬ stantly being rehomogenised, possibly due to the effects of giant impacts. The Nd isotopic results are therefore consistent with a relatively extended interval (=100 Ma) for the formation of the Earth. Discussion and Conclusions The interpretation of AT's determined using the Pb-Pb and Rb-Sr chronologies is not straightforward. Essen¬ tially, the Pb-Pb system records the integrated history of core formation while Rb-Sr records the volatile/refrac¬ tory fractionation with neither system discriminating be¬ tween various rates of core formation or accretion. Thus the AT inferred for the Earth of =100 Ma does not neces¬ sarily imply that the Earth accreted continuously throughout this interval. For example, the Earth may have formed essentially catastrophically in a period of <107 vrs, but =80 to 100 Ma following the differentiation of ADOR. Alternatively, if accretion and core formation had been essentially continuous throughout this period, then the formation interval would have been signifi¬ cantly longer than the mean ages given by the Pb-Pb and Rb-Sr isotopic systems. In the latter case, the mean age for accretion and core formation of =4.49 Ga, implies a formation interval of =140 Ma (Fig 5). Thus a formation interval for the Earth of =100 Ma is a relatively conser¬ vative lower estimate. What are the implications of a relatively prolonged formation interval for the Earth on the origin of the Moon? Assuming that the Earth and Moon were derived from similar precursor materials, then to account for the Moon's lower initial Sr ratio (Alibert et al. 1994), it is required to have formed first, probably via a giant im¬ pact on the proto-Earth, within =20 Ma of ADOR. The strong terrestrial geochemical signature of the Moon (Ringwood 1986) would have been inherited from the proto-Earth at this time, indicating that the Earth's core had started to form. Immediately following the giant im¬ pact, both the proto-Earth and proto-Moon are molten, but if large-scale differentiation occurs on the Earth, its effects were shortlived due to ongoing accretion and thorough mixing. From =4.54 Ga to ==4.48 Ga, the Earth was rehomogenised while continuing to accrete the re¬ maining —30% of its final mass. Core formation prob¬ ably occurred concomitantly with accretion rather than being triggered by a late-stage giant impact. Both the terrestrial Sr, Nd and Pb-Pb isotopic systematics are thus consistent with an extended (=100 Ma) interval for the accretion of the Earth. 137 Journal of the Royal Society of Western Australia, 79(1), March 1996 Final accretion and core formation EARTH Crystallisation of lunar magma ocean ('"A MOON ^ Partially molten accreting Earth Partially molten Moon ~ Giant (?) impact molten proto-Earth Impactor Growth of terrestrial planetismals Proto-Moon (molten) t # Formation of meteorites at 4.56 Ga I AT=8 Ma Formation of Solar System Earth formation 100 million years - ► to 4.40 Ga to 4.3 Ga Accretion of the Earth-volatile loss (Rb-Sr) jHH Melting and rehomogenisation of the Earth. No record of planetary scale differentiation (i46Sm-i42Nd) 4.42 4.44 4.46 4.48 4.50 4.52 4.54 4.56 TIME (Ga) Figure 5. Schematic diagram illustrating the sequence of events during accretion and early differentiation of the Earth-Moon system. Meteorites first formed at 4.563 Ga with an =8 Ma range of ages (Allegre ct al 1995). The mean age for core formation and accretion of the Earth is 4.49 ± 0.02 Ga and 4.48 ± 0.04 Ga respectively (solid shading). Accretion and core formation may have continued until M.42 Ga (open boxes) consistent with the Earth forming over at least an M00 Ma interval. Constraints from e142-e143 values indicate that the differentiation of the mantle into distinctive long-lived reservoirs did not commence until after ‘ M.3 Ga, consistent with rehomogenisation of the Earth's mantle during the first *»200-300 Ma of its history. Acknowledgements: Professor John deLaeter's integrity, generosity and above all enthusiasm for doing science at the highest level has provided a superlative example for all his students to emulate. I'm indebted to John for my initiation into isotope geochemistry References Alibert C, Norman M & McCulloch M T 1994 An ancient Sm- Nd age for a ferroan noritic anorthosite clast from lunar brec¬ cia 67016. Geochimica et Cosmochimica 58:2921-2926. Allegre C J, Manhes G & Gopel C 1995 The age of the Earth. Geochimica et Cosmochimica Acta 59:1445-1456. Appel P W U, Moorbath S & Taylor P N 1978 Least radiogenic terrestrial lead from Isua, West Greenland. Nature 272:524- 526. Bennett V C, Nutman A P & McCulloch M T 1993 Nd isotopic evidence for transient, highly depleted mantle reservoirs in the early history of the Earth. Earth Planetary Science Letters 119:299-317. Brannon J C, Podosek F A & Lugmair G W 1987 Initial 87Sr/86Sr and the Sm-Nd chronology of chondritic meteorites. Proceed¬ ings of the Lunar Planetary Science Conference 18:555-564. Compston W & Pidgeon R T 1986 Jack Hills, evidence of more very old detrital zircons in Western Australia. Nature 321:766-769. Gancarz A J & Wasserburg G J 1977 Initial Pb of the Amitsoq gneiss, West Greenland, and implications for the age of the Earth. Geochimica et Cosmochimica Acta 41:1283-1301. Harper C L & Jacobsen S B 1992 Evidence from coupled 147Sm- 143Nd and 14,>Sm-142Nd systematics for very early (4.5 Gyr) differentiation of the Earth's mantle. Nature 360:728-732. Jahn B & Shih C 1974 On the age of the Onverwacht Group, Swaziland Sequence, South Africa. Geochimica et Cosmochimica Acta 38:873-885. Lugmair G W & Galer S J G 1992 Age and isotopic relationships among the angrites Lewis Cliff 86010 and Angra dos Reis. Geochimica et Cosmochimica Acta 56:1673-1694. Machado N, Brooks C & Hart S R 1986 Determination of initial ^Sr/^Sr and l43Nd/144Nd in primary minerals from mafic 138 Journal of the Royal Society of Western Australia, 79(1), March 1996 and ultramafic rocks: Experimental procedure and implica¬ tions for isotopic characteristics of the Archean mantle under the Abitibi greeenstone belt, Canada. Geochimica et Cosmochimica Acta 50:2335-2338. McCulloch M T 1994 Primitive *7Sr/Sr>Sr from an Archean barite and conjecture on the Earth's age and origin. Earth and Plan¬ etary Science Letters 126:1-13. McCulloch M T & Bennett V C 1993 Evolution of the early Earth: constraints from 142Nd-143Nd isotopic systematics. Lithos 30:237-255. McNaughton N J, Comston W & Barley M E 1993 Constraints on the age of the Warrawoona Group, eastern Pilbara Block, Western Australia. Precambrian Research 60:69-98. Nutman A P, McGregor V R, Friend C R L, Bennett V C & Kinny P D 1995 The ltaq gneiss complex of southern West Greenland; the worlds most extensive record of early crustal evolution (3,900- 3,600 Ma). Precambrian Research: in press. Nyquist L E, Hubbard N J, Cast P W, Bansal B M, Wiesmann H ' & Jahn B 1973 Rb-Sr systematics for chemically defined Apollo 15 and 16 materials. Proceedings of the Lunar Sci¬ ence Conference 4:1823-1846. Paterson C C 1956 Age of meteorites & the Earth. Geochimica et Cosmochimica Acta 10:230-237. Pidgeon R T 1978 Big Stubby and the early history of the Earth. In: Short papers of the 4th ICOG (ed R E Zartman). U.S. Geological Survey, Open File Report 78-701, 476. Prinzhofer A, Papanastassiou D A & Wasserburg G J 1992 Sa¬ marium-neodymium evolution of meteorites. Geochimica et Cosmochimica Acta 56:797-815. Richards J R 1986 Lead isotopic signatures: further examination of comparisons between South Africa and Western Australia. Transactions of the Geological Society of South Africa 89:285- 304. Richards J R & Appel P W U 1987 Age of the "least radiogenic" galenas at Isua, West Greenland. Chemical Geology 66:181- 191. Rutherford E 1929 Origin of actinium and the age of the Earth. Nature 123:313-314. Ringwood A E 1986 Terrestrial origin of the Moon. Nature 322:323-328. Stacey ] S & Kramers J D 1974 Approximation of terrestrial lead isotope evolution by a two-stage model. Earth and Planetary Science Letters 26:207-221. Stevenson D J 1987 Origin of the Moon: the collision hypothesis. Annual Reviews of the Earth and Planetary Science 15:271- 315. Tatsumoto M, Knight R J & Allegre C J 1973 Time differences in the formation of meteorites as determined from the ratio of lead-207 to lead-206. Science 180:1278-1283. Taylor S R & Norman M D 1990 Accretion of differentiated planetismals to the Earth. In: Origin of the Earth (eds H E Newsom & J H Jones). Oxford University Press, New York, 29-43. 139 Journal of the Royal Society of Western Australia, 79:141, 1996 A tale of two cratons: Speculations on the origin of continents A F Trendall Retired Director of Geological Survey of Western Australia 29 Troy Street, Applecross WA 6153 Abstract The origin of continents is not well understood, perhaps because each of the two usual ways of address¬ ing the problem has methodological difficulties. Isotope geochronology has enabled the condition of the early Earth to be reconstructed from the direct evidence of successively older rocks, working backwards in time. But, since a point is reached where there are no older rocks available, much speculation about the immediately pre-geological history of the Earth (the period when its identity was well established but of which no direct trace remains in the form of existing surface rocks) has come from planetology; and its approach has been, properly, to work forwards in time, starting from some preferred model of accretion. This paper is an attempt to bridge these two approaches, by focusing on what characteristics of the earliest rocks actually need to be explained by the planetological models. The restriction of these oldest rocks to a small number of core areas within early cratons, of which the Kaapvaal Craton (KC) of southern Africa and the Pilbara Craton (PC) of Western Australia are taken as typical, is first noted. The closely parallel evolution of both of these Cratons starts at about the same time, with the sudden appearance of vast quantities of evolved silicic magma (now orthogneisses), accompa¬ nied by the deposition of thick successions of mixed, mainly volcanic, rocks differing in only a few respects from those that have continued to come from the Earth's interior; although the early Earth must have been hotter, this is not reflected in the metamorphism of the oldest strata. Following this abrupt beginning, these protocontinents grew steadily through continuous, but diminishing, magmatic activity, manifest as both silicic plutonism and bimodal eruption. All of these features are consistent with the following simple model of the early Earth. A completely molten immediately pre-geological Earth had, outside its metallic core, a pattern of huge, polygonal, convection cells with ascending central plumes and descending peripheries; there were possibly as few as 12 such cells, forming a pattern of irregular pentagons, with 20 main centres of convective descent (COCDs) at their triple junctions. Before the continents formed, the thin skin of cooled and solidified material, formed on the surface of each cell during radial surface flow, was carried down at the margins and remelted as it descended. Such melting produced, by well known petrogenetic processes, a variety of differentiated magmas which during rapid early convection were carried down and remixed with their parent materials. As the whole Earth cooled, and the rate of convection slowed, a point was reached where the buoyancy of these silicic magmas enabled them to rise faster than the descending columns below the COCDs, resulting in the appearance at the surface of a plug of orthogneiss. The subsequent geological development of the KC and PC, jointly taken as archetypes of the early continents, fits comfortably with a concept that later continental growth was a simple response to mainte¬ nance of the same large-scale convective system which remained in force long enough to establish, below the oldest parts of the continents, the very deep roots for which there is increasing evidence. At some stage of lateral growth, possibly associated with both a change of global convective patterns and the development of a layered mantle, the early continents became unstable and were subject to break-up and relative displacement over the Earth's surface. The existing areas of oldest rocks are not simply the few surviving remnants of some early (but not earliest) continental crust that covered the whole Earth, but really do represent both the initial and sole centres of continental growth. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 141 Journal of the Royal Society of Western Australia, 79:143-148, 1996 Neutrinos: Ghosts of creation J R de Laeter Department of Applied Physics, Curtin University of Technology, Perth WA 6001 Abstract The magnitude of the mass of an electron neutrino is still uncertain, yet this parameter is of vital importance to our understanding of cosmology, particle physics and neutrino oscillations. The process of double beta decay offers a means of testing grand unification theories as it may yield limits for the electron neutrino mass and indicate the degree of lepton conservation in elementary particle reactions. A number of measurements of the half-life of double beta decay of 130,12HTe/ mu*Xe and H2Se/H2Kr have been obtained, but the values are almost certainly inaccurate because of leakage of the gaseous daughters over geological time. This may be overcome by measuring the double beta decay half-lives of nuclides which have solid daughter products. Improvements in solid source mass spectrometry now allow such measurements to be successfully attempted. The results of one such experiment will be described and other double beta decay systems which may yield experimental results suggested. Introduction On December 4th 1930, the Austrian physicist Wolfgang Pauli suggested that a particle might be re¬ sponsible for carrying off the varying amounts of energy that seemed to be missing in beta decay. However, over two decades were to elapse before neutrinos were finally discovered by Reines & Cowan (1956). It has since been shown that there are different types of neutrinos - elec¬ tron neutrinos which are emitted during beta decay, muon neutrinos which are associated with charged pion decay, and tau neutrinos. The only way that neutrinos can interact with normal matter is through an extremely feeble force known as the weak interaction. Thus, in the "standard model of par¬ ticle physics, neutrinos are mass-less particles that do not disintegrate. There is some experimental evidence (albeit controversial) that neutrinos can oscillate from one form to another (Athanassopoulos et al. 1995). This can only occur if these neutrinos possess a mass. In 1987, super¬ nova SN 1987 A exploded in the Large Magellanic Cloud. From the spread in the arrival times of the burst of neu¬ trinos emitted by SN 1987A, it could be calculated that the mass of the electron neutrino is less than 25eV. One of the most successful developments of modern science has been the rapid evolution of nuclear astro¬ physics to what is now a mature scientific study. A study of nuclear astrophysics has enabled us to unravel many of the secrets of nucleosynthesis — those nuclear constraints which determine the abundances of the chemical elements. The isotope abundances of the ele¬ ments, which we measure so precisely by mass spectrom¬ etry today, have been fashioned in the stars and interstel¬ lar medium over billions of years. Some are vestiges of the Big Bang itself. © Royal Society of Western Australia, 1996 de Laeter Symposium on Isotope Science Curtin University of Technology, Perth, 1995 The three basic process involved in synthesising the heavy elements (Z>31) are the slow (s) and rapid (r) neu¬ tron capture processes, and the proton (p) process, a re¬ action that synthesises the neutron-deficient isotopes. The s-process synthesises nuclides along the valley of nuclear stability during the red giant phase of a star's evolution, whilst the r and p processes occur on a much shorter time scale in cataclysmic, non-equilibrium envi¬ ronments, such as supernova. De Laeter (1990) has de¬ scribed these nucleosynthetic processes in greater detail. Figure 1 shows a section of the Chart of the Nuclides in the vicinity of tin. The r-process nuclides mimic fis¬ sion product decay chains and produce neutron-rich iso¬ topes, in some cases far from the valley of stability. From the nuclear stability of these r-process even-even nu¬ clides, it can be shown that they cannot be absolutely stable, yet the fact remains that they have sustained their existence in measurable quantities over geological time. There is therefore a need for a nuclear decay mechanism to exist that is sufficiently slow to enable these nuclides to survive in significant quantities over a period of many billion years. Double beta decay is just such a mechanism in that its half-life is in excess of 1017 years (Mayer 1935). The nuclear systematics of double beta decay is applicable to even-even nuclides on the neutron-rich side of the valley of nuclear stability, thus enabling survival of these nu¬ clides over geological time scales. Double beta decay Double beta decay is the rarest nuclear phenomenon in Nature. In double beta decay, a nucleus undergoes a transmutation from one element to another such that two electrons are emitted rather than the single electron emit¬ ted, in the more commonly observed single beta decay. All of the parent and daughter isotopes involved in double beta decay are even-even nuclei. The pairing force acting between like nucleons is responsible for the 143 Journal of the Royal Society of Western Australia, 79(1), March 1996 Figure 1. A section of the Chart of the Nuclides in the vicinity of tin. The arrows represent nuclides produced by the r-process, which indicates the various modes of nucleosynthesis. increase in the binding energy of these nuclei relative to the odd-odd isotopes, thus preventing a single beta decay. The long standing question in double beta decay is whether the electron neutrino should be described by a Dirac or by a Majorana field (Primakoff & Rosen 1981). In the Dirac formalism, the neutrino has a distinct anti¬ particle such that: (A, Z) -> (A, Z + 2) + 2e- + 2ve However, if the neutrino is a Majorana particle then the neutrino and anti-neutrino are indistinguishable, so that double beta decay can occur with the net emission of no neutrinos such that; (A, Z) — > (A, Z + 2) + 2e* Grand Unified Theories of particle physics allow the neutrino to have a finite mass, thus allowing the possibil¬ ity of neutrino-less double beta decay. However, double beta decay systems have such long half lives (1017- 1024 years; Doi et al 1985), that enormous experimental con¬ straints are imposed on the detection of double beta de¬ cay daughter products. There are two approaches to measuring the half-life of a double beta decay reaction. In the direct detection method, coincidence counting is used to measure the ra¬ dioactivity at the energy value predicted from precise atomic mass measurements (Dyck et al. 1990). The disad¬ vantage of the radioactive counting method is that intrin¬ sically faster radioactivities, such as those associated with uranium or thorium decay chains, produce significant backgrounds which may mask the beta decay product. Instrumental experiments are designed to detect the elec¬ trons as they are emitted and are thus capable of not only measuring the lifetime of the double beta decay, but also the electron energy spectrum. Doi et al (1988) have summarised a number of experimental results for the half-life of double beta decay on 7hGe, 82Se, 1D0Mo, l36Xe and r,0Nd conducted by a number of research groups (Table 1). The half lives range from >6 x 1018 years to >1.4 x 1021 years. One of these direct detection experiments was con¬ ducted by the University of California, Santa Barbara and Lawrence Berkeley Livermore, 600m underground with eight 0.9kg Ge detectors. The detectors are inside a cav¬ ity formed by 10 blocks of 15 cm thick sodium iodide crystals providing an active volume with a 30 keV threshold. The sodium iodide crystals are in turn inside Table 1 The data of the half-lives of the neutrinoless double-beta decay with Majorana emission for various nuclei. (Doi 1988 and pri¬ mary references therein). Experimental group Half-life (yr) 76Ge Osaka Moscow (ITEP) Batelle-South Carolina Caltech-SIN-Neuchatel Santa Barbara-LBL >2 1020 >2 1020 6 (±1) 1020 >12 1020 >14 1020 82Se Irvine >4.4 1020 1(M,Mo Osaka Irvine >6 1018 >7.5 1018 136Xe Milano Moscow (INR) >1.6 1019 >1.0 1020 150Nd Moscow (INR) >7.0 1019 a pure Pb shield of 20 cm thickness. The detectors are kept at liquid-nitrogen temperatures under vacuum. The detectors are tuned to search for electrons emitted from the 7hGe — > 7hSe double beta decay with a 2.041 MeV summed electron energy, which is the total kinetic en¬ ergy available for this particular reaction. The second approach to measuring the half-life of a double beta decay reaction is based on the cumulative effects of double beta decay over geological time. This geochemical method depends on the mass spectrometric determination of the stable double beta decay product that has accumulated in geologically old minerals. The geochemical method does not directly determine the de¬ cay mode as is possible in the direct determination method. It indicates only the total decay probability of two neutrino decay X2v and no neutrino decay Xov X 2v + X u . However, for double beta decay accompanied by the emission of two neutrinos, the theoretically predicted de¬ cay probability X 2v is much smaller than that predicted in neutrinoless double beta decay X n by perhaps six orders of magnitude (Doi et al 1985), although the actual value depends on the mass of the neutrino. The geochemical method provides an upper limit to the double beta decay half-life because the gas retention age invariably post-dates the time of mineralisation; the direct detection method yields lower limits because of the possibility of background events. The direct detection 144 Journal of the Royal Society of Western Australia, 79(1), March 1996 method has the potential advantage of being able to dis¬ tinguish between the two double beta decay modes. In the case of the two neutrino double beta decay mode, the energy deposited in the detectors will show a continuous energy distribution up to the total transition energy, in the same manner as single beta decay. However, if neutrino-less double beta decay occurs, the sum of the energies of the two emitted electrons will be the total transition energy, and the energy spectrum will be a single narrow peak. The energy spectrum therefore pro¬ vides a dear, unambiguous signature of the nature of double beta decay. The high resolution, double focusing mass spectrometer at the University of Manitoba has been used to determine double beta decay energy values for 128Te and l30Te of 867.2 ±1.0 and 2528.8 ±1.3 keV respectively (Dyck et al. 1990). The total kinetic energy available for any double beta decay system can be ob¬ tained from the atomic mass table of Wapstra & Audi (1985). Geochemical half-life determinations of double beta decay The geochemical method of measuring double beta decay depends on the accumulation of decay products from this rare nuclear process in a natural mineral of the parent element. The method is most likely to succeed if little of the daughter element was incorporated when the mineral of the parent element formed. The likelihood of this separation occurring in nature is enhanced if the geo¬ chemical nature of the double beta decay product is radi¬ cally different from that of the parent element. Success also depends on the mineral being geologically old and having survived intact over a long period of time so that detectable quantities of the daughter have accumulated in the mineral. Hard, high temperature minerals have a favoured chance of surviving temperature/pressure variations to which the minerals may have been exposed during their geological history. Geochemical measurements of double beta decay have, until recently, been confined to the measurement of gaseous daughter products (by gas source mass spec¬ trometry), which have accumulated in old tellurium and selenium minerals from the following reactions; 130Te -> 33 -> 130Xe ± 2v e 128Te -» 33 -» ,28Xe + 2v e 82Se -» 33 -> 82Kr + 2v e The double beta decay of 130Te to 130Xe is a favourable experimental situation because 130Te is the most abun¬ dant nuclide of Te, old telluride minerals of high concen¬ tration are available, and the mass spectrometry of Xe can be carried out with high sensitivity. Inghram & Reynolds (1949) first estimated the half-life of 130Te for double beta decay to be 1.4 x 1021y. Since that time, many ,30Te decay experiments have been carried out and a partial list of the results are given in Table 2. The current best estimate of the half-life of 130Te is 8 1020y (Manuel 1991). Figure 2 (taken from Kirsten et al. 1968) shows the dominance of 1%Xe in the mass spectrum of Xe extracted from an old telluride ore sample. Similar mass spectrometric determinations have been made for the double beta decay half-life of 128Te and 82Se. Manuel (1991) gives best estimates of these half-lives as 2 1024y and 1 1020y respectively. The results from these isotopic systems support the lepton-conserving, two neutrino mode of decay. Table 2 Values reported for the half-life of 13(,Te Reference Half-life Inghram & Reynolds (1949) 1.4 1021 Takaoka & Ogata, (1966) 8.2 1020 Srinivasan et al. (1972) 2.5 1021 Hennecke et al. (1975) 9.7 1020 Kirsten (1983) 2.6 1021 Richardson et al. (1986) <1 1021 Kirsten et al (1986) 1.6 1021 Manuel (1986) 7 1020 Chiou & Manuel (1988) 7 1020 Current best estimate 8 1020 128 130 132 134 136 Mass Number A Figure 2. Isotopic composition of xenon extracted from tellu¬ rium ore (Kirsten et al. 1968). The horizontal lines indicate the maximum contribution of atmospheric xenon. The major analytical shortcoming of the Te/Xe and Se/Kr decay systems is that their daughters are gaseous and therefore may have suffered some leakage from the ore sample over the long time period since the minerals were formed. Minerals of tellurium and selenium are particularly susceptible to gas loss because they are soft, low temperature minerals that may have recrystallised and lost radiogenic Xe or Kr during any deformation event or high temperature phase subsequent to mineralisation. Richardson et al. (1986) have shown that gas retention ages are significantly less than mineralisation ages. Lin et al. (1986) discuss the prob¬ lems involved with gas retention in the mineral kitkaite. 145 Journal of the Royal Society of Western Australia, 79(1), March 1996 Although there may well be uncertainties in the abso¬ lute values of 128Te, ,30Te and 82Se, there are compelling reasons why the ratio of the half-lives of 128Te, 130Te for double beta decay should be determined. The great ad¬ vantage of the Te/Xe double beta decay system is that both the daughter products are noble gases and the ratio 128Xe/130Xe, which is directly proportional to the ratio of the half lives of 128Te/l30Te, can be determined with rela¬ tively high precision by noble gas mass spectrometry, thus eliminating many sources of systematic error. Pontecorvo (1968) has shown that the ratio of the de¬ cay rates of pairs of similar nuclei can be used to test various theoretical predictions because the ratio of their respective matrix elements should be near unity. In the case of Dirac decay, the ratio p2v = 128X2v/130X2v is much smaller than the Majorana decay ratio poi = 128Xov/130\Ov/ where X is the probability of double beta decay. Thus even a small Majorana decay contribution would cause a drastic increase in the measured ratio p^ = 128Xx/130\x above the base level of Dirac decay p2v, (where £ repre¬ sents the sum of the double beta decay mode compo¬ nents). The measurement of the half-life of the double beta- decay of 128Te is far more difficult than for 130Te because the probability of double beta decay is so much smaller. However, a recent measurement of the double beta decay half lives of l28Te and 130Te (Bematowicz et al. 1993), has determined the daughter product ratio 128Xe/130Xe with greater confidence than previous measurements listed in Table 3 (Manuel 1991). The mean value for the ratio of half-lives in Table 3 is 4 10"4, which compares favourably with the Bematowicz et al (1993) value of 3.52 (±0.11) 10"4. The data of Bematowicz et al (1993) give a limit to the effective Majorana mass of the neutrino of 1.5eV. Table 3 Values reported for the ratio of the double-beta decay rate of 128Te relative to that of 1320%) should be culled out. Jones RAC & Cowling WA 1995 Resistance to seed transmis¬ sion of cucumber mosaic virus in narrow-leafed lupins {Lupinus angustifolius). Australian Journal of Agricultural Re¬ search 46:1339-1352. Examination of the canopy arthropod communities of eucalypt trees in eastern and western Australia, deter¬ mined at 3 month intervals using chemical knock-down procedures, revealed that arthropods were more com¬ mon in eastern Australian trees and that there was a different seasonal pattern. The phenological patterns of canopy arthropods appear to be related to the condition of the host trees and/or climatic factors. Detection of variability in seasonal and annual patterns of canopy invertebrate communities by long-term sampling is re¬ quired to evaluate the impact of disturbances on forest communities. Recher HE, Majer ]D & Ganesh S 1996 Seasonality of canopy invertebrate communities in eucalypt forests of eastern and western Australia. Australian Journal of Ecology 21:64-80. Volume 4 of Monographs on Australian Lepidoptera, The Checklist of the Lepidoptera of Australia, provides the first complete documentation of this taxon, including nomenclature and classification for all of the entire named Australian lepidopterans. Generated from a com¬ puter database compiled with strict protocols, it includes over 24660 names. There is an introduction to each fam¬ ily. It records over 850 significant misspellings, over 1250 new synonymies and over 1500 new combinations. A CD-ROM of all text files is included. Nielsen ES, Edwards ED & Rangsi TV (eds) 1996 Checklist of the Lepidoptera of Australia. Vol 4, Monographs on Austra¬ lian Lepidoptera. CSIRO Publishing, East Melbourne. Resource partitioning within an assemblage of seven species of insectivorous birds, inhabiting remnant Melaleuca woodland on Rottnest island, was examined by researchers from Murdoch University. Foraging habits of birds were found to be both species-specific and associated with the month of observation, indicating that foraging partitioning occurred but the pattern varied temporally. The diversity of foraging habitats of each species varied widely in each foraging dimension, al¬ though species which were generalists in one foraging dimension tended to also be generalists in other dimen¬ sions. Wheeler AG & Calver MC 1996 Resource partitioning in an island community of insectivorous birds during winter. Emu 96:23-31. Assessment of the growth and yield of 97 seedlots of Gungurru narrow-leaf lupins, for seed size, percentage germination, cucumber mosaic virus (CMV) seed infec¬ tion, and several seed nutrients, showed that plant den¬ sity, shoot dry weight at 6 weeks, and grain yield varied significantly among seedlots. Seed size influenced stand density and shoot dry weight, but not grain yield. No seed nutrient was strongly associated with grain yield. There was no association of seedlot with the geographi¬ cal source of the seed. High germination percentage and low CMV infection were the main predictors of high grain yield, but accounted for only 40% of the variance among seedlots in yield. Tapscott HL & Cowling WA 1995 Predictors of yield of Lupinus angustifolius (cv. Gungurru) seedlots from different sources in Western Australia. Australian Journal of Experi¬ mental Agriculture 35:745-751. Evidence of the natural fire frequency to which Bank- sia hookeriana was best adapted, was sought through analysis of its seed demography. Plants begin flowering at 3-4 years, but storage of viable seeds in the canopy was rare for plants less than 5 years old. Cones con¬ tained an average of 12 follicles, holding 2 seeds; 62% of seeds were estimated to be viable. Cone production in¬ creased with age, to about 17 cones per year at an age of 15 years. Seed was lost by spontaneous follicle rupture, granivory, and decay. A seed bank computer model pre¬ dicted continued seed accumulation to at least 25 years of age, but by this time more seeds had been lost than were stored. Over many generations, the likelihood of successful recruitment occurs for an average fire fre¬ quency of 15-18 years. Recent fires have recurred within 7-11 years, indicating a human impact on the natural fire regime which may ultimately threaten Banksia hookeriana. Enright NJ, Lamont BB & Marsula R 1996 Canopy seed bank dynamics and optimum fire regime for the highly serotinous shrub, Banksia hookeriana. Journal of Ecology 84:9-17. Cooperative research between the University of West¬ ern Australia and the Western Australian Museum used allozyme and morphometric variability to provide a ge¬ netic perspective of the Indonesian fruit bat genus Cynopterus. The genetic distance between populations of 150 Journal of the Royal Society of Western Australia, 79(2), June 1996 C. nusatenggara was strongly correlated with both the contemporary sea-crossing distance between islands, and the estimated sea-crossing distance at the last Pleis¬ tocene glacial maximum. This, and the low levels of population substructure within islands, indicates that the sea is a primary and formidable barrier to gene ex¬ change. Morphometric variability did not show any of the main effects seen in the genetic data, suggesting that genetic and morphometric variability were not associ¬ ated at the level of individuals. Schmitt LH, Kitchener DJ & How RA 1995 A genetic perspec¬ tive of mammalian variation and evolution in the Indonesian Archipelago: Biogeographic correlates in the fruit bat genus Cynoptems. Evolution 49:399-412. The present environment of Australia represents a pal¬ impsest that records a history of past climates, nutrient poor soils, burning and increasing aridity, but these details of history are not easily disentangled. Present ob¬ servations cannot be readily used to interpret the past environments, and to infer these it is necessary to con¬ sider geological, climatic and weathering histories, as well as the physiological development of the structures being used as evidence. It is argued that the origins of fire-adaptation may predate the arrival of Aborigines, whose use of fire in cultural and hunting practices has only influenced the biota for about 50000 years. This is too short a time to have influenced the biota, other than to eliminate or winnow out those elements unable to cope with the new fire regimes. Main AR 1996 Ghosts of the past: Where does environmental history begin? Environment & History 2:97-114. Change in the number of red kangaroos in the pasto¬ ral zone of South Australia is related by a researcher from the University of Melbourne to rainfall and popula¬ tion size. Because of uncertainty about the responses of red kangaroo populations to rainfall, it is difficult to analyse the effects of harvesting on the kangaroos. The effects of rainfall and population size depend on the scale of observation; the orevious year's rainfall is closely related to population changes over large areas, but summer and autumn rainfall is more critical over smaller areas. McCarthy MA 1996 Red kangaroo ( Macropus rufus) dynamics: Effects of rainfall, density dependence, harvesting and environmental stochasticity. Journal of Applied Ecology 33:45-53. The authorative volumes on Protura, Collembola and Diplura (Volume 22), and Echinodermata (Volume 33) of Australia are part of the Zoological Catalogue of Austra¬ lia, a computer database of taxonomic and biological knowledge of the Australian Fauna. This 90-odd volume Catalogue, produced as a public inquiry database, serves as a comprehensive directory to recent information on each species of the Australian fauna. Information on each species includes synonomy, literature citation, loca¬ tion and status of type material and type locality, a brief summary of geographical distribution, ecological attri¬ butes, and important references including basic biology. Houston WWK & Greenslade P 1994 Protura, Collembola and Diplura. Volume 22. Zoological Catalogue of Australia. Aus¬ tralian Biological Resources Study, Canberra. Rowe FEW & Gates 1996 Echinodermata. Volume 33. Zoologi¬ cal Catalogue of Australia. Australian Biological Resources Study, Canberra. Papers from the Sixth International Theriological Congress have been published in the March issue of Vol¬ ume 21 of the Australian journal of Ecology: Oates JF 1996 Habitat alteration, hunting and the conservation of folivorous primates in African forests. Australian Journal of Ecology 21.1-9. Cork SJ 1996 Optimal digestive strategies for arboreal herbivo¬ rous mammals in contrasting forest types: Why koalas and colobines are different. Australian Journal of Ecology 21:10-20. Braithwaite LW 1996 Conservation of arboreal herbivores: The Australian scene. Australian Journal of Ecology 21:21-30. Smith AP & Ganzhorn JU 1996 Convergence in community structure and dietary adaptation in Australian possums and gliders and Malagasy lemurs. Australian Journal of Ecology 21:31-46. Physical Sciences Note from the Hon Editor: This column helps to link the various disciplines and inform others of the broad spectrum of achievements of WA scientists (or others writing about WA). Contributions to "Recent Advances in Science in Western Australia" are welcome, and may include papers that have caught your attention or that you believe may interest other scientists in Western Aus¬ tralia and abroad. They are usually papers in refereed journals, or books, chapters and reviews. Abstracts from conference proceedings will not be accepted. Please sub¬ mit either a reprint of the paper, or a short (2-3 sen¬ tences) summary of a recent paper together with a copy of the authors' names and addresses, to the Hon Editor or a member of the Publications Committee: Dr P C Withers (Hon Editor), Department of Zoology, University of Western Australia, Nedlands WA 6907 Dr S D Hopper (Life Sciences), Kings Park and Botanic Garden, West Perth WA 6004 Dr A E Cockbain (Earth Sciences), PO Box 8114, Angelo Street, South Perth WA 6151 Assoc Prof G Hefter (Physical Sciences), Mathemati¬ cal and Physical Sciences, Murdoch University, Murdoch WA 6150 Final choice of articles is at the discretion of the Hon¬ orary Editor. "Letters to the Editor" concerning scientific issues of relevance to this journal are also published, at the dis¬ cretion of the Hon Editor. Please submit a word process¬ ing disk with letters, and suggest potential reviewers or respondents to your letter. P C Withers, Honorary Editor , journal of the Royal Society of Western Australia. c/o Western Australian Museum, Francis Street, Perth WA 6000 151 journal of the Royal Society of Western Australia, 79:153-154, 1996 Installation of Professor Michael Jones as the new President of the Royal Society of Western Australia. His Excellency Major General Michael Jeffery, AC MC Governor of Western Australia. Vice-Patron of The Royal Society of Western Australia Monday 15 July, 1996 Dr Steve Hopper, Outgoing President of the Royal Society of Western Australia Professor Michael Jones, Incoming President of the Royal Society of Western Australia Distinguished Guests, Ladies and Gentlemen It is my pleasure to be here tonight to attend your Annual General Meeting and to install the Incoming President, Professor Michael Jones. The Royal Society of Western Australia has the very real potential, and indeed I would suggest the responsi¬ bility, to play a very important role in Society in influ¬ encing the community and governments of the critical importance of science and research, if Australia wishes to remain the 'Lucky Country'. In recent years, science appeared to be losing its ap¬ peal for children, and while many young people were concerned, few seemed to be aware of the role that sci¬ ence played in their daily lives. Many appeared to asso¬ ciate science with pollution and nuclear accidents, like Chernobyl, and saw scientists as underpaid and under¬ valued. Thankfully, places like Scitech and the weekend morning children's television show 'Hot Science', which present science as a wonderful, glamorous and exciting subject, and organisations like the Royal Society of West¬ ern Australia, a body of distinguished scientists and scholars whose primary objective is to promote learning and research in the natural and social sciences and in the humanities, are doing much to change this image. The Western Australia Education Department has in¬ stigated its own policy to encourage science, while sci¬ ence teachers and engineering associations hold compe¬ titions such as the annual 'Tournament of Minds' to en¬ courage and stimulate young students to discover cre¬ ative solutions to problems. For example, from February 1995 to last month, a total of 240 State Primary Schools implemented, or were in the process of implementing, the Australian Academy of Science 'Primary Investigations' Program. The suc¬ cess of the program has been due to cooperation and collaboration between the Education Department and the science education community. An external evalua¬ tion of the effectiveness of the program indicated that, in general, the program • raised the status of science in primary schools; • resulted in a whole school approach to science; • increased student interest in science; and • resulted in reluctant teachers teaching science. Its effectiveness has been recognised nationally as in¬ dicated by the following statement from the Australian Academy of Science; "By 1999 , at least 50% of Australian primary schools should have structured , whole-school science programs, with 100% involvement by 2002. One state, Western Australia, is already close to achieving the 1999 target." A science content course for science teachers has also been established. At the high school level, a Secondary School Teacher Leaders Program was established last year, to train 14 experienced secondary teachers in science education theory and its applications to classroom best practice. Each of these teachers was supported in testing out the ideas presented, sharing their work with others in the group and staff at their schools, and in initiating pro¬ grams to suit the needs of their students. Many of the teachers trained last year are continuing their training this year, and two additional programs have been established. One is for experienced teachers, and the other is for science Heads of Department. Last year, the number of students awarded grades for year 12 science courses were; Biology 2476 (14.3% of year 12 students), Chemistry 4132 (23.87%), Geology 95 (0.54%), Human Biology 5070 (29.29%), Physical Science 532 (3.07%), Physics (3132 (18.09%), and Senior Science 2052 (11.85%). In a world where the population is expected to reach 10 billion over the next few decades and where people will live to an older age, we are going to need more scientists to tackle problems such as nuclear radiation from damaged power plants, such as at Chernobyl, and leaks from sunken nuclear submarines; agricultural shortages in South East Asia, where China for example will have to annually import 370 million tonnes of grain by the year 2030, and Indonesia which was self-sufficient in rice production from 1984 to 1994 has had to start importing rice again, to make up for shortfalls caused by drought, disease and conversion of rice land to other uses; the salinity problems in the Murray/Darling Basin and our own Wheatbelt; vehicle emissions and pollution worldwide; the Greenhouse effect, marine and air pollution, land degradation, conservation of the natural 153 Journal of the Royal Society of Western Australia, 79(2), June 1996 environment, retention of biodiversity, protection of for¬ ests, over-population, water quality and energy options. There will be substantial opportunities for new environ¬ ment-friendly technologies, such as waste munchers, water purifiers and solar power. Western Australia offers wonderful opportunities for research in a range of sciences, from those that support industries like mining and petroleum, through to bota¬ nists who can explore one of the richest floras in the world, with the potential to yield new natural medicines. Australia is so different in its geological origins and history that scientists have a superb opportunity to make new discoveries. On the home front, we will need to maintain our place in competitive international markets and meet the immediate challenge of being 'APEC-ready' by 2010, when the expected free trade agreement will have maxi¬ mum impact on Australia's international performance. Australia needs to address many difficult issues in man¬ aging the rapid pace of scientific and technological change, as it confronts the forces shaping our long-term economic future. We can no longer rely on natural resources for Australia's wealth and long-term prosperity in a com¬ plex world where technologies that did not exist a few decades ago are playing important roles in our social and economic life. For example, twenty years ago many students could expect to leave school and move into a job in the clerical field while continuing their studies. The advent of electronic mail, fax machines, mobile phones, CD-ROM, the Internet, multi-media and computer networks is seeing the rapid demise of the office boy or girl, thus resulting in fewer jobs for today's school leavers. While the development, acquisition and application of knowledge through science, technology and innovation can create new sources of wealth and improve the quality of our lives, it also raises some serious social issues. For example, it has been suggested that partly due to technology, of our 757,100 Australians out of work as of last month (74,200 in Western Australia), many may never work again. They simply will not be employable, and nor might their children. Is this going to be an irreconcilable fact of life, with its serious social consequences for the foreseeable future, or can we through genuine intellectual endeavour meet the technology challenges of a modern society with a corresponding social conscience? It will, I think, be important for the scientist of the future to have a well- developed social conscience. In all its work. The Royal Society of Western Australia aims to bring together an informed and scholarly approach to scientific and technological questions. Its work aims to improve the links between the science and technology community and the education sector from pre-school to tertiary levels. The Royal Society of Western Australia contributes to the advancement of science for the good of Western Aus¬ tralia by; • providing a forum for discussion of scientific and technological issues relevant to the community, • bridging the communication gap between scientific disciplines, and • bringing significant scientific and technological is¬ sues to the attention of government and other deci¬ sion makers. I believe that The Royal Society has the potential to utilise its wealth of talent to influence decision makers toward a more dedicated approach towards science edu¬ cation. The Royal Society of Western Australia has been very fortunate in having Dr Steve Hopper as President for the past year. Dr Hopper is Director and Chief Executive Officer of Kings Park and Botanic Garden. After 14 years as a research scientist working on the conservation of Western Australian flora, Dr Hopper moved into re¬ search administration as officer-in-charge of the Western Australian Wildlife Centre, before his appointment at Kings Park and Botanic Gardens in 1992. He has devel¬ oped specialist expertise in the fields of rare and endan¬ gered plants, urban bushland conservation, and the biol¬ ogy of eucalypts, orchids and kangaroo paws. In 1990, he travelled to the United States of America on a Full- bright Senior Scholar's Award, and served as a Miller Visiting Research Professor at the University of Califor¬ nia, Berkeley. At Kings Park and Botanic Gardens, Dr Hopper has led strategic planning and new works aimed at delivering world-class services and facilities within 10 years. As Vice-Patron of The Royal Society of Western Australia, I would like to extend my thanks on behalf of The Society to Dr Hopper for his dedication to The Soci¬ ety, especially in the past year as President. Professor Michael Jones attended Cambridge Univer¬ sity, graduating in Natural Sciences (Biochemistry) fol¬ lowed by a Ph. D. in plant biochemistry. This was fol¬ lowed by postdoctoral research fellowships at the Uni¬ versity of Missouri (Plant Pathology), the Australian Na¬ tional University (Developmental Biology) and then back to Cambridge University (Biochemistry). He was then appointed senior scientific officer at the Welsh Plant Breeding Statiuon (1979-1982) and principle Scientific Of¬ ficer at Rothamsted Experimental Station, United King¬ dom (1982-1990). In November 1990, Professor Jones was appointed to Murdoch University and became Head of the Plant Sciences discipline within the School of Biological and Environmental Sciences. In 1994, he was appointed Director of the Western Australia State Agricultural Biotechnology Centre, and from this month Professor Jones has been released from his duties within Plant Science to act as full-time director of the Western Australian State Agricultural Biotechnology Centre. I ex¬ tend my best wishes to Professor Jones for his term as President. As Vice-Patron of The Royal Society of West¬ ern Australia, I very much look forward to furthering my association with The Society and its new president in the coming year. It is now my very agreeable duty to call upon you. Professor Michael Jones, to be installed as the President of The Royal Society of Western Australia. As the Presi¬ dent, you are a member of the Council and you preside over all meetings of The Council and Society. I charge you to carry out diligently and faithfully your responsi¬ bilities, and to foster the study of science for the good of the global community. Professor Michael Jones, it gives me great pleasure to formally install you as the President of The Royal Society of Western Australia. 154 Journal of the Royal Society of Western Australia, 79:155-160, 1996 The impact of vegetated buffer zones on water and nutrient flow into Lake Clifton, Western Australia P M Davies1 & J A K Lane2 1 Department of Zoology, The University of Western Australia Nedlands, WA 6907 2 Department of Conservation and Land Management Wildlife Research Centre, PO Box 51 Wanneroo, WA 6065 Manuscript received October 1995; accepted March 1996 Abstract Lake Clifton, a saline lake situated on the Swan Coastal Plain, south of Perth, is recognised under the Ramsar Convention as a "Wetland of International Importance". Intensification of rural and urban developments along the eastern catchment of the lake have been implicated in increased nutrient levels and consequent algal growth in the water. This has adversely affected a highly- restricted microbialite population which is the largest in the southern hemisphere. Existing buffer zones of native vegetation separate the cultural developments from the lake, and have the potential to reduce impacts both by the uptake of surface-water flow and the assimilation of nutrients. These buffer zones can be divided into three categories based on width; small (20-50 m), medium (100- 150 m), and large (500-600 m). Surface-wrater flow into Lake Clifton from the buffer zones was highly episodic, occurring predominantly during and immediately after rainfall events of >10 mm rainfall over a 48 hour period. During six substantial rainfall events (May to September 1993), surface flow into the lake (located using a low-flying aircraft) was measured for rates of instantaneous water discharge and water samples were taken for nutrient analyses. Surface water discharge from small buffer zones was about 100 times higher and total nitrogen content was significantly greater compared with the other buffer zones. Although not statistically different, mean levels of total phosphorus ranged from 0.16, 0.04 and 0.07 mg L1 from the small, medium and large zones respectively. Our results suggest that the existing buffer zones, particularly those classified as small, are inadequate to limit nutrient input into Lake Clifton. The inadequacy of these buffer zones is mitigated, in part, by the small volumes of water flowing into the lake via surface discharge. However, unless wide buffer zones are provided, further development in the catchment will increase the rate of eutrophication of Lake Clifton. Introduction Lake Clifton is part of the Peel-Yalgorup system on the western edge of the Swan Coastal Plain, approxi¬ mately 100 km south of Perth. The environmental signifi¬ cance of this Ramsar-listed lake is based, in part, on an extensive (400 ha) living microbialite (thrombolite) community that is the largest known in the southern hemisphere (Moore et al. 1983; CALM 1990). The pres¬ ence of microbialites along the eastern shore of the Lake has been attributed to constant discharge of low salinity, highly alkaline groundwater (Moore et al 1983; Moore 1987; Moore & Turner 1988). Lake Clifton lies within Yalgorup National Park, but only the lake basin and a very narrow margin of land are protected. Much of the land on the eastern catchment is privately owned and used for livestock and horticul¬ tural purposes (Moore & Turner 1988), and more re¬ cently for urban developments. There has been consider¬ able concern expressed that survival of the microbialite community may be threatened by increasing nutrient inputs from these cultural developments, and that buffer zones may not be substantial enough to reduce nutrient © Royal Society of Western Australia 1996 movement into the lake. Excessive nutrient inputs have already been implicated in increased growth of a green alga ( Cladophora vagabunda) which, along with epiphytic algae and plankton blooms, smothers and spatially com¬ petes with microbialites (Moore 1990). There are a number of mechanisms whereby nutri¬ ents enter the lake, the two major ones being overland flow and groundwater discharge. Overland flow is epi¬ sodic, occurring after substantial rainfall events, while groundwater input is relatively more constant. Prelimi¬ nary results have shown that the groundwater typically has a lower total nitrogen concentration than Lake Clifton water (Moore & Turner 1988). However, surface water has not yet been monitored for rates of discharge and associated nutrient levels. Surface flow may be in¬ tercepted by vegetated buffer zones, reducing the im¬ pact of land practices on a wetland. Buffer zones are generally described as areas of unde¬ veloped, vegetated land extending from the banks or high water level of a wetland to some landward point (Palfrey & Bradley 1990); the function of a buffer zone is to protect a wetland from negative impacts of catchment land-uses (Carter 1992). Buffer zones around wetlands are valuable as biological filters of both sediment and nutrients. These zones can reduce the extent of channelised surface-flow into a lake and have an inher- 155 Journal of the Royal Society of Western Australia, 79(2), June 1996 Figure 1. The study sites along the eastern edge of Lake Clifton. Sites as follows: 2,5,8 (large [L]), 1,3,7 (medium [M]) and 4,6,9 (small [S]). 156 Journal of the Royal Society of Western Australia, 79(2), June 1996 ent capacity to assimilate nutrients and sediment (Davies & Lane 1995). Buffer zones of native vegetation, of vary¬ ing widths, have been created along the eastern edge of Lake Clifton around rural land-use areas and, recently, urban sub-divisions. This presented an opportunity to compare the effectiveness of buffer zones of different widths in reducing both water discharge and nutrient inputs into Lake Clifton. Methods Site description Lake Clifton is an elongate (approximately 28 km x 1 km), shallow and saline lake about 100 km south of Perth on the Swan Coastal Plain (Fig 1). The extensive shoreline means there is considerable potential for nutri¬ ent enrichment from adjacent land practises. Sampling program There are no major drains flowing into Lake Clifton (Moore et al. 1983). However, previous observations re¬ vealed that surface water becomes channelised and flows into the lake after substantial rainfall events. Substantial rainfall events were defined by analysing historical rainfall records for Mandurah (Bureau of Meteorology, Perth). On the basis of this analysis, events of >10 mm rainfall over a 48 hour period were determined to be "substantial". Sampling was then conducted on six of these events (Table 1), which accounted for almost 20% of the 1993 total rainfall. During each sampling occasion, the eastern edge of the lake was observed from a low-flying aircraft to deter¬ mine areas where surface water was channelised and flowed into the lake. Channelised surface flow was de¬ fined as stream-flow originating within each buffer zone and flowing into the lake. Channelised flow did not originate from outside the boundary of any buffer zone. Existing buffer zones along the eastern edge of the lake, from the airstrip in the north (site 1) to Location 52 in the south (site 9) (approximately 10 km), were placed into three width categories (Fig 1); small, 20-50 m; medium, 100-150 m; and large, 500-600 m wide. A total of nine buffer zones were investigated, with three from each of the above categories. The zones were approximately the same length (see Fig 1). These buffer zones were interspersed, minimising the effects of spatial Table 1 Sampling dates and associated 48 hour rainfall data for Yalgorup National Park during 1993. SAMPLING DATES ACRONYM RAINFALL (mm over 48 h) 28 May 1993 MAY 28 29 June 1993 JUN 19 28 July 1993 JUL 34 25 August 1993 AUG 12 7 September 1993 SEP 17 30 September 1993 OCT 12 arrangement confounding the subsequent statistical comparisons. Additionally, there were no land-use areas associated with only a single buffer category, which could also have confounded the results. Flow from these zones accounted for the total visible surface flow into the lake. The extreme south-east corner was inaccessible to sampling; however, surface water flow from this area into the lake appeared to be minimal. Instantaneous channelised discharge from each site was estimated from the rate of water flow and its width and depth. The rate of flow was measured by a Marsh- McBirney 201 M model field water-flow meter. Where the water was not of sufficient depth for the meter to operate, discharge was estimated by the time taken to fill a plastic 1 litre measuring cylinder. Water samples were collected from each site for the analyses of total nitrogen, ammonia, nitrite and nitrate, and total phosphorus and orthophosphate. Unfiltered water samples were collected into sterile 250 ml plastic bottles for measurement of total nitrogen and total phos¬ phorus. Water samples filtered through a 0.22 pm millipore filter into a 100 ml sterile plastic bottle were collected for the measurement of ammonium, nitrate and orthophosphate. Due to costs of each analyses, ammo¬ nium, nitrate and orthophosphate were not measured from all sites, with analysis restricted to one sample from each buffer-width category on each occasion. Within each buffer zone where channelised flow was evident from more than one area, nutrient sampling was conducted from a randomly-determined location. However, total discharge was estimated by measuring flow from all channelised flow within each buffer zone. Samples of lake water for nutrient analyses were taken adjacent to one of the large buffer zones (site 8). All samples were kept on ice prior to delivery to the Chemistry Centre of Western Australia for analyses. Data analyses All data were plotted prior to parametric analyses to check for normality, and bivariate plots were examined to determine possible non-linear relationships between variables. Statistical comparisons between means were made by Analysis of Variance (ANOVA); Cochran's C was used to test for homogeneity of variances and ap¬ propriate transformations were made for biased data, e.g. log(x), or log(x+l) if data contained zero values. After ANOVA, if there was a significant difference, a Tukey's test was used to find which factors or sites were different. Tukey's tests have been recommended for post hoc significance testing as a method using the correct experiment-wise error rate (Day & Quinn 1989). Results Lake Clifton is within a region of Western Australia typified by a mediterranean climate (Seddon 1972) where rainfall occurs predominantly from late autumn to early spring. Rainfall in 1993 for Yalgorup National Park (Fig 2) followed this pattern although storms during early autumn and late spring delivered substantial rainfall. The total rainfall in 1993 was 688 mm, which is substantially less than the average 1982-1992 annual rain¬ fall of 780 mm. 157 Journal of the Royal Society of Western Australia, 79(2), June 1996 200 r MONTHLY RAINFALL (mm) FOR YALGORUP NATIONAL PARK MONTH Figure 2. Monthly rainfall for Yalgorup National Park (which incorporates Lake Clifton) during 1993. Table 2 Mean (SEM) levels of total nitrogen (mg L ’) from the three different buffer sizes and from Lake Clifton during the six sampling periods. Sampling size for the small and medium buffers, n = 3 on each occasion. One sample was collected from the lake and from a large buffer zone (site 8) on each occasion. MONTH SMALL MEDIUM LARGE LAKE MAY 31.56 (26.67) 4.17 (1.24) 11.00 1.7 JUN 4.50 (0.40) 4.43 (1.58) 20.00 1.1 JUL 5.90 (1.10) 5.53 (1.01) 0.48 1.3 AUG 13.50 (7.99) 3.80 (1.45) 0.27 1.1 SEP 13.43 (9.76) 0.90 (0.15) 1.50 1.0 OCT 3.17 (0.91) 1.28 (0.44) 4.80 1.5 Surface runoff Rates of discharge from the small buffer zones were about four times those of the medium-sized buffer zones and about 100 times those of the large buffer zones (Fig 3). Runoff from the large buffer zones was negligible, with the exception of site 8 (see Fig 1) which was subse¬ quently the only large buffer zone sampled. This area had been burnt prior to the May sampling period, re¬ moving the majority of the understorey. This may have contributed to the extent of the measured surface runoff. Figure 3. Mean (±SEM) surface runoff (in kilolitres per day) from the three different-sized buffer zones during each sampling oc¬ casion. Sampling sizes for the small and medium buffers, n = 3 on each occasion. A single sample was taken from the large buffer zones (site 8). The measured surface water runoff contributed little to the total lake water levels. Determining the area of Lake Clifton from aerial photographs, and assuming an annual mean depth of 1 m, enabled the estimation of the mean water volume of the lake. Assuming that the mean surface flow measured during the substantial rainfall events was constant for six months of the year, this flow would contribute only about 0.005% to the total lake volume. This emphasises the importance of sources of water inputs other than surface flow for the maintenance of lake water levels. Nitrogen concentration in surface runoff In surface runoff, total nitrogen concentration ranged from low values of 0.59 mg L'1 from site 3 (a medium¬ sized buffer) during September to a high value of 85.0 mg L'1 from site 9 (a small-sized buffer) during May (Table 2). Total nitrogen values of 11.0 and 20.0 mg L'1 were recorded in May and June from site 8 (a large-sized buffer) which was burnt in May. Concentrations at site 8 were lower in subsequent months, presumably because the regenerating understorey was taking up water and nutrients and the release of nutrients due to the fire had diminished. Mean values of total nitrogen from the three differ¬ ent-sized buffer zones and the lake are shown in Table 2. On most occasions, total nitrogen concentration was substantially greater in the flowT from the buffer zones compared with the lake water. Statistical analysis showed the total nitrogen concentration measured in wa¬ ter from the small buffer zones was significantly greater (ANOVA, P<0.05) than the medium and large buffer zones (Table 3). Total phosphorus level was low, with highest concen¬ trations measured from small buffer zones during September. The concentration of total phosphorus in the lake was 0.01 mg L 1 on all occasions with the exception of August and September, when the values were less than detection limits (Table 4). The mean values of total phosphorus from the small buffer zones were about three to four times levels measured in the other buffer zones. There were no statistically significant differences (Table 5). Table 3 One-way ANOVA for total nitrogen concentrations for water originating from the three different-sized buffer zones. Data were transformed by a log function, which was the closest approximation (cf untransformed, square-root) to homoscedasticity, Cochran's C=0.606. Source df MS F P Buffer size 2 4.54 3.73 <0.05 Residual 39 1.22 Results of Tukey's tests on transformed data. Untransformed means (total [N] in mg L'1) are presented in parentheses. Buffer Size Small > Medium = Large (12.07) (3.18) (6.34) 158 Journal of the Royal Society of Western Australia, 79(2), June 1996 Table 4 Mean (SEM) levels of total phosphorus (in mg L1) from the three different buffer sizes and from Lake Clifton during the six sampling periods. Sampling size for the small and medium buffers, n = 3 on each occasion. One sample was collected from the lake and from a large buffer zone (site 8) on each occasion. MONTH SMALL MEDIUM LARGE LAKE MAY 0.03 (0.01) 0.02 (0.00) 0.02 0.01 JUN 0.06 (0.02) 0.02 (0.01) 0.16 0.01 JUL 0.06 (0.04) 0.11 (0.05) 0.04 0.01 AUG 0.22 (0.10) 0.06 (0.02) 0.01 <0.01 SEP 0.52 (0.44) 0.02 (0.01) 0.06 <0.01 OCT 0.07 (0.05) 0.03 (0.01) 0.15 0.01 Table 5 One-way ANOVA for total phosphorus concentrations for water originating from the three different-sized buffer zones. Data were transformed by a log function, which was the closest approximation ( cf untransformed, square -root) to homoscedasticity, Cochran's C=0.436. Source df MS F P Buffer size 2 2.46 2.27 NS Residual 39 1.08 Extractable nutrient components Analyses of extractable nutrients showed nitrate and ammonia concentrations were always a small component of total nitrogen concentrations (Table 6). In all measure¬ ments, orthophosphate was below detection limits. Concentrations of each nutrient in the lake and in surface flows were similar except for nitrate in July and October when values in the surface flows were 5.5 and 7.5 times those of the lake water respectively. Table 6 Levels of extractable nutrients (in mg L1) from the lake and from surface flows during each sampling occasion. Note, values for surface flows are means (SEM), n = 3 and lake values are from single samples taken on each occasion. LAKE SURFACE FLOWS no3 NH/ PO / 4 NO, nh4+ po43 MAY 0.10 0.18 <0.01 0.11 (0.03) 0.53 (0.34) <0.01 JUN 0.02 0.10 <0.01 0.01 (0.01) 0.02 (0.02) <0.01 JUL 0.02 0.02 <0.01 0.11 (0.01) 0.05 (0.03) <0.01 AUG 0.04 0.05 <0.01 0.04 (0.00) 0.05 (0.00) <0.01 SEP 0.07 0.02 <0.01 0.05 (0.02) 0.03 (0.01) <0.01 OCT 0.04 0.03 <0.01 0.29 (0.14) 0.02 (0.00) <0.01 Discussion Our results showed the inadequacy of the existing small buffer zones on the eastern catchment of Lake Clifton to control nutrient inputs. Both surface water discharge and the nutrient concentration of the water from the small buffer zones were substantially greater than from the other zones measured. The inadequacy of the small zones and the impact of surface flow into the lake is mitigated, in part, by the low volumes of water flowing from the buffer zones. Measured surface dis¬ charge from all nine buffer zone sites accounts for a small amount of the total volume of Lake Clifton. How¬ ever, total rainfall in Yalgorup National Park during 1993 was significantly less than average. During wetter years, surface flow is expected to make a greater contri¬ bution to the water level of the lake. A previous fire in a large buffer zone probably re¬ sulted in substantial nutrient inputs into the lake. This declined over the next few months, presumably as re¬ vegetation of the understorey buffer zone increased, re¬ ducing both water and associated nutrient levels. The concentrations of phosphorus and nitrogen in streams in northwestern United States increased after a wildfire from 5 to 60 fold over background levels (Spencer & Hauer 1991). Fortunately, fires are probably relatively infrequent in buffer zones around most wetlands of the Swan Coastal Plain, although management practices need to recognise the effects of nutrient release after fires around both lakes and streams. The total nitrogen content of surface runoff from the small buffer zones was significantly greater than for the other buffer zones. Similarly, total phosphorus concen¬ tration in the smaller buffers was about 4 times the con¬ centration from the other buffers. Nutrient concentra¬ tions in surface water were high compared with ground- water levels (e.g. Moore & Turner 1988). The mean nu¬ trient concentrations in pore water were: total phospho¬ rus 0.03 mg L1; total nitrogen 0.05 mg L*1; ammonium 0.05 mg L 1 and nitrate 0.02 mg L*1 (Moore & Turner 1988). Concentrations of total phosphorus and nitrate in pore water were similar to the lowest values recorded in surface water while total nitrogen was two orders of magnitude lower in groundwater than surface water. Many of the wetlands of the Swan Coastal Plain re¬ ceive high levels of nutrients from their surrounding catchments and are becoming increasingly eutrophic. The following factors contribute to the increased rates of eutrophication (Birch 1984; Yeates et al. 1985; Schofield et al 1985; Sanders et al 1988; Humphries & Bott 1988): limited nutrient-holding capacity of the sandy Coastal Plain soils; a high and seasonal rainfall; the shallow wa¬ ter-table; extensive drainage of waterlogged soils; the over-use of fertilisers; inadequate treatment of effluent and inappropriate drains into wetlands. Nutrient enrichment of wetlands may be reduced by adequate buffer zones; however, in the absence of scien¬ tific information, arguments for larger buffer zones have had little influence (Lane 1991). This study provides the first empirical data on the capacity of buffer zones to reduce nutrient input into wetlands of the Swan Coastal Plain and complements research elsewhere ( e.g . Doyle et al. 1977; Overcash et al 1981). Based on the results from nutrient enrichment of Lake Clifton, an adequate buffer zone would be considered to be greater than the "me¬ dium" buffer width (100-150 m) category of this study. Davies & Lane (1995) recommend a buffer zone of at least 200 m to minimise nutrient enrichment of wetlands 159 Journal of the Royal Society of Western Australia, 79(2), June 1996 on sandy soils. That recommended width is supported by the results of this study. Many pools formed in depressions on the shoreline became continuous with the main body of the lake wa¬ ter during late winter. Some of these pools, particularly those adjacent to the rural lots with small buffer zones, would result in the input of nutrients into the lake. Fu¬ ture research should measure the nutrient levels in these pools and determine the contribution they make to the total lake nutrient concentrations. At this stage. Lake Clifton could be classified, on the basis of trophic status ( sensu Wetzel 1975), as meso- eutrophic. As the surface run-off from the small buffer regions was eutrophic to hyper-eutrophic, future devel¬ opments need to provide wider buffers to reduce the rate of eutrophication of the lake. Acknowledgments: We thank the Water Authority of Western Australia (J Kite and L Moore); the Department of Conservation and Land Manage¬ ment (S Dutton); the Royal Aero Club of Western Australia; the Chemistry Centre of Western Australia (R Schulz); and particularly for field assistance, I Craig. Financial support from the Land and Water Resources Research and Development Corporation through the National Riparian Program is acknowledged. References Birch P B 1984 Catchment management to reduce phosphorus discharge into the estuary- setting the scene. In: Potential for Management of the Peel-Harvey Estuary. Department of Conservation and Environment, Perth. Bulletin 160, 33-42. CALM 1990 Wetlands nominated by the Government of Western Australia for inclusion on the list of Wetlands of International Importance (Ramsar Convention). Department of Conservation and Land Management, Perth. Carter R 1992 Management and policy perspectives. In: The Role of Buffer Strips in the Management of Waterway Pollution from Diffuse Urban and Rural Sources (eds Woodfull J, Finlayson B & McMahon T). Proceedings of a Workshop, University of Melbourne. Land and Water Resources Research Development Corporation and Centre for Environmental Applied Hydrology, Melbourne, 103-106. Day R W & Quinn G P 1989 Comparisons of treatments after an analysis of variance in ecology. Ecological Monographs 59: 433-463. Davies P M & Lane J A K 1995 Effective buffers for wetlands. Wetlands Australia 2: 9-10. Doyle R C, Stanton G C & Wolf D C 1977 Effectiveness of forest and grass buffer filters in improving the water quality of manure polluted runoff. ASAE Paper 77-2501. Humphries R & Bott G 1988 Intensive animal industries on the Swan Coastal Plain and their associated pollution problems. In: The Swan Coastal Plain in Crisis: Agriculture and the En¬ vironment. The Australian Institute of Agriculture Science, Perth. Occasional Publication 10:59-66. Lane J 1991 The wise use of wetlands-managing wildlife habitat. In: Educating and Managing for Wetlands Conservation (eds Donohue R & Phillips B). Proceedings of the Wetland Conser¬ vation and Management Workshop, Newcastle, NSW. Aus¬ tralian National Parks and Wildlife Service, Canberra, 151- 155. Moore L S 1987 Water chemistry of the coastal saline lakes of the Clifton-Preston lakeland system, South-western Australia, and its influence on stromatolite formation. Australian Jour¬ nal of Marine and Freshwater Research 38: 647-660. Moore L 1990 Lake Clifton - an internationally significant wet¬ land in need of management. Land and Water Research News 8: 37-41. Moore L S & Turner J V 1988 Stable isotopic, hydrogeochemical and nutrient aspects of lake-groundwater relations at Lake Clifton. In: Proceedings of the Swan Coastal Plain Groundwa¬ ter Management Conference (ed G Lowe). Western Austra¬ lian Water Resources Council Publication 1/89, Perth, 201- 213. Moore L, Knott B & Stanley N 1983 The stromatolites of Lake Clifton, Western Australia. Search 14:309-314. Overcash M R, Bingham S C & Westerman P W 1981 Predicting runoff pollutant in buffer zones adjacent to land treatment sites. Transactions of the American Society of Agricultural En¬ gineers Volume 14:430-435. Palfrey R & Bradley E 1990 The Buffer Area Study. Maryland Department of Natural Resources, Tidewater Administration, Maryland, USA. Sanders C, Robinson S, McAlpine K & Bott G 1988 Environmen¬ tal objectives: what are the criteria and how can we achieve change? In: The Swan Coastal Plain in Crisis: Agriculture and the Environment. The Australian Institute of Agriculture Sci¬ ence. Occasional Publication 10: 95-105. Schofield N J, Bettenay E, McAlpine K W, Height M I, Ritchie G S P & Birch P B 1985 Water and phosphorus transport processes in permeable grey sands at Talbot's Site near Harvey, West¬ ern Australia. Department of Conservation and Environment, Perth. Bulletin 209. Seddon G 1972. Sense of Place. University of Western Australia Press, Perth. Spencer C N & Hauer F R 1991 Phosphorus and nitrogen dynam¬ ics in streams during a wildfire. Journal of the North Ameri¬ can Benthological Society 10:24-30. Wetzel R G 1975 Limnology. Saunders, Philadelphia. Yeates J S, Arkell P T, Russell W K, Deeley D M, Peek C & Allen D 1985 Management of agricultural losses from the soils of the Peel-Harvey catchment. In: Peel-Harvey Estuarine System Study Management of the Estuary. Proceedings of a symposium at The University of Western Australia, February 19-20, 1985. Department of Conservation and Environment, Perth, Bulletin 195: 59-76. 160 Journal of the Royal Society of Western Australia, 79:161-164, 1996 A new species of Lerista (Lacertilia: Scincidae) from Western Australia, Lerista eupoda L A Smith Western Australian Museum, Francis Street, Perth WA 6000 Manuscript received September 1995; accepted March 1996 Abstract The new species of skink, Lerista eupoda , described here is most like Lerista gerrardii, from which it differs in having more digits (two fingers and three toes rather than one finger and two toes) and more supracilaries (usually five rather than four). Lerista eupoda is confined to the semi- arid interior of Western Australia. Introduction The late GM Storr's last seven taxonomic papers dealt with the speciose skink genus Lerista. Two of these pa¬ pers reviewed the macropisthopus species-group (Storr 1991a, 1991b). Together, they revised all of the taxa as¬ signed to the group except Lerista gerrardii, which was presumably omitted because the accession of many more specimens since the previous revision (Storr 1972) had done little to change the concept of gerrardii. The new species described here was brought to Storr's attention by G Harold who collected three specimens near Cue, Western Australia. The subsequent collection of more specimens, which are all consistent in their digital formula, indicates they represent an undescribed taxon. None of the ten species of Lerista already described from Western Australia, that have a digital formula of 2+3, approach the new species in terms of their gross morphology or colour pattern. Lerista borealis, L. bunglebungle and L. zvalkeri are endemic to the Kimberley region, and have dorsal patterns reduced to rows of dots; L. planiventralis is a highly specialised, sharp-snouted burrower with an acute ventrolateral flange, paravertebral rows of dots and a dorsolateral stripe; L. lineata has a fixed eyelid, paravertebral lines, a dorsolat¬ eral stripe and only 16 midbody scale rows; and L. allochira is a small (up to 37mm SVL), almost patternless species, allied to L. muelleri. The remaining species, Lerista axillaris, L. desertorum, L. puncticauda and L. macropisthopus occur in the arid and semi-arid regions of southern Western Australia. None of these species has a solid vertebral stripe like the new species; in fact, L macropisthopus, the only species that has a digital formula of 2+3 and is sympatric with the new species, is patternless. The new species is compared with Lerista gerrardii. Although L. gerrardii has fewer fingers and toes (1+2, sometimes 0+2), their dorsal patterns are identical That the new species has a distribution that is almost sympat¬ ric with L. gerrardii suggests that it is a new species rather than a subspecies. Notation used by Storr (1984, 1991a) to score supraciliary fusions is adopted here to describe certain © Royal Society of Western Australia 1996 conditions in L. eupoda and L. gerrardii. The various de¬ grees of fusion are as follows; second supracilary fused with first supraocular (1+3); first and second supracilary fused with first supraocular (0+3); first supracilary fused with first supraocular (0+4); second supracilary fused with first supraocular, third and fourth supracilary fused (1+2); first and second supracilary fused to first supraocular, third and fourth supracilaries fused (0+2); second supracilary fused with first supraocular and a very small extra supracilary posterior to fused scales (1+4). All specimens examined are housed in the Western Australian Museum. Systematics Lerista eupoda sp. nov. Holotype R103943, a male in Western Australian Museum, col¬ lected by G Harold on 15 February 1990 at 14 km NNE Cue, Western Australia, at 27°19'S, 117°57,E. Paratypes North-west Division (WA); 16 km NNE Cue (103944); 35 km SSW Meekatharra (108853-54); 70 km NNE Cue (104363, 120153-54); Telegoothera Hill (87814). Diagnosis Distinguished from Lerista gerrardii, which has 1 fin¬ ger (sometimes a stump) and 2 toes, and rarely more than 4 supracilaries, by having 2 fingers, 3 toes and mostly 5 supracilaries, and from Lerista desertorum by having more supracilaries and a solid blackish-brown vertebral stripe (desertorum usually has 0+3, very rarely five supracilaries, and a paravertebral series of brownish dots). Description Snout-vent length 40-87 mm; sexed males 60-87 mm (n=5), females 56-77 mm (n=2). Length of appendages (% of SVL): foreiimb 3.4-5.0%, hindlimb 13.0-17.5%, snout to forelimb 23-25%. Eyelid moveable. Nasals narrowly separated (n=6) or in contact (n=2). Prefrontals widely separated. Frontoparietals narrowly 161 Journal of the Royal Society of Western Australia, 79(2), June 1996 separated, smaller than interparietal. Nuchals 2 or 3. Supraoculars 3, first 2 in contact with frontal. Supracilaries 5 (1+3 one specimen), second and last smallest. Upper labials 6. Three temporals, upper sec¬ ondary much the largest, lower secondary much the smallest (see Figure 1). Midbody scale rows 20. Paravertebrals; males 82-92 (n=5, mean 84.0), females 89- 91 (n=2, mean 90.0). Lamellae under longest toe 11-16. Upper surfaces brownish-white or very pale brown. Broad brownish-black vertebral stripe (more than one scale wide) from nape to base of tail, after which it breaks up into two series of angular blackish-brown spots. Broad brownish-black upper lateral stripe (nearly two scales wide) from nasal to base of tail, after which it breaks up into three series of angular blackish-brown spots or a series of narrow, curving vertical bars. Upper surfaces of limbs stippled blackish-brown. Lower sur¬ faces whitish (see Fig 2). Distribution Only known from the arid southern interior between Cue and Meekatharra (Fig 3). Habitat Open mulga on red loams and sandy loams. Comparison with other species Despite the fact that two of the other seven members of the Lerista macropisthopus species-group with two fin¬ gers and three toes (L macropisthopus macropisthopus and desertorwn) are sympatric or geographically close to the new species (Storr 1991a), it is more pertinent to compare it with L. gerrardii . Digital formulae aside, the details of colour pattern for eupoda are almost identical to gerrardii. This, together with the observation that L. gerrardii and L. eupoda are probably sympatric NW of Cue (Woolgorong Rock where L. gerrardii has been collected, is only about 20 km SW of Telegoothera Hill, the most south-westerly locality for L. eupoda) led to the comparison of eupoda with a series of 72 L. gerrardii specimens (see specimens examined). As is the case for other members of the L. macropisthopus species-group, fusion of supracilaries with supraoculars occurs in L. gerrardii. Most of the supracilary-supraocular fusions involve the first supraocular and the first and second supracilary. Fusion of the first supracilary, first two supracilaries, and the second supracilary (0+4, 0+3 and 1+3 respectively) to the first supraocular account for 93% of the variation in the supracilary series in L. gerrardii, which without fusions, would have to complete series of five supracilaries. The remaining seven percent of variation is caused by the occasional fusion of the third and fourth supracilary (0+2, 1+2) or, in rare cases, the splitting of a supracilary (1+4). Proportions of various conditions of fusion in L. gerrardii were as follows (where possible condition of both sides of each specimen scored); 1+3 (n=53); 0+3 Figure 2. Lerista eupoda specimen R104363 (photograph by B Maryan). 162 Journal of the Royal Society of Western Australia, 79(2), June 1996 Figure 3. Map showing the distribution of Lerista gerrardii (•) and Lerista eupoda (O). (n=34); 0+2 (n=3); 1+2 (n=3); 0+4 (n=3); 1+4 (n=l). The supraciliary configuration for 13 (19%) of specimens was different on either side of the head; the combinations included 1+3, 5 (5 specimens); 1+3, 1+2 (2 specimens); 0+3, 1+2 (2 specimens); 0+3, 0+2 (1 specimen); 0+3, 1+3 (1 specimen); 0+3, 1+4 (1 specimen); 1+3, 0+4 (1 speci¬ men). Snout- vent lengths of male L. gerrardii ranged from 36-80 mm (n=25, mean 70.2), females 56-91 mm (n=29, mean 76.7 mm). Paravertebral counts for males ranged from 75-102 (n=22, mean 90.5), females 83-107 (n=24, mean 92.2). The distributions of L eupoda and L. gerrardii abut and it could be argued that the distinguishing characters of L. eupoda are only part of the variation of L. gerrardii. If having extra digits was all that characterised L. eupoda then this could be the case, but the localised shift in digital formula is matched, almost exactly, in a shift in the proportion of specimens with a complete series of supracilaries (87% in L. eupoda, 5% in L. gerrardii). Fur¬ thermore, the digital and supracilary variation present in L. gerrardii is scattered and not concentrated in the northeast where it would be expected were there in¬ tergradation between the two taxa. The only specimen of L. gerrardii with two uninterrupted series of supracilaries (5+5) is from Georgina; the five specimens of gerrardii with a series of five supracilaries on one side are from widely scattered localities such as the Northampton district, Pindabunna and Toomey Hills while specimens with fingers reduced to a stump came from Northampton, Wubin, Jibberding and Southern Cross. Specimens examined North-west Division (WA): Walganna Rock (95290); Woolgerong Rock (97028-29); 15 km N Mt Magnet (88801-02); 40 km NE Paynes Find (83187, 83204-206); 30 km NNE Paynes Find (108855); 19 km NNE Paynes Find (117306); 11 km NNE Paynes Find (84145); 10 km NNE Paynes Find (91133); 10 km NNE Pindabunna (83802, 83805-06). South-west Division (WA): 8 km W Tallering Peak (115068-69); 5 km N Northampton (115866); Northampton (176, 25960, 31973-74, 66193-94, 71047, 73102-03); Naraling (119168-70); Howetharra Hill (95864); East Chapman (4430); Chapman Valley Research Station (49947-49950); Newmarracarra (1729, 3847); Geraldton (8597); 10 km S Mugga Mugga Hill (98161-98164); Moonyanooka (53693); 5 km ESE Mt Kenneth (108295); Georgina (114681); Mingenew (34103); Lochada (96628); Morowa (39160); 29 km SW Morawa (56831); Perenjori (943); 45 km NE Wubin (113567-68); 66 km NE Wubin (11004); 13 km NE Jibberding White Well (28263); Jibberding (81629); Winchester (3847); Coorow (6941, 10145); Maya (27914); 26 km NE Dalwallinu (58176); Merredin (7351). Eastern Division : 7.5 km E Yuinmery HS (66052, 66059); 13 km NE Bungalbin Hill (94485); 16 km SSW Mt Jack- son (hill) (76105); Southern Cross (34577); Yellowdine (97758); 15 km E Toomey Hills (78795, 78805); 15.5 km S Toomey Hills (71841); Toomey Hills (117372). Etymology From Greek eu (well) + pod (foot). An allusion to the extra digits (compared to L. gerrardii). Acknowledgments: 1 am grateful to Brad Maryan for the photograph of specimen R1 04363 and to RE Johnstone for assistance with figures. 163 Journal of the Royal Society of Western Australia, 79(2), June 1996 References Storr GM 1972 The genus Lerista (Lacertilia, Scincidae) in West¬ ern Australia. Journal of the Royal Society of Western Aus¬ tralia 54: 59-76. Storr GM 1984 Revision of the Lerista nichollsi complex (Lacertilia: Scincidae). Records of the Western Australian Mu¬ seum 11:109-118. Storr GM 1991a Partial revision of the Lerista macropisthopus group (Lacertilia: Scincidae). Records of the Western Austra¬ lian Museum 15: 149-161. Storr GM 1991b Revision of Lerista picturata (Lacertilia: Scincidae) of southern Australia. Records of the Western Australian Mu¬ seum 15: 529-533. 164 Journal of the Royal Society of Western Australia, 79:165-173, 1996 Biogeography of the herpetofauna of the Archipelago of the Recherche, Western Australia L A Smith & R E Johnstone Western Australian Museum, Perth WA 6000 Manuscript received September 1995; accepted April 1996 Abstract Published and unpublished herpetological data collected up until 1995 are collated for 47 islands in the Archipelago of the Recherche, off the south coast of Western Australia. The 20 species of reptile and one species of frog found on the Archipelago's islands represent only 42% of the herpetofaunal diversity of the Esperance Plain on the adjacent mainland. The six most common and widely distributed species are the geckos Phyllodactylus marmoratus marmoratus and Underwoodisaunis milii, and the skinks Ctenotus labillardieri, Egernia kingii, E. napoleonis and Hemiergis peronii. The herpetofauna of individual islands ranges from just one of these common species on the smallest islands, to 16 species (including all of the core suite of common species listed above) on the largest island. The snake, Pseudonaja affinis tanneri is the only reptile endemic to the Archipelago. A similarity index based on average linkages (UPGMA) indicates that the Archipelago's herpetofauna is most similar to that of the eastern end of the Esperance Plain. Introduction The islands of the Archipelago of the Recherche off the southern Western Australian coast (Fig 1) lie between longitudes 120°30’S and 124°10'E and up to 60km offshore in the west (Termination Island in 35°28’S, 121°59’E) and 30km offshore to the east (Daw Island in 33°51'S, 124 06E). There are about 200 islands in the Archipelago, rang ag in size from over 1000 ha (Middle Island, having most vegetation types and most plant species) to many small islands less than lOha that have little vegetation or are no more than exposed rocks. The Archipelago, with the exception of High, Station and Woody islands, was vested in the Western Australian Wildlife Authority as an A class reserve on 19 April, 1980. European knowledge of the natural history of these islands began nearly 200 years ago with the investiga¬ tions of Labillardiere and Riche on the "Recherche" and "Esperance" under the command of d'Entrecasteaux. Matthew Flinders charted the islands in 1802 and Cap¬ tain Phillip Parker King of the Royal Navy made a short visit in 1818. The remainder of the nineteenth century saw the Archipelago exploited by sealers arid pastoralists, the latter with minimal success. The little biological data collected during this time was anecdotal and imprecise (Bechervaise 1954). The beginning of the twentieth century saw the first attempt to systematically investigate the natural history of the islands. JT Tunney, the Western Australian Museum's professional collector at the turn of the cen¬ tury, made mammal and bird collections from some of the largest islands in 1904 and 1906 (Whittell 1954). This was followed by an expedition in 1921 which included Messrs Hull, Wright and Grant (Hull 1922). DL Serventy made visits in 1944 (Serventy 1947) and again in 1947 and 1948 (Serventy 1952). In 1950, an expedition organised by the Australian Geographical Society sys- © Royal Society of Western Australia 1996 tematically worked its way through the Archipelago, vis¬ iting 20 islands (Fairbridge & Serventy 1954; Glauert 1954 & Willis 1953). More recently, naturalists such as Abbott & Black (1978), Goodsell et al (1976), Lane (1982a,b,c), Lane & Daw (1985), Maryan & Robinson (1987) and Tingay & Tingay (1982a,b) have published on collections and observations on the reptiles of the Archipelago. Other contributors, including AA Burbidge, I Cook, J Dell, A Hopkins, D Knowles, NL McKenzie, D Pearson & A Williams, VN Serventy, AN Start and A Weston, have lodged specimens and unpublished observations with the Western Australian Museum. Since 1979, we have had the opportunity of visiting many more islands to collect birds and reptiles. Despite all efforts, Storr (1987) lists bird data from only about 50 islands, while Western Australian Museum records together with published ob¬ servations give herpetological data for only 47 islands. The aim of this paper is to collate and review the scattered published and unpublished herpetological data available for the Archipelago of the Recherche, and to compare it with the herpetofauna of the opposite main¬ land. Results The particular species that have been recorded on the various islands of the Recherche Archipelago are listed in Appendix 1 and summarised in Table 1. Recent col¬ lecting has thrown doubt on some earlier observations; these earlier and doubtful observations, and certain uni¬ dentifiable published descriptions of reptiles from par¬ ticular islands, are excluded from this list. Substantial efforts by a number of collectors have not confirmed the sight record of E. kingii from Daw Island (Glauert 1954). Specimens of Egernia napoleonis from the south coast of Western Australia, east of Bremer Bay (including the Archipelago of the Recherche), are larger than specimens from Western populations (Storr 1978). Specimens of E. napoleonis from Daw Island are excep- 165 V 9 Journal of the Royal Society of Western Australia, 79(2), June 1996 Westall 9 D Glennie Round Douglas Cooper <7 herche una of the Arch.pelago of the Recherche is compared; Two Peoples Bay (1), Fitzgerald River National Park (2), Cape Le Grand National Park (3), Journal of the Royal Society of Western Australia, 79(2), June 1996 Table 1 The number of islands which have one species, two species, and so on, up to 16 species supported by the largest island. The area of the largest island, within a suite of islands supporting the same number of species, is also given. Number of species Number of islands Area of largest island (ha) 1 7 100 2 8 140 3 6 130 4 7 100 5 9 280 6 3 285 7 2 190 8 1 125 9 1 315 10 0 - 11 1 300 12 0 - 13 0 _ 14 0 - 15 1 780 16 1 1080 tionally large (up to 145 mm SVL and 94 grams) so it is possible that Glauert's observation is based on these large E. napoleonis. Despite extensive work on Mondrain Island by Abbott & Black (1976), Johnstone (unpublished observa¬ tions) and Pearson & Williams (pcrs. comm.), there has not been a specimen of Hemiergis peronii to substantiate a sight record (Glauert 1954). It is possible that Glauert's sight record is based on Lerista dorsalis, because H. peronii and L. dorsalis have the same digital formula (4+4) and many H. peronii from the Archipelago develop dorsal stripes similar to those of L. dorsalis. Names of some lizards have changed since the publi¬ cation of earlier lists. The skink Lerista dorsalis is the Lerista frosti of Abbott & Black (1978) and the Ablepharus elegans of Glauert (1954). Gemmatophora norrisi is the Amphibolous muricatus of Abbott & Black (1978). The names of some islands have also changed. In the past, MacKenzie Island (34°12 S, 122°06 E) has been called Round Island; Westall Island (34°05’S, 122°58’E) has been called Combe Island; Cull Island (33°55’S, 121°54’E) has been called Gull Island; Kermadec Island (34°05'S, 122°05'E) has been called Wedge Island; Wickham Island (34°0TS, 123°17’E) has been called Stanley Island; and Daw Island (33°5TS, 124°06'E) has been called Christmas Island. Anvil Island (33°44'S, 1241>05'E), Skink Island (33°59'S, 123°09’E), Harlequin Island (34°01'S, 123°14'E), Table 2 Habitat of 19 reptile and amphibian taxa in the Archipelago of the Recherche. Numbers in columns indicate the number of individuals found in each habitat. T3 G X 3 G (V Ih 50 T3 G 3 G O T3 C 3 T3 G 3 Species Habitat G D c O G o c O Litoria cyclorhynchus 2 Phyllodactylus m. marmoratus 75 1 2 Underzvoodisaurus milii 29 2 Ctenophorus omatus 29 Gemmatophora norrisi 4 Crypt oblepharus virgatus clarus 1 Ctenotus labillardieri 65 Egemia kingii 18 1 Egemia multiscutata bos 2 Egemia napoleonis 2 80 1 Hemiergis peronii 9 Bassiana trilineata 3 5 Morethia obscura 1 Lerista microtis intermedia Til i qua rugosa rugosa 2 Varanus rosenbergi 1 Morelia spilota imbricata 1 Acanthophis antarcticus 2 Notechis coronatus 10 2 g- 8 OJ TJ C D T3 G X G o o c/i C 3 X G 168 Journal of the Royal Society of Western Australia, 79(2), June 1996 Hull Island (33°58'S, 122°50'E, Six Mile Island (33°38'S, 123°57'E) and Tunney Island (33°58'S, 122°49'E) have only recently been named. Discussion Species distribution within the Recherche Archipelago Less than half of the islands in the Recherche Archi¬ pelago have reptiles or frogs collected from them, and the collecting effort varies from many days in the case of the larger islands to only an hour or so for the smaller islands. Such incomplete data only allows the most gen¬ eral of conclusions regarding the distribution of species within the Archipelago and species diversity-island area relationships to be drawn. Table 1, which summarises the data in Appendix 1, shows the number of islands which have one species, two species and so on up to 16 species, and the area of the largest island within a suite of islands supporting the same number of species. The most widely distrib¬ uted species in the Archipelago, the gecko Phyllodactylus marmoratus marmoratus, occurs on 32 islands. It is fol¬ lowed by various skinks; Egernia kingii (31), Egernia napoleonis (30), Ctenotus labillardieri (21), and the gecko Underwoodisaurus milii (19) and the skink Herniergis peronii (19). We consider these six species to comprise the ''core" herpetofauna of the Archipelago. The most common and widely distributed snake ( Notechis coronatus) is found on nine islands. Morelia spilota imbricata and Pseudonaja affinis tanneri are the least com¬ mon snakes, with the python only occurring on the two largest islands (Middle and Mondrain), and the endemic P. affinis tanneri only on Boxer and Figure of Eight Is¬ lands, at the western end of the Archipelago. The skink Lerista microtis intermedia is restricted to the predomi¬ nantly sandy Wickham Island, while the monitor Varanas rostmbergi is only found on Middle Island. The four largest islands, of Middle (1080 ha). Mondrain (780 ha), Salisbury (315 ha) and North Twin Peak (300 ha), support 16, 15, 9 and 11 species respec¬ tively, and all of them except Salisbury has a complete suite of the six core species. Fewer species utilise the habitats created by limestone than by granites (Table 2), which may explain why Salisbury Island, which is largely Quaternary eolianite (Bechervaise 1954), supports the fewest species of the four largest islands. The remaining 43 smaller islands are less than 300 ha and support only up to eight species. Comparison with mainland herpetofauna Lack of data may preclude a thorough analysis of the intra-island herpetofauna for the Archipelago of the Re¬ cherche but it does not hinder a biogeographic compari¬ son with the opposite mainland because the herpetofauna assemblages of both areas should now be nearly completely known. Although many small islands in the Archipelago remain unvisited, they probably only support a few of the common species (most likely some, but not all, of the six core suite of species). It is unlikely that few, if any, amphibian or reptile species remain to be added to the 21 species so far recorded from the Ar¬ chipelago. Details of the lower west and south coast mainland herpetofauna have been obtained from studies of the coastal strip between Busselton and Two Peoples Bay (How et al. 1987), Cape Le Grand National Park (Kitchener et al 1975) and the Great Australian Bight (Storr et al. 1981; Congreve 1985; McKenzie & Robinson 1987; Greer et al. 1991). These data, which are summarised in Table 3 and augmented with data from Western Australian Museum records pertaining to the plain below Wylie Escarpment (see Figure 1), indicate that the coastal strip from Augusta to Eucla supports a moderately rich herpetofauna. The herpetology of the long coastal strip between Two Peoples Bay and Roe Plains, to which we have compared the herpetofauna of the Archipelago of the Recherche (Table 4), transects several ecophysical zones. Two Peoples Bay, the westernmost site, is by far the wettest ( ca . 850 mm rainfall p.a .) and represents the southeastern extremity of the jarrah-marri woodlands that is generally Table 3 The number of species and subspecies recorded from the Archipelago of the Recherche and 5 regions of the adjacent south coast mainland. Numbers in brackets refer to site numbers (Appendix 2). Two Fitzgerald Cape Le Grand Recherche Below Wylie Roe Plain (5) Peoples River National National Park (3) Archipelago Escarpment (4) FAMILY Bay (1) Park (2) Leptodactylidae 7 Hylidae 2 Cheluidae 1 Gekkonidae 2 Pygopodidae 3 Agamidae 0 Scincidae 14 Varanidae 1 Typhlopidae 0 Boidae 1 Elapidae 6 9 2 1 5 5 2 14 0 1 0 5 4 2 0 3 3 3 14 1 1 1 4 0 1 0 2 0 2 11 1 0 1 3 2 1 0 5 2 3 17 1 0 0 4 0 0 0 5 1 7 12 0 0 1 3 TOTAL 37 44 36 21 35 29 169 Journal of the Royal Society of Western Australia, 79(2), June 1996 associated with leached sands over ironstone gravels (Beard 1990). Two Peoples Bay is the eastern limit of southwestern Australian reptile endemics, such as Egernia luctnosa, E. pulchra and Glaphyromorphus australis. ^ The mainland sites of Fitzgerald River National Park, Cape Le Grand National Park and the plain below Wylie Escarpment are situated on the Esperance Plain which generally becomes more arid with increasing longitude. Ravensthorpe has a mean average rainfall of 420 mm, Esperance 740 mm and Israelite Bay 388 mm ( pers . comm., Western Australian Bureau of Meteorology). This undulating plain with scrub heaths and occasional thick¬ ets of eucalypts is relieved by numerous granitic and gneissic hills and domes. Most prominent of these is the Barren Range (Fitzgerald River National Park), Mt Le Grand and Frenchman Peak (Cape Le Grand National Park) and Russell Range (plateau above Wyle Escarp¬ ment). Israelite Bay is the eastern limit of granitic out¬ crops. Offshore, where the Esperance Plain is sub¬ merged, granite domes continue to protrude above sea level and form the islands of the Archipelago of the Re¬ cherche. Dortch & Morse (1984) estimate that Middle Island was formed by rising sea levels at least 9,000 and perhaps as much as 11,000 years ago. Increasing aridity and the lack of granite, one of the more important microhabitats in the southwest of West¬ ern Australia, induces major changes to the herpetofauna of the south coast particularly in the east¬ ern portion of the Esperance Plain. The Western Swamp Turtle ( Chelodina oblonga) has its eastern limit in the Fitzgerald River drainage, while the lack of lithic com¬ plexes terminates the distribution of a number of am¬ phibians and reptiles. Cape Le Grand is the eastern limit of Ctenophorus ornatus while the lack of soaks and pools generated by granite outcrops limits the number of am¬ phibians found east of Israelite Bay (Table 3). Israelite Bay is also the eastern limit for the Tiger Snake (. Notechis scutatus occidentals ) and the pygopod Delrna fraseri frasen. The python Morelia spilota imbricata and the skink i ilicjua occipitalis have recently been recorded from the vicinity of Eyre (Griffin & Hunt 1993). The loss of these south-western herpetofaunal ele¬ ments is compensated by the western intrusion of semi- arid and arid zone reptiles, in particular agamids and Lerista, although the species diversity of the latter in no way approaches the diversity exhibited by this genus on the mid-west coast of Western Australia. Three species of agamid and four species of Lerista (Ctenophorus maculatus, C. griseus , C. pictus, Gemmatophora norrisi, Lerista dorsalis, L. baynesi and L. picturata) are found on the plain below Wylie Escarpment but not further west on the mainland. Still further northeast along the coast the Roe Plains (Nullarbor Region, Eremian Province, Beard 1990) is even more arid (Eyre 296 mm rainfall p.a .) and is isolated by the Baxter Cliffs in the west, the Hampton Plateau in the north and the cliffs at the head of the Great Austra¬ lian Bight in the east. It is vegetated with areas of cheno- pod steppe, scattered Myall and copses of mallee on heavier soils and melaleuca and dense mallee on lighter soils. There are also extensive areas of dune with little or no vegetation. The reptiles are represented by two skinks endemic to the Roe Plains ( Ctenotus brooksi euclae and Lerista arenicola) as well as elements of the Nullarbor herpetofauna not found further southwest ( Morethia adelaidensis and Pogona nullarbor). A Simpson's similarity index (Biodiversity 1993) based on the taxa listed in Appendix 2 for the mainland sites of Two Peoples Bay, Cape Le Grande and Fitzgerald National Parks, Esperance Plain below Wylie Escarpment and Roe Plains, and the geckos, pygopods, agamids, skinks and elapids in Appendix 1 using aver¬ age linkages (UPGMA), indicates that the Archipelago's herpetofauna is least like the western and easternmost mainland sites (Two Peoples Bay and Roe Plain respec¬ tively) and most similar to the herpetofauna of Cape Le Grand National Park, Fitzgerald River National Park and Wylie Escarpment of the Esperance Plain (Fig 2). What is, perhaps, a little surprising is the close affinity of the Archipelago's herpetofauna with that of the Wylie Escarpment. The analysis could have been influenced by the fact that three of the four largest islands (Middle, Salisbury, North and South Twin Peak), which collec¬ tively support about 90% of the Archipelago's herpetofaunal diversity are situated towards the more arid, eastern end of the Archipelago, and as a conse¬ quence are likely to be ecophysically and zoologically most similar to the eastern end of the Esperance Plain. Figure 2. Simpson s similarity index for Archipelago of the Re¬ cherche (AR) and the mainland sites. Two Peoples Bay (TPB) Fitzgerald River National Park (FRNP), Cape Le Grand National Park (CLGNP), Wylie Escarpment (WE) and Roe Plain (RP). 170 Journal of the Royal Society of Western Australia, 79(2), June 1996 With the exception of the endemic race of the dugite (. Pseudonaja affirm tannerii), the species of amphibians and reptiles on the Archipelago of the Recherche are moderately common to common on the adjacent main¬ land. In our view, none of these species, including the dugite, require more protection than that currently pro¬ vided by the ' A ' Class Reserve status of the majority of the islands in the Archipelago. Current South Australian Museum records (A Edwardes, South Australian Museum, pers. comm. ) indi¬ cate that there are now about twice as many species of reptile recorded from the Nuyts Archipelago than indi¬ cated by the most recent published paper on the area (Robinson & Smyth 1976). Furthermore, apart from a list of the amphibians and reptiles of Eyre Peninsula (Schwaner et al. 1985) there are no published faunal lists for discrete sites on the west coast of South Australia. This lack of recent published data precluded the exten¬ sion of this study to include island and adjacent main¬ land amphibian and reptile faunas, east to Eyre Penin¬ sula, South Australia. Nuyts Archipelago is of particular interest because it offers a glimpse of a fauna in an other¬ wise inundated terrain at the eastern end of the Great Australian Bight. This group of islands mirrors the Ar¬ chipelago of the Recherche at the western end of the Great Australian Bight. Acknowledgments: We are grateful to Mr L Spurr of Israelite Bay for his hospitality and help in getting us to many islands, to Mr and Mrs WH Butler and Mr N Kohchis whose grants helped defray the costs of some of our fieldwork in 1985 and 1986, and Dr RA How (Department of Biogeog¬ raphy and Ecology, Western Australian Museum) for assistance with Fig¬ ure 2. Ms A Edwardes, South Australian Museum kindly provided us with data on the Nuyts Archipelago. We also thank Mr B Haberley, Department of Conservation and Land Management, Esperance for assistance in 1987 and Mr B Goodchild of Department of Lands and Survey for help with island names. References Abbott I & Black R 1978 An ecological reconnaissance of four islands in the Archipelago of the Recherche, Western Austra¬ lia. Journal of the Royal Society of Western Australia 60:115-128. Beard JS 1990 Plant Life of Western Australia. Kangaroo Press, Kenthurst, NSW. Bechervaise JM 1954 The Archipelago of the Recherche, la. Gen¬ eral History. Australian Geographic Society Report 1:1-7. Biodiversity 1993. Computer program for calculating biological diversity parameters similarity, niche overlap and duster analysis. Pensoft, Sofia. Chapman A & Dell J 1975 Reptiles, amphibians, fishes. In: A Biological Survey of Cape Le Grande National Park (ed AF Lovell). Records of the Western Australian Museum Supple¬ ment 1:34-40. Congreve P 1985 Reptiles recorded from the vicinity of Eyre Bird observatory. Eyre Bird Observatory Report 1981- 1983:125-128. Dortch CE & Morse K 1984 Prehistoric stone artefacts on some offshore islands in Western Australia. Australian Archaeol¬ ogy 19:31-47. Fairbridge RW & Serventy VN 1954 The Archipelago of Recher¬ che. lb. Physiography. Australian Geographic Society Report 1:9-28. Glauert L 1954 The Archipelago of the Recherche 5. Reptiles and Frogs. Australian Geographic Society Report 1:29-35. Goodsell J, Tingay A & Tingay SR 1976 A resource study of Woody Island, Archipelago of the Recherche, WA Depart¬ ment of Fisheries and Wildlife, Perth, Report 21. Greer AE, Thorpe R & Malhotra A 1991 Natural history notes on lizards from the Roe Plain, Western Australia. Western Australian Naturalist 18:178-184. Griffin P & Hunt T 1993 First record of two reptiles Morelia spilota imbricata (Boidae) and Tilicjua occipitalis (Scincidae) from the vicinity of Eyre Bird Observatory, Western Austra¬ lia. Western Australian Naturalist 19:186-187. How R A, Dell J & Humphreys WF 1987 The ground vertebrate fauna of coastal areas between Busselton and Albany, West¬ ern Australia. Records of the Western Australian Museum 13:553-574. Hull AFB 1922 A visit to the Archipelago of the Recherche, S.W. Australia. Emu 1:277-289. Kitchener DJ, Chapman A & Dell J 1974 A Biological Survey of Cape Le Grand National Park. Records of the Western Aus¬ tralian Museum, Perth. Supplement 1. Lane SG 1982a Seabird Islands No 117. MacKenzie Island, Ar¬ chipelago of the Recherche, Western Australia. Corella 6:57- 58. Lane SG 1982b Seabird Islands No. 119. Frederick Island, Archi¬ pelago of the Recherche, Western Australia. Corella 6:61-62 Lane SG 1982c Seabird Islands No 121. Remark Island, Archi¬ pelago of the Recherche, Western Australia. Corella 6: 65-66. Lane SG &: Daw AK 1985 Seabird Islands No 147. Charley Is¬ land, Archipelago of the Recherche, Western Australia. Corella 8:119-120. Maryan B & Robinson RD 1987 Notes on the herpeto fauna of Woody Island, Archipelago of the Recherche. Western Aus¬ tralian Naturalist 17:3-4. McKenzie NL Rolfe JK & Carter DB 1984 Reptiles In: A biologi¬ cal Survey of the Nullarbor region South and Western Austra¬ lia in 1984 (eds NL McKenzie & AC Robinson). Government Printer, Adelaide, 179-210. Robinson AC & Smyth MEB 1976 The vertebrate fauna of Nuyts Archipelago South Australia. Transactions of the Royal Soci¬ ety of South Australia 100:171-176. Schwaner TD, Miller B & Tyler MJ 1985 Reptiles and Amphib¬ ians. In: Natural History of Eyre Peninsula (eds CR Twidale & MJ Tyler). Royal Society of South Australia, Adelaide, 160- 167. Serventy DL 1947 Notes from the Recherche Archipelago. Emu 47:44-49. Serventy VN 1952 The Archipelago of the Recherche. 2. Birds. Australian Geographic Society Reports 1:4-23. Storr GM 1978 The genus Egemia (Lacertilia, Scincidae) in West¬ ern Australia. Records of the Western Australian Museum 6:147-187. Storr GM, Hanlon TMS & Harold G 1981 The herpetofauna of the shores and hinterland of the Great Australian Bight, Western Australia. Records of the Western Australian Museum 9:23-39. Storr GM 1987 Birds of the Eucla Division of Western Australia. Records of the Western Australian Museum, Perth, Supple¬ ment 27. Tingay A & Tingay SR 1982a Seabird Islands No. 118. Hood Island, Archipelago of the Recherche, Western Australia. Corella t>:59-60. Tingay A & Tingay SR 1982b Seabird Islands No. 120. Sandy Hook Island, Archipelago of the Recherche, Western Austra¬ lia. Corella 16:63-64. Whittell HM 1954 The Literature of Australian Birds. Paterson Brokensha, Perth. Willis JH 1953 The Archipelago of the Recherche. 3a. Land Flora. Australian Geographic Society Reports 1:3-35. 171 journal of the Royal Society of Western Australia, 79(2), June 1996 Appendix 1. The herpetofauna of the islands of the Archipelago of the Recherche on which species have been found. * indicates specimens in Western Australian Museum; letters indicate observations as follows: a, Australian Geographical Society Expedition (Glauert 1954); b, Lane & Daw (1985); c, Lane (1982a); d, Lane (1982b); e, Lane (1982c); f, Tingay & Tingay (1982a); g, Tingay & Tingay (1982b); h, Maryan & Robinson (1987); i, Goodsell et al (1976); j, Smith & Johnstone (unpublished notes); k, Pearson & Williams (pers. comm.). ISLAND ANVIL 25 j * * * ♦ 5 BELLINGER 40 * * + * * 5 BEN 55 * * ♦ * 4 BOXER 200 * a * * * * 6 CAVE 3 a * * a 4 CHARLEY 100 * * a 3 CORBET 100 * b * * 4 CULL 70 * * * 3 DAW 180 * * * a * * if 7 DOUGLAS 30 a 2 FIG. OF EIGHT 200 ♦ * a a * 5 FORREST 20 * * * * 4 FREDERICK 80 d 1 GOOSE 85 * if ♦ if 5 GULCH 100 * * * * if 5 GUNTON 95 * a 2 HARLEQUIN 10 * * 3 HIGH 10 * * * * if 5 HOOD 130 f f f 3 HOPE 25 * 1 HULL 20 * * if * 4 INSHORE 35 * * * * 5 KERMADEC 30 a a 2 LONG 140 a a 2 MACKENZIE 50 c c 2 MIDDLE 1080 * * * ♦ * * * ♦ * * * + ♦ * if 16 MONDRAIN 780 * * * ♦ * ♦ * * * a * * * k if 15 NEW YEAR 20 ♦ 1 NARES 5 * * * 3 N TWIN PEAK 300 »f * * a ♦ * ♦ * * >f 11 S TWIN PEAK 115 * * a 3 OWEN 25 * * 2 PASCO 85 a a a a a 5 RABBIT <10 * 1 REMARK 100 * a * a&e 4 SALISBURY 315 * * * * * * * ♦ if 9 SANDY HOOK 285 * * g ♦ g a&g 6 SIX MILE 10 * 1 SKINK 10 * * * 3 TAYLOR 15 * a ♦ ♦ * if 6 TERMINATION 85 * * * a a 5 THOMAS 100 * * 2 TUNNEY 25 * + * * 4 WESTALL 100 ♦ 1 WICKHAM 40 ♦ * 2 WILSON 125 * * ♦ ♦ * * * >f 8 WOODY 190 h&i h&i h h&i * h&i h 7 172 Journal of the Royal Society of Western Australia, 79(2), June 1996 Appendix 2. Species and subspecies of geckos, pygopods, agamids, skinks and elapid snakes recorded from 5 sites on the south coast mainland adjacent to the Archipelago of the Recherche. Numbers in brackets refer to site numbers (Fig 1). Two Peoples Bay (1) Fitzgerald River Cape Le Grand Wylie Roe Plain (5) National Park (2) National Park (3) Escarpment (4) Gekkonidae Crenadactylus ocellatus ocellatus Gehyra variegata Phyllodactylus marmoratus marmoratus * Phyllodactylus marmoratus alexanderi Diplodactylus granariensis * Diplodactylus spinigerus inornatus Underwoodisaurus milii * * * * * * * * ♦ * * * * * * * * * Pygopodidae Aprasia inaurita Aprasia repens Aprasia striolata * Delma australis * Delma fraseri fraseri Pygopus lepidopodus lepidopodus * * * * * * * * * * * Agamidae Ctenophorus maculatus griseus Ctenophorus omatus Ctenophorus pictus Gemmatophora norrisi Pogona minor minor Pogona nullarbor Tympanocryptis adelaidensis chapmani * * * * * * * * * * * * * * Scincidae Cryptoblephurus virgatus clarus Bassiana trilineata * Ctenotus catenifer * Ctenotus gemmula Ctenotus labillardieri * Ctenotus irnpar Ctenotus brooksi euclae Egemia bos * Egemia kingii * Egemia luctuosa * Egem ia napoleo nis * Egemia pulchra * Glaphyromorphus australis * Hemiergis initialis brookeri Hemiergis peronii * Lerista baynesi Lerista distinguenda Lerista microtis microtis Lerista microtis intermedia Lerista dorsalis Lerista picturata Lerista arenicola Menetia greyii Pseudomoia baudini Tiliqua occipitalis Tiliqua rugosa rugosa Morethia adelaidensis Morethia obscura * * ** ******** ***** * * **** *** ** * **** *** *** **** ***** * * **** * ** Elapidae Notechis coronatus Notechis curtus Notechis mastersii Notechis minor Notechis scutatus occidentalis Pseudonaja affinis affinis Rhinoplocephalus bicolor Rhinoplocephalus nigriceps Rhinoplocephalus spectabilis * * * * * * * * * * * x- * * * * 173 Journal of the Royal Society of Western Australia, 79:175-181, 1996 Population and plant growth studies of six species of Eremophila (Myoporaceae) from central Western Australia G S Richmond1 & E L Ghisalberti2 1 School of Environmental Biology, Curtin University of Technology, GPO Box U1987, Perth WA 6001; present address: BSD Consultants, PO Box 155, Subiaco WA 6008 2 Department of Chemistry, The University of Western Australia, Nedlands WA 6907 manuscript received January 1996; accepted July 1996 Abstract Growth of six Eremophila (Myoporaceae) species occurring in central Western Australia were investigated over a period of three years. Field studies at Mt Keith Station (Wiluna) and Mt Weld Station (Laverton) indicated that Eremophila showed both germination and growth responses after heavy rainfalls. Eremophila spectabilis brevis increased in number from 2407 to 17684 plants ha 1 after a one in fifteen year rainfall event. Eremophila fraserii galeata increased from 2650 to 9640 plants ha1 in the same period. The other species monitored, E. exilifoha, E , forrestii, E. latrobei latrobei and E. margarethae, also showed significant population increases. Eremophila germination strategies in¬ cluded generation of multiple seedlings as well as staggered germination from the same fruit over a one vear period. Seedling survival was as high as 72% for E. latrobei latrobei after one year and seedlings reached a mean height of 8.3 cm in this period. Adult plants increased in height and leaf cover even during adverse seasons. A study of the seed bank of an E. fraseri galeata community revealed a mean of 56 fruits nr2 under shrubs compared to 2 fruits nv2 in bare ground. Introduction Eremophila species (Myoporaceae) are hardy peren¬ nial shrubs and small trees which occur throughout the arid and semi-arid regions of Australia (Chinnock 1981, 1986), and currently number over 315 species and sub¬ species (Chinnock pers. comm.). The main centre of di¬ versity is the Austin phytogeographic region of Western Australia (Beard 1980). Since many species are tolerant to drought, fire, frost, salinity and grazing, interest in this genus is centred on its potential in rangeland reveg¬ etation and minesite rehabilitation programmes (Rich¬ mond & Ghisalberti 1994a). There have been number of studies on the taxonomy and biology of Eremophila (Beard 1965; Barlow 1971; Bowen 1975; Smith 1975) and their propagation potential (Lothian & Holliday 1964, Beard 1968; Wrigley & Fagg 1979). A detailed examina¬ tion of the reproductive biology, vegetative and floral morphology of Eremophila has also been reported (Chinnock 1982). Knowledge of the population dynamics of species of this genus is limited. Fruit production and germination in E. gilesii and E. mitchellii were most noted after rains (>40 mm) during the winter months, March-September (Burrows 1971, 1972; Beeston & Webb 1977). Seed bank studies of E. gilesii have demonstrated that the soil can contain more than 400 fruits nv2. Eremophila spectabilis increased to 700 plants ha*1 yr1 in a grazed paddock during a wet period in 1973-76, and decreased to 116 plants ha1 yr1 during a drought period in 1979-82 (Gardiner 1986a, b). Eremophila delisseri and E. maculata responded favourably to grazing pressure while E. forrestii declined in numbers (Hacker 1987). This study describes plant density, seedling recruitment, survival © Royal Society of Western Australia 1996 and growth habits of six species in the Laverton and Wiluna areas of WA over a period of three years. Methods The sites selected were at Mt Keith Station (27°17'S, 120°31'E) and Mt Weld Station (28°38'S, 122°24'E) in cen¬ tral Western Australia. An initial survey showed that these sites contain a wide range of Eremophila species. From the Mt Weld site, two species were selected for detailed study; E. forrestii (MWS1) and E. margarethae (MW S3) which are co-dominant species throughout the area located within populations of Atriplex (saltbush) and Maireana (bluebush) species. From the Mt Keith site, five species were chosen; E. exilifolia F. Muell. (MKS1), E. forrestii (MKS2), E. fraseri galeata Chinnock (MKS3), E. latrobei latrobei F. Muell. (MKS4) and E. spectabilis brevis Chinnock (MKS5). Eremophila fraseri galeata dominates the understorey throughout this area, with E. spectabilis brevis as a sub-dominant species. The other three species occur at lower densities and are scattered throughout the region (Speck 1963). General details of the location of the selected species and asssociated flora are given in Table 1. Adult plant density, seedling recruitment, survivorship, and growth habits were evaluated for each species. The data were collected over the period Septem¬ ber 1990-June 1993 Assessment of seasonal plant growth period The climate of the two study areas can be described as arid. The average rainfall at Mt Keith and Mt Weld is 218 and 221 mm respectively. Winter rainfall at Mt Keith is due to the passage of westerly frontal systems, while tropical cyclones and thunderstorms account for the pre¬ cipitation during the summer months. Average monthly rainfall ranges between 4 mm in September to 37 mm in 175 Journal of the Royal Society of Western Australia, 79(2), June 1996 Table 1 Location and description of selected Eremophila species at the two study sites. Species Location Description Associated flora Mt Weld Station Site 30 km west of Laverton 28°38'S, 122°24'E E.forrestii (MWS 1) (Wilcox bush) Jubilee paddock 28058'96S, 122°27'22"E shrub to 2 m, hairy branches, obovate-oblong leaves, tubular flowers Acacia aneura, E. latrobei , Templetonia egena, Maireana triptera, Leichardia australis E. margarethae (MWS3) (Sandbank Poverty bush) Brook paddock 28°5275"S, 122°25’04"E gray shrub to 1 m, narrowly linear leaves, mauve flowers A. aneura, M. triptera, Atriplex nummularia, Ptilotus obovatus Mt Keith Site 90 km south of Wiluna 27°17'S, 120°31'E E. exilifolia (MKS1) Jumps Up paddock 27°16,51”S/ 120"27'16"E shrub to 2 m, resinous branches, pink-violet tubular flowers A. aneura-A. linopylla, E. latrobei, Calytrix glaucophylla, P. obovatus E.forrestii (MKS2) (Wilcox bush) Red paddock 27°18'96"S, 120n33'56"E shrub to 2 m, ovate to oblong leaves, pale pink flowers A. aneura, E. latrobei, P. obovatus , Cassia sp. E.fraserii galeata (MKS3) (Turpentine bush) Rocky Hills paddock 26°10'82"S, 120°36'20"E shrub to 3 m, resinous leaves, ovate leaves, red tubular flowers A. aneura, Cassia sp. E. latrobei latrobei (MKS4) (Warty fuchsia bush) Red paddock 27t’19'0rS, 120°33'54"E gray, green shrub to 2.5m, warty branches, linear-oblong leaves A. aneura, P. obovatus, Cassia sp., Triodia basedowii E. spectabilis brevis (MKS5) (Showy poverty bush) Two Tanks paddock 27°13'98"S, 120U36'84"E shrub to 2 m, resinous branches, linear-lanceolate leaves, purple flowers, dark grey bark A. aneura, E. foliosissima, Eragrostis eriopoda, Prosantha sp. March. The highest mean maximum monthly tempera¬ ture was 37.7 °C in January, and the lowest mean mini¬ mum monthly temperature was 5.3 1>C in July (Bureau of Meteorology, Perth); evaporation exceeds 5440 mm yr1. As a consequence of episodic rainfall and high evapora¬ tion rates, short growing seasons and frequent droughts are characteristic. Rainfall at Mt Weld is more depend¬ able during the late summer with precipitation being brought about by cyclonic activity. Average monthly rainfall ranges between 7 mm in October to 31 mm in March. The highest mean maximum monthly tempera¬ ture was 35.9 °C in January, and the lowest mean monthly minimum temperature was 5.2 °C in July; evaporation exceeds 3473 mm yr1 (Beard 1974). Study site design and assessment Circular study plots (diameter 25 m) were partitioned into 30° sectors in which the exact location of individual plants could be monitored (Lindsey et al. 1958; Mueller- Dombois & Ellenberg 1974). The MKS study sites were visited four times a year in the first year (1991), but this was reduced to two per year (December and June) in Table 2 Monthly rainfall (mm) for the period 1989-1993 recorded at Mt Keith (Mt Keith Station Exploration Camp, Western Mining Corporation) and Mt Weld station (Granny Smith Gold Mine, Placer Pacific Ltd). Mt Keith station Year Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total 1989 9.5 31.0 32.5 32.5 13.5 44.0 0.0 0.0 0.0 0.0 13.2 0.0 173.7 1990 100.8 18.3 7.8 1.4 9.2 6.2 12.7 26.5 1.8 1.6 1.0 0.0 187.3 1991 0.0 0.0 14.0 8.1 0.0 33.5 22.6 0.0 5.0 0.0 0.0 8.8 92.0 1992 20.3 5.6 75.6 126.4 16.8 32.2 0.0 29.8 40.4 0.0 4.2 0.9 398.7 1993 0.0 62.0 7.8 41.0 37.2 35.0 - - - - - - Mt Weld station 1989 4.4 0.0 0.0 23.0 40.6 29.8 0.0 0.0 0.0 0.0 19.6 2.2 119.4 1990 79.6 2.4 0.0 24.0 21.6 11.8 33.4 343.7 0.3 29.4 0.2 12.1 248.5 1991 0.4 0.0 3.5 25.5 52.0 37.7 12.8 0.8 2.8 7.0 0.5 40.6 183.6 1992 13.5 56.5 64.4 102.7 47.1 29.6 3.2 99.4 46.1 13.5 3.5 1.5 481.0 1993 0.1 20.5 36.0 28.2 100.2 43.7 _ _ _ _ _ . 176 Journal of the Royal Society of Western Australia, 79(2), June 1996 1992 and 1993 due to minimal seedling establishment and unfavourable seasonal conditions. The MWS study sites were visited twice per year. All study sites (exclud¬ ing MKS1 in breakaway country) were fenced to reduce the impact of sheep and kangaroos. Plant height (H), widest point (W,) and perpendicu¬ lar (W2) width wTere measured and categorised into two groups. The inverted cone shape described £. exilifolia, E. latrobei latrobei and E. spectabilis brevis , and canopy volume was calculated using the geometric formula (*/3 r2h) (Ludwig et al 1975). The upper half spheroid form (Witkowski et al. 1991), 4/3 n (W,/2) (W2/2) (H/2), was used for E. fraseri galeata, E. forrestu and E. margarethae. Changes in mean shrub height (cm) and canopy volume (m3) between summer 1990 and winter 1993 were calculated and tested using a one-way ANOVA. Differences between means were contrasted using Scheffe's test. Germination and seedling establishment Newly established seedlings were counted on each visit, tagged, labelled and the heights measured. The number of seedlings produced per fruit was individually assessed. The proportion of seedling which had germi¬ nated during the winter of 1992 and survived into winter 1993 was assessed. Seed bank variation of E. fraseri galeata (MKS3) The seed bank variation for an E. fraseri galeata com¬ munity was investigated during December 1992. Thirty adult shrubs within the height range of 1. 5-2.0 m were selected adjacent to MSK3. A further 30 sites, each situ¬ ated 3 m from the selected shrubs within a bare scalded area, were chosen at random for seed-bank sampling. Since the shrub canopy of this species overhangs the main stem with a radius of approximately 0.5 m, aim soil pit (depth 10 cm) immediately around the central stem was selected. Soil and litter fractions were sepa¬ rated, bagged, and taken to the laboratory for sieving. The soil was sieved using an automated sieve (aperture 140 mm, mesh size 2.5 mm (No. 14), diameter 21.5 cm). This fraction was further sieved by hand (wooden sieve, diameter 48.5 cm, mesh size 2.0 mm) for fruit collection. A statistical comparison of the fruit bank number be¬ tween the combined soil and litter fraction in vegetated and non-vegetated sites were assessed using a t-test. An evaluation of fruit density within the two litter and soil fractions from the bare ground and under E. fraseri galeata shrubs were assessed using a Duncan's Multiple Comparison Test. Results Although germination occurred for a number of species during the winter and summer of 1991, these values were minimal and are not discussed in detail. Plant growth and seedling establishment occurred after favourable rainfall events during winter 1992 (Table 1). Mt Keith received 400 mm (twice the yearly average), a situation that has occurred only on eight occasions over the last 100 years. Mt Weld received 480 mm, the highest recorded rainfall since records began in 1900 (Table 2). Table 3 Shrub community density, recruitment and mortality of Eremophila species at the Mt Keith (MKS) and Mt Weld sites (MWS); Su = summer, Au = autumn, Wi = winter, Sp = spring. Seedling survived refers to those seedling that had germinated in previous seasons and were still surviving at the time of the observation Season Shrub number per plot adults seedlings germinated survived dead E. exilifolia (MKS1) Su90 73 0 - - Au91 73 0 0 0 Wi91 73 0 0 0 Sp91 72 1 0 0 Su91 72 0 1 0 Wi92 69 98 1 0 Su92 64 63 86 12 Wi93 64 21 80 69 E.forrestii (MKS2) Su90 32 0 - - Au91 32 0 0 0 Wi91 31 0 0 0 Sp91 31 0 0 0 Su91 31 0 0 0 Wi92 19 32 0 0 Su92 17 17 32 0 Wi93 17 14 49 18 E. fraseri galeata (MKS3) Su90 132 0 - - Au91 132 0 0 0 Wi91 131 0 0 0 Sp91 130 0 0 0 Su91 130 0 0 0 Wi92 122 148 0 0 Su92 115 237 121 27 Wi93 108 11 183 175 E. latrobei latrobei (MKS4) Su90 42 0 - - Au91 42 0 0 0 Wi91 42 0 0 0 Sp91 42 0 0 0 Su91 42 2 0 0 Wi92 33 108 2 0 Su92 33 145 100 8 Wi93 31 33 66 79 E. spectabilis brevis (MKS5) Su90 118 0 - - Au91 118 0 0 0 Wi91 118 0 0 0 Sp91 118 0 0 0 Su91 118 0 0 0 Wi92 111 756 0 0 Su92 110 34 625 131 Wi93 109 16 349 310 E.forrestii (MWS1) Su90 31 0 - - Wi91 31 2 0 0 Su91 31 1 2 0 Wi92 28 27 0 3 Su92 28 42 27 0 Wi93 28 3 42 27 E. margarethae (MWS3) Su90 42 0 - - Wi91 42 4 0 0 Su91 42 0 2 2 Wi92 38 13 0 2 Su92 38 13 13 0 Wi93 38 0 12 14 177 Journal of the Royal Society of Western Australia, 79(2), June 1996 Eremophila exilifolia (MKS1). The adult plant density was 73 plants in 1991 (Table 3). Following significant rainfall (103 mm) in winter 1992, seedling recruitment was high (98 seedlings), increasing the population to 213, and this was largely maintained into winter 1993 (165 plants). Between winter and summer 1992, a period with little rain, adult plant mortality was 5. The majority of fruits produced individual seedlings during the period winter 1992-winter 1993, consisting of 87, 89 and 100% of total fruit recruitment. Twin seedlings were recorded during the winter and summer of 1992, and comprised 10% and 8% of total seedling production. All fruits with multiple seedlings (twins to quadruplets), which survived for one year, lost seedlings and had become individual seedlings by the winter of 1993. Eremophila forrestii (MKS2). The population of 32 adult plants during 1990 at Mt Keith declined marginally to 31 plants by summer 1991 (Table 3). During the first half of 1992, the adult population was reduced to 19 plants. However, this trend was offset by favourable rains of March (76 mm) and April 1992 (126 mm). When the 32 new seedlings were examined, all the fruits were found to produce individual seedlings. New germinants were also recorded during summer 1992 (63 seedlings) and winter 1993 (14 seedlings). The total population of 31 plants in summer 1991 had risen to 80 in winter 1993. E. fraseri galeata (MKS3). This shrub community was characterised by two age cohorts, four adult "mother" plants (average height 1.8 m) and 128 young shrubs (av¬ erage height 16 cm), estimated to be 4-5 years old. The adult community appeared stable, with a gradual de¬ cline in numbers from 132 to 108 plants from summer 1990 to winter 1993 (Table 2). This decline was compen¬ sated by a marked germination event of 148 seedlings during winter 1992. These germinants were clustered around the "mother" plants and wTere characterised pri¬ marily by individual seedlings per fruit (103 individuals; 70% of total germinants). However, twin (24%), triplet (5%) and quadruplet germinants (1%) were also re¬ corded. Germination was observed in summer 1992 when 237 germinants emerged. While the majority of multiple seedlings were reduced in number, primarily to individual seedlings per fruit, some fruits which ini¬ tially were individual seedlings continued to produce more seedlings. For example, of the 78 single seedlings which germinated in winter 1992 and survived during the summer of 1992, five were recorded as twins while one individual fruit produced triplet seedlings. In the area containing the majority of germinants, shade, nutri¬ ents and cover from annuals, e.g. Aristida arenaria, Eragrostis dielsii, Eriachne pulchella and Calandrinia spe¬ cies, were abundant. Eremophila latrobei latrohei (MKS4). The plant com¬ munity of 42 plants remained stable from summer 1990 to summer 1991 (Table 3). A marked decline in adult numbers (33 plants) occurred during winter 1992 and into winter 1993 (31 plants). However, this pattern of adult mortality was compensated by recruitment of 108 and 145 germinants during winter and summer 1992. The majority of germinants comprised individual seed¬ lings though multiple germinants were recorded, rang¬ ing from twins to quintuplet (one) fruits. In terms of total shrub density, this community peaked at 278 plants during summer 1992, but declined to 130 plants during winter 1993, a decline due primarily to seedling mortali¬ ties (44% loss). Eremophila spectahilis brevis (MKS5). The adult popu¬ lation remained stable at a density of 118 plants from summer 1990 to summer 1991, declining slightly to 111 plants by winter 1993 (Table 3). Germination following favourable early winter rains during 1992 offset this de¬ cline and 756 seedlings was recorded. High survival rates (83%) of these seedlings during the summer 1992 period with the addition of a further 34 germinants, resulted in a population increase to 769 plants. However, by winter 1993 only 349 seedlings (53%) survived. The majority of fruits produced individual seedlings (98, 91 and 94% for the winter, summer 1992 and winter 1993 respectively) and the remaining germinants were twins and triplets. Table 4 Mean shrub height and canopy volume for Eremophila species between summer 1990 and winter 1993 at Mt Keith and Mt Weld stations. Means between species with the same letters do not differ significantly (p<0.05) by Scheffe's test. Values are mean with the standard error in parentheses. Site Species subspecies Sample size (n) MKS1 E. exilifolia 67 MKS2 E. forrestii 17 MKS3 E. fraseri galeata 108 MKS4 E. latrobei latrobei 30 MKS5 E. spectabilis brevis 108 MWS1 E. forrestii 28 MWS3 E. margarethae 4 Mean shrub height (cm) Summer 1990 Winter 1993 Mean 78.13 (5.23) 85.96 (5.18) 82.04 (3.91)a 55.06 (9.40) 58.70 (10.16) 56.88 (1.82)b 24.51 (3.21) 31.88 (3.26) 28.20 (3.86)c 155.93 (10.29) 156.50 (10.03) 156.21 (0.28)d 54.40 (1.86) 58.85 (1.79) 56.47 (1.57)b 77.89 (7.13) 81.57 (6.10) 79.73 (1.84)a 62.91 (2.81) 69.14 (2.11) 66.02 (3.11)" Mean shrub canopy volume (m3) Summer 1990 Winter 1993 Mean 0.44 (0.08) 0.56 (0.09) 0.50 (0.06)a 1.24 (0.51) 1.42 (0.54) 1.33 (0.09)b 0.36 (0.17) 0.45 (0.21) 0.40 (0.04)a 0.48 (0.10) 0.61 (0.10) 0.54 (0.06)ac 0.07 (0.01) 0.11 (0.01) 0.09 (0.02)d 1.76 (0.36) 1.94 (0.38) 1.85 (0.09)" 0.46 (0.09) 0.60 (0.09) 0.53 (0.07)ac 178 Journal of the Royal Society of Western Australia, 79(2), June 1996 Eremophila forrestii (MWS1). A similar trend was ob¬ served at Mt Weld station with the most significant ger¬ mination event occuring after the winter rain of 1992. Following this period, a total of 27, 69 and 45 seedlings were recorded to winter 1993 (Table 3). The majority of new germinants were noted to be single seedlings. Eremophila margarethae (MWS3). The adult plant den¬ sity remained relatively stable (42 to 38 plants) from win¬ ter 1991 to 1993 (Table 3). Germination peaked in the winter and summer 1992, but was followed by a seedling mortality of 54% between winter 1992 and winter 1993. Mean shrub height and canopy volume The smallest mean shrub height was 28.20 ± 3.68 cm (n=108) for E. fraseri galeata and the largest shrub height was recorded for E. latrobei latrobei with 156.2 ± 0.3 cm (n=30; Table 4). In terms of canopy volume, E, spectabilis brevis and E. forrestii (MWS3; MKS2) had the lowest and highest canopy volume, 0.09 and 1.85 (1.33) m3 respec¬ tively. The remaining species were closely related with a range of 0.40 - 0.54 m3 (Table 4). Germination and seedling establishment The survival rates of all single seedlings which germi¬ nated during the winter period of 1992 was observed for 1 year (Table 5). One year after seedling establishment, 60, 69 and 30% respectively of the seedlings of E. forrestii (MWS1; MKS2) and E. margarethae (MWS3) survived. All the other species followed a constant mortality rate. The greatest proportion of mortalities was related to the spe¬ cies which produced the highest number of germinants; E. exilifolia, E , fraseri galeata , E. latrobei latrobei and £. spectabilis brevis. Seed bank variation for an E. fraseri galeata community The number of fruits in bare ground (2.2 ± 0.4) and under an E. fraseri galeata canopy layer (56.0 ± 16.4) was significantly different (p<0.01). There was also a signifi¬ cant difference (p<0.05) in fruit numbers under the plant within the soil (42.9 ± 14.3) and litter zones (13.1 ± 3.4) compared to the non-vegetated soil (1.3 ± 0.3) and litter zone (0.9 ± 0.2). If the adult population density of 4 "mother" plants is extrapolated to 81 adults ha'1, with each plant possessing a mean of 56 fruits per plant, the estimated total is 4536 fruits ha Assuming the remain¬ ing area consists of either bare ground or £. fraseri galeata community other than "mother" plants, the estimated fruit count is 21822. Since each fruit may contain up to 8 seeds, the maximum potential is 210864 seeds ha1. The Table 5 Proportion of seedlings surviving over 12 months (Wi = winter, Su = summer) Species Site Wi92 Su92 Wi93 E. exilifolia MKS1 1.00 0.75 0.42 E. forrestii MKS2 1.00 1.00 0.69 E.fraserii galeata MKS3 1.00 0.76 0.49 E. latrobei latrobei MKS4 1.00 0.93 0.72 E. spectabilis brevis MKS5 1.00 0.83 0.46 E. forrestii MWS1 1.00 1.00 0.60 E. margarethae MWS3 1.00 1.00 0.30 mean litter weight under the canopy was estimated as 5328 kg ha1, in contrast to 2693 kg ha'1 for the bare areas. Discussion Eremophila species are well adapted to arid condi¬ tions, some capable of surviving without any rainfall for approximately 2 years (Elliot & Jones 1984). Adult shrub mortalities were recorded at all sites in the seasons 1990- 1991, probably due to lack of rain. Favourable rainfall events at both sites during autumn and winter of 1992 resulted in significant levels of seedling recruitment for all species. The density of germinants at Mt Weld was markedly lower than at Mt Keith. This may be due to the competitive effect of Atriplex and Maireana species which dominated as an understorey species at the former site. At Mt Keith, with the exception of scattered communities of Eragrostis erapoda and Triodia basedoivii, the understorey consists primarily of Eremophila species. Following the high rainfall of winter 1992, the E. spectabilis brevis (MKS5) community (15420 germinants ha1) increased to a shrub population of 17684 ha'1. All fruits which produced seedlings appeared to be highly weathered, indicating that they were several years old. This, combined with the rain which leaches out germina¬ tion inhibitors in the fruit wall (Richmond 1993; Rich¬ mond & Ghisalberti 1994b), might explain the high re¬ cruitment rate. E. spectabilis brevis appears to be a highly competitive and aggressive species when favourable sea¬ sonal conditions occur (Richmond 1993). The fruits can contain up to 12 seeds compared to a maximum of eight for the other species under investigation. This may be a contributing factor to the higher germination rate. Its seedling recruitment pattern resembles that of the Queensland species E. gilesii, to which it is taxonomically closely related (Section Eremaeae; Chinnock pers. comm.). Many seedlings appeared from one fruit, illustrating the highly competitive nature of this species. While most germination involved fruits possessing single seedlings, some fruits produced multiple seed¬ lings, with quintuplets being recorded for fruits of E. fraseri galeata during the summer of 1992. Burrows (1971) also observed multiple seedlings (11.8% twin and 2.5% triplets) produced from fruits of £. gilesii after a 40 mm rainfall event. An interesting observation in the present study is that, when climatic variables are favourable, the fruits may continue to produce multiple germinants several months after initial germination. Seedlings of E. forrestii (MKS2 and MWS1), £. latrobei latrobei (MKS4) and E. margarethae (MWS3) appear resil¬ ient to desiccation, since 100, 93 and 100% survived 6 months after germination. Field observations suggested that their survival may be due to a characteristic thicken¬ ing of the hypocotyl and rapid root establishment. These hypocotyls were hirsute, thus possibly reflecting damaging light and reducing evapo-transpiration (Clifford 1987). Eremophila forrestii and E. latrobei latrobei appeared to have the fastest growth rates in terms of height gained over 1 year, with a mean of 8.3 cm, com¬ pared to 5.9 cm for the remaining species. Whilst germination was observed for all species within the study sites, seedling survival was 179 Journal of the Royal Society of Western Australia, 79(2), June 1996 characterised by an aggregation around old plant re¬ mains and under canopies of existing shrubs and small trees. Within the E. fraseri galeata community, germinants aggregated within the canopy drip line of adult plants as noted by Hacker (1984). Increased accu¬ mulation of nutrients, water availability and soil depth beneath adult plants may contribute towards the estab¬ lishment and maintenance of £. forrestii within the arid shrublands (Hacker 1979, 1984). A study of the seed bank variation within the soil and litter layer of an E. fraseri galeata community revealed a mean concentration of 56 fruits nr2 under each shrub compared to 2 fruits nr2 in bare ground. This highlights the point that the litter and detritus layers may play a role in maintaining favourable edaphic factors and microclimate, thereby promoting seedling germination and survival when better environmental conditions oc¬ cur (Hacker 1984). The total fruit count was 26402 fruits ha'1 and is an important topsoil seed bank for this com¬ munity. The fruit bank for £. gilesii is far greater, with 4285000 ha'1 being calculated by Burrows (1971) for a plant community comprising 60000 adults ha1. It is inter¬ esting to note that £. gilesii is a prolific seeder with a life span of approximately 10 years (Burrows 1974), whereas E. fraserii has a life span of 100 years (Chinnock 1974). For £. sturtii and £. mitchellii , 170000 fruits ha 1 have been recorded (Hodgkinson et al. 1980). Germination and growth of Eremophila species occur on relatively impoverished soils when favourable condi¬ tions occur. This has been illustrated by the large increase in population of £. spectabilis brevis after high rainfall, and is indicative of the ability of this genus to regenerate at high levels. A species' ability to survive both in time and space is associated and enhanced by the germination characteristics, particular if multiple seedlings emerge from one fruit. Acknowledgements: J Fox, B Tan (Curtin University of Technology) and R Chinnock (Adelaide Botanic Gardens and State Herbarium) are thanked for ecological discussions. The project was supported by the Minerals and Energy Research Institute of Western Australia (Project M156) and the following organisations; Australian Consolidated Minerals-Mt Keith op¬ eration (Western Mining Corporation); Central Kalgoorlie Gold Mines, Coolgardie Gold; Goldfan; Kaltails Mining Sendees, Placer Pacific; and Western Mining Corporation at Kambalda. Special thanks to C Woolard (WMC), P Davidson and D Ferguson for logistical support at Mt Keith. A Leeds (Mt Keith) and L Polomer (Mt Weld) gave permission for fieldwork to be carried out on the properties. The dedication of the many field assistants, G Verrall, L-A Stewart, L Rroadhurst, M Pennacchio, P Middleton, D Fletcher. J Brand, T Grubba and P Yarns, is acknowledged. References Barlow B A 1971 Cytogeography of the genus Eremophila. Austra¬ lian Journal of Botany 19:295-310. Beard ] S 1965 Descriptive catalogue of West Australian plants. Kings Park Board and the Society for Growing Australian Plants, Perth. Beard J S 1968 Eremophila. Australian Plants 4:252-254. Beard J S 1974 Vegetation survey of Western Australia. Great Victoria Desert. 1:1 000 000 Vegetation Series, Explanatory notes to sheet 3. The vegetation of the Great Victoria Desert. Map: Great Victoria Desert Sheet 3. The University of West¬ ern Australia Press, Perth. Beard J S 1980 A new phytogeographic map of Western Austra¬ lia. Western Australian Herbarium Research Notes 3:37-58. Beeston G R & A A Webb 1977 The ecology and control of Eremophila mitchellii , Queensland Department of Primary In¬ dustry Technical Bulletin 2:1-84. Bowen S E 1975 Taxonomic studies in the Australian Myoporaceae. BSc (Hons) Thesis, The University of New En¬ gland, Armadale. Burrows W H 1971 Studies in the ecology and control of Green Turkey Bush (Eremophila gilesii F.Muell.) in south-west Queensland. MAgrSc Thesis, The University of Queensland, Brisbane. Burrows W H 1972 Productivity of an arid zone shrub ( Eremophila gilesii ) community in south-western Queensland. Australian Journal of Botany 20:317-329. Burrows W H 1974 A study of the phenology and germination of Eremophila gilesii in semi-arid Queensland. In: Plant morpho¬ genesis as a basis for scientific management of rangelands (eds KW Kreitlow & RH Hart). Miscellaneous Publication 1271. Agricultural Research Service, USDA, Washington, DC, 150-159. Chinnock R J 1981 Myoporaceae. In: Flora of Central Australia (ed JP Jessop). AH & AW Reed Publications, Sydney, 338-348. Chinnock R J 1982 Taxonomy and relations in the Myoporaceae. PhD Thesis, The Flinders University of South Australia, Adelaide. Chinnock R J 1986 Family Myoporaceae. In: Flora of South Aus¬ tralia. Part III. Polemoniaceae - Compositae (eds JP Jessop & HR Toelken). South Australian Printing Division Publications, Adelaide, 1325-1345. Clifford T 1987 Seedling characteristics of Australian native and naturalised species. In: Germination of Australian native plant seed (ed P Langkamp). Inkata Press, Melbourne, 224- 231. Elliot W R & D L Jones (eds) 1984 Encyclopaedia of Australian plants suitable for cultivation. Vol 3. Lothian Publications, Melbourne. Gardiner H G 1986a Dynamics of perennial plants in the Mulga (Acacia aneura F.Muell.) zone of Western Australia. I. Rates of population change. Australian Rangeland Journal 8:18-27. Gardiner H G 1986b Dynamics of perennial plants in the Mulga (Acacia aneura F.Muell.) zone of Western Australia. II. Survival of perennial shrubs and grasses. Australian Rangeland Jour¬ nal 8:28-35. Hacker R B 1979 Studies in the ecology and range condition of two arid ecosystems. PhD Thesis, The University of New South Wales, Sydney. Hacker R B 1984 Vegetation dynamics in a grazed mulga shrubland community. I. The mid-storey shrubs. Australian Journal of Botany 32:239-249. Hacker R B 1987 Species responses to grazing and environmental factors in an arid halophytic shrubland community. Austra¬ lian Journal of Botany 35:135-50. Hodgkinson K C, G N Harrington & G E Miles 1980 Composi¬ tion, spatial and temporal variability of the soil seed pool in a Eucalyptus populnea shrub woodland in central New South Wales. Australian Journal of Ecology 23-29. Lindsey A A, J D Barton & S R Miles 1958 Field efficiencies of forest sampling methods. Journal of Ecology 39:428-444. Lothian N & I Holiday 1964 Growing Australian Plants. Rigby, Adelaide. Ludwig J A, J F Reynolds & P D Whitson 1975 Size-biomass rela¬ tionships of several Chihuahuan desert shrubs. American Midland Naturalist 94:451-461. Mueller-Dombois D& H Ellenberg (eds) 1974 Aims and Methods of Vegetation Ecology. John Wiley & Sons, Sydney. Richmond G S 1993 Seed dormancy, germination and ecology of Eremophila (Myoporaceae) in Western Australia. PhD Thesis, Curtin University of Technology, Perth. Richmond G S & E L Ghisalberti 1994a The Australian desert 180 Journal of the Royal Society of Western Australia, 79(2), June 1996 shrub Eremophila (Myoporaceae): Medicinal, cultural, horticul¬ tural and phytochemical uses. Economic Botany 48:35-59. Richmond G S & E L Ghisalberti 1994b Seed dormancy and ger¬ mination mechanisms in Eremophila (Myoporaceae). Austra¬ lian Journal of Botany 42:705-715. Smith L.S 1975 The genus Eremophila (Myoporaceae) in Queensland with notes on the genus Myoporum. Contribu¬ tions from the Queensland Herbarium. 19. Queensland Herbarium, Department of Primary Industries, Brisbane Speck N H 1963 Vegetation of the Wiluna-Meekatharra area. In: Lands of the Wiluna-Meekatharra area. Western Australia, 1958 (eds J A Mabbutt, W H Litchfield, N H Speck, J Sofoulis, D G Wilcox, J M Arnold, M Brookfield & M Wright). Land Research Series 7, CSIRO, Melbourne, 143-161 . Witkowski E T F, B B Lamont & S J Connell 1991 Seed bank dynamics of three co-occurring Banksias in south coastal Western Australia: The role of plant age, cockatoos, senes¬ cence and interfire establishment. Australian Journal of Botany 39:385-397. Wrigley ) W & M Fagg 1979 Australian Native Plants. A Manual for their Propagation, Cultivation and Use in Landscaping. Collins, Sydney. 181 Journal of the Royal Society of Western Australia, 79:183-194, 1996 Plant breeding for stable agriculture: Presidential Address 1994 W A Cowling Agriculture Western Australia, 3 Baron-Hay Court, South Perth WA 6151 Manuscript received October 1996 Introduction Plant breeding contributes to stable agriculture by improving adaptation of plants to the agricultural environment. There are many successes and failures in modern plant breeding - the key is to learn from the mistakes and build on the successes. It is my privilege to talk on the subject of plant breed¬ ing and its role in developing stable agricultural systems. Plant breeding has fascinated the human race and has been pivotal to its development since the early hunter- gatherers selected the first domesticated plants about 10,000 years ago (Duvick 1996). Plant breeding may also provide a key to survival of the human species. There is a widely held view that agriculture is intrinsically unstable and that plant breeding can do little to help, especially with regard to disease resistance. My talk challenges this view and provides a more optimistic outlook for o. ri culture. In order to be "stable", agricul¬ tural systems aust be dynamic, have minimal impact on the surrounding environment, and be well buffered against attacks of pests and diseases. 1 believe that plant breeding will underpin future progress towards stable agriculture by developing crops that are better adapted to their environment and have more durable disease re¬ sistance. However, some changes in the approach to plant breeding may be needed. Plant breeding is a very public activity, and the result of a plant breeder's labour - a new crop variety - is subject to more critical review by the public than results of most other biological science professions. There have been some spectacular failures in modern plant breeding, most notably the failure of disease resistance genes. This has over-shadowed the steady progress in yield and quality that continues to be made in modern plant breed¬ ing. I define "modern plant breeding" as the era of Mendelian genetics from the early 1900's to the present day. Resistance to rust in wheat was one of the first economically important characters proven to be under the control of a single gene. Such genes had major quali¬ tative effects that followed Mendel's Laws. There was much excitement with the discovery of the gene as the basic unit of inheritance, and zealous Mendelian crusad¬ ers used disease resistance genes to verify the impor¬ tance of their theory, and to criticise the proponents of Darwinian "gradualism" (Robinson 1996). It was soon learned that some characters were not controlled by single genes, and that not all genetic effects were qualitative. Some characters displayed quantitative rather than discrete variation. Bitter arguments erupted between the new Mendelian geneticists and followers of Darwinian "gradualism", and these arguments were not resolved until the development of the science of popula- © Royal Society of Western Australia 1996 tion genetics. Population genetics theory explained the movement of genes in populations, and also explained the genetic control of quantitative characters. Such char¬ acters did not form discrete groups in Mendelian tests, and were influenced by the environment. Biometricians developed the statistical theory that formed the basis of population breeding methods in animals and cross- fertilising plants such as lucerne. However, breeders of self-fertilising crops remained largely locked into pedi¬ gree breeding procedures developed by the early Mende¬ lian geneticists 90 years ago - especially with respect to disease resistance (Robinson 1996). In many crops, single "Mendelian" genes for disease resistance have often "broken down" due to the develop¬ ment of new races of the pathogen. The situation with wheat stem rust and potato late blight is documented and discussed by Vanderplank (1963). This has happened so many times and in so many crops in the past 90 years that most farmers and scientists alike have come to believe that all disease resistance is temporary - that eventually all disease resistance breaks down. There is no doubt that Mendelian disease resistance genes have been of great economic value in some crops, although the hidden costs of such breeding may be high. Neverthe¬ less, the failure of some types of disease resistance genes have prompted many authors such as Vanderplank (1968) and Robinson (1987) to promote alternative approaches to breeding for durable resistance. Plant breeders are reluctant to change from pedigree breeding methods because they continue to make breed¬ ing progress with these methods (Kannenberg & Falk 1995). In pedigree breeding, selfing occurs for several generations to allow selection on near-homozygous and uniform lines. The tendency is to restrict parents to a few "tried and proven" varieties, and cycles of crossing and selection are long (10 to 15 year cycles). This has lead to narrowing of the genetic base in modern cultivars, which is a concern for many breeders (Shands & Weisner 1991, 1992). In the common bean there has been no yield progress for several decades due to the narrow germplasm base (Silbernagel & Hannan 1992; McClean et al. 1993). Population breeding, on the other hand, draws on a broad genetic base and employs rapid cycles of crossing and selection (1 to 4 year cycles). Parents may be het¬ erozygous early generation lines rather than "tried and proven" varieties. The introduction of new germplasm raises the genetic ceiling on yield improvement, im¬ proves ecological adaptation, and decreases vulnerability to pests and diseases (Kannenberg & Falk 1995). Popula¬ tion improvement is a powerful procedure for breeding programs to exploit genetic variability (Frey 1983). Population breeding methods have been shown to im¬ prove polygenic and durable resistance, and to secure 183 Journal of the Royal Society of Western Australia, 79(3), September 1996 long-term improvements in yield and adaptation (Carver & Bruns 1993; Cassler & Pederson 1996; Robinson 1987). Such methods should contribute to the long term stability of agriculture. Stable agriculture - past and present Instability in agricultural systems has caused major social and environmental problems since the beginning of civilisation. Despite this , some stable agricultural systems are evident and some have involved lupins. Is stable agriculture possible? To answer this question, it is important to examine the factors that have rendered agriculture unstable. These include genetic vulnerability in crop plants (usually referring to vulnerability to new diseases or changes in strains of pathogens), cultivation techniques or grazing pressures that cause soil erosion by wind or water, rising water tables and salinity caused by land clearing or irrigation, continuous cropping leading to soil acidity and depletion of soil fertility and structure, and more recently, chemical contamination of ground water and produce and development of pesticide resistance. The genetic vulnerability of crop plants to disease was highlighted as the level of international trade increased in the 18th and 19th century. Potatoes provide an interest¬ ing example of extreme susceptibility in a crop plant that is exposed to a new pathogen for the first time. Potatoes were introduced into Europe from South America by the Spanish in the 16th century. Two centuries of selection by European horticulturalists altered the potato from a tropical plant to a long-day plant that matured before the frosts of winter (Robinson 1996). Potatoes grew particularly well in Ireland where the climate was moist and there was little frost. Social changes at the time of the industrial revolution demanded plentiful supply of cheap food, and potatoes became cheaper than bread. In Ireland, where it was difficult to grow cereals, the population flourished as the potato flourished. However, the potato in South America (where the European potato originated) had never encountered the late blight fungus, and when the late blight fungus arrived in Europe from Mexico in the 1840's, it rapidly spread through vast regions of uniformly susceptible potato crops - with devastating consequences (Large 1940) . Ireland's social structure and economy collapsed as millions died or emigrated to the USA or Australia. Instability of agriculture in WA is very evident as a result of rising water tables and salinity following land clearing. This is a major threat to long-term sustainability of agriculture. Increasing water use from agricultural lands and nature reserves is a major priority for reducing the problems in WA (George et al 1996). Alley farming and agro-forestry will help to lower salt-laden water tables and provide useful shelter for stock and wind breaks for crops (Lefroy & Scott 1994). The cost of revegetation will be large, and must be accompanied by the development of profitable and sustainable cropping systems on the best soils. Some agricultural systems have remained relatively stable for many years - take for example cork oak planta¬ tions in Portugal (Pinto 1994). The cork oak soils in southern Portugal are very poor - not unlike soils in the south-west of WA. Cork is harvested from the trees ev¬ ery nine years. During this nine year cycle, up to two crops of semi-bitter yellow lupins ( Lupinus luteus) with shattering pods are planted and allowed to self-seed for a second year. Sheep graze the lupin seed and stubble at the end of each year. Cultivation occurs only to plant the lupin and to control flammable vegetation. This agro¬ ecosystem has survived without chemicals or fertilisers and has proven to be a very sustainable and low-input agricultural system (Pinto 1994). Plant breeding and stable agriculture Despite the repeated break-downs in disease resistance, modern plant breeding has resulted in steady improvements in yield and quality of many crops However, genetic diversity in modern crops is low and rates of improvement may fall. Population breeding methods may improve the yield and stability of crops through more diversified gene pools and durable polygenic disease resistance. Plant breeding should underpin progress towards stable agriculture by providing farmers with well adapted and high yielding crop varieties. But how well has plant breeding achieved stable improvements in the past, and how should breeding practices be altered to maintain or accelerate these improvements in the future? Examples will be given to show that improvements in grain yield have been occurring steadily in most crops since modem plant breeding began about 90 years ago. Biotechnology will be one more addition to the tool kit of plant breeders to help them to continue this progress into the future (Duvick 1996). In the process of achieving these improvements, plant breeders have restricted their genetic diversity to a nar¬ row range of elite parents. There is a growing concern that genetic diversity is dangerously restricted in mod¬ ern crop cultivars (Shands & Weisner 1991, 1992). Crops that demonstrate continued improvements with modern plant breeding have achieved this progress in the early cycles of selection. In most crops only 5 to 8 "effective" cycles of selection have occurred since the early 1900's. There is a strong incentive to restrict the crossing parents to tried and proven varieties in pedigree breeding and pure line methods. It is very difficult to find improve¬ ments in single crosses outside of the main adapted cultivars. It is possible that many crops have not yet reached a yield plateau because of these slow cycles of selection. The yield plateau in common bean (Silbernagel & Hannan 1992; McClean et al. 1993) provides a timely warning - there is an urgent need to introduce additional genetic variability into breeding systems (Kannenberg & Falk 1995). Yield improvements in modern crop cultivars are of¬ ten dependent on major genes for disease and pest resis¬ tance, and the extensive use of pesticides. Pedigree breed¬ ing attempts to maintain disease resistance through a procedure known as backcrossing, whereby new disease resistance genes are transferred from a "good source" into the elite cultivar following the "break down" of previous resistance genes (Robinson 1996). In order to contribute to stable agriculture, plant breeding should improve yield and quality in target en¬ vironments in tandem with durable disease resistance. 184 Journal of the Royal Society of Western Australia, 79(3), September 1996 Most economically-important traits, including yield, quality and disease resistance, may be considered as polygenic characters that are subject to the laws of popu¬ lation genetics (rather than Mendel ian genetics). In many cases, the rate of yield improvement may be increased and genetic vulnerability decreased by the application of population breeding methods. Recurrent selection is a common population breeding method that increases the frequency of favourable genes in a population (Frey 1983). The rate of improvement depends on genetic di¬ versity, selection pressure and the duration of selection cycles. Population breeding methods should enhance the development of favourable combinations of genes that are likely to contribute to the long-term stability of agri¬ culture. Plant breeding progress Breeding progress has been apparent in the major crops in most countries over the past 90 years. Examples include wheat, barley, soybean, cotton, sorghum and corn. Grain yield of soft red winter wheat varieties in Ohio increased by 2259 kg ha 1 from 1910 to 1991 (1.3% per year from 1947-1987), based on historical variety trials conducted at high rates of fertilisation but without fungi¬ cides (Fig 1). Higher yields were associated with earlier flowering, reduced height, and greater resistance to lodging in modern varieties (Berzonsky & La fever 1993). Yield of barley varieties in Eastern Canada increased steadily by 1.0% per year from 1956 to 1988, with no signs of a yield plateau. Harvest index increased from 0.44 to 0.51, and biomass also increased. Modern culti- vars are heavier, have stronger stems and tend to be shorter, with greater lodging resistance (Bulman et al. 1993). Similar genetic improvements are noted in yield of corn (Tollenaar et al 1994), soybean, sorghum, cotton and wheat in the USA (Fehr 1984), and in yield and quality of wheat in South Africa (van Lill & Purchase 1995). It is important to remember that varieties bred in one country are not necessarily adapted in another. Estima¬ tions of breeding progress are affected by the locations of historic variety trials. Genotype x environment interac¬ Year of Cultivar Release tions affect estimates of breeding progress, or to use an analogy, there are "horses for courses" - each variety has its preferred region of adaptation. The "horses for courses" analogy applies at the international level as well as at the regional level within WA. Genetic improvement must be measured in relevant environments. Breeding progress in WA Plant breeding in WA has an impressive record of crop improvement in recent years. Breeding progress is usually measured in historic variety trials, and the results are highly dependent on how, where and what measurements are made. Improvements in lupins, barley and wheat in WA demonstrate these points. The techniques chosen to estimate breeding progress can have a major influence on the outcome of the experi¬ ments. In 1991 and 1992, I conducted a series of historic variety trials in the cropping region of southwest WA using an historical set of narrow-leafed lupin (L. angustifolius) varieties released earlier by Dr John Gladstones. Each variety was sown at a wide range of seeding rates. Merrit, released in 1991, was significantly higher yielding than Unicrop, released in 1973. At a tar¬ get density of 70 plants nr2, the yield of Merrit was 43% higher yielding than Unicrop, a yield improvement of 2.4% per year from 1973 to 1991 (Cowling & Speijers 1994). At a seeding rate of 120 kg ha1, the estimate of breeding progress was 1.9% per year between Unicrop and Merrit. However, at low seeding rates of 30-40 kg ha1, there was no difference in yield between the two varieties, and no apparent progress in yield improve¬ ment (Fig 2). It follows that estimates of breeding progress depend on the seeding rates at which varieties are tested. As with lupins in WA, modern varieties of corn perform better than older varieties at high densities (Tollenaar et al. 1994). It makes good economic sense for a farmer to sow Merrit lupins at higher seeding rates than Unicrop. Seeding Rate (kg ha 0 Figure 2. Influence of seeding rate (kg ha1) on yield (kg ha1) of narrow-leafed lupin cultivars Merrit (solid line; released in 1991) and Unicrop (broken line; released in 1973) in historic variety trials in Western Australia. (Source: Cowling & Speijers 1994). Figure 1. Grain yield (kg ha1) of soft red winter wheat varieties in Ohio based on historical variety trials (Source: Berzonsky & La fever 1993). 185 Journal of the Royal Society of Western Australia, 79(3), September 1996 Figure 3. Results of historic variety trials in Western Australia of narrow-leafed lupin cultivars Unicrop (released 1973), Illyarrie (1979), Danja (1986), Gungurru (1988) and Merrit (1991): top Relative yield (kg ha1) at predicted maximum yield; centre Bio¬ mass (g plant-1 dry weight) at harvest; and bottom Harvest In¬ dex. (Sources: Cowling & Speijers 1994, and Tapscott et al. 1994). varieties perform better in low rainfall regions, others perform better in high rainfall regions of WA (R F Gilmour, pers. comm.). In the high rainfall region, progress in barley breeding in WA has been very impressive - yield has improved by 2.5% per year from Dampier (released 1966) to Onslow (released 1990). Onslow yields more than twice that of Prior, released in 1900 (Fig 4 top). In the low rainfall region, yield improvement has not been as rapid, but is still very reasonable (Fig 4 bottom), increasing 1.5% per year from Dampier (1966) to O'Connor (1984). Onslow is not adapted to low rainfall regions, and yields less than Prior in low rainfall! It would be quite misleading to measure breeding progress in barley based on the average yield of these varieties across rainfall regions. Breeding progress in barley is occurring in different gene pools in the high rainfall than in the low rainfall. When estimating breeding progress, it is also impor¬ tant to define what is being measured. The wheat breed¬ ing program in Agriculture WA is the first in the world to breed a variety specifically for the Japanese noodle market. The variety Cadoux was released in 1992. Cadoux is higher yielding than its predecessor, and gives WA a competitive edge in the Japanese noodle market over its rivals in Canada and the USA. Growers receive a price bonus for Cadoux wheat. This is equivalent to 10- 3200 to 2400 O) 2 0) 1600 800 Conversely, any factors that decrease the plant density of Merrit, such as poor germination or root rot disease, will decrease its relative yield advantage. While there has been a progressive improvement in yield of lupin varieties from Unicrop to Merrit (Fig 3 top), total plant weight or biomass of the historical culti¬ vars has not changed (Fig 3 centre) over this 18 year period (Tapscott et al. 1994). Total weight includes seed, leaves and stems at harvest. Harvest index, or the pro¬ portion of the mature plant dry weight in seed, has in¬ creased steadily from Unicrop to Merrit (Fig 3 bottom) (Tapscott et al. 1994). Merrit is more efficient at convert- ing vegetative mass to seed than Unicrop. However, har¬ vest index is still low in comparison with other grain legumes, and harvest index of lupins in WA may be higher in future higher-yielding varieties. Estimations of breeding progress are also affected by locations of variety trials. In this example, some barley Figure 4. Yield (kg ha1) of barley varieties in historic variety trials in high rainfall (top) and low rainfall (bottom) regions of Western Australia (Source: R F Gilmour, pers. comm.). 186 Journal of the Royal Society of Western Australia, 79(3), September 1996 20% yield improvement in normal Australian Standard White varieties. The market potential was recognised by the breeders and cereal chemists at least 10 years in ad¬ vance. Selection techniques for Japanese noodle qualities had to be developed and working effectively long before the new wheat was released (G Crosbie, pers. comm.). The Japanese noodle wheat has lifted the profitability of crop¬ ping and provided cash for stability projects on the farm. Lupins and stable agriculture in WA Lupins have rapidly expanded in area and production in WA over the past two decades. Lupins filled the need for a produc¬ tive legume in rotations with cereal crops on light sandy soils. Improvements in crop rotations due to lupins will improve the profitability and stability of agriculture. Improvements in farm profits may allow increased expenditure on land and soil conservation measures. Narrow-leafed lupin production increased exponen¬ tially during the 1980's in WA to more than one million tonnes per annum by the early 1990's. Average lupin yield per hectare in WA improved by about 20% during the 1980's (Fig 5). This was partly the result of better lupin agronomy and partly due to better lupin varieties. Lupins are now the second largest crop in WA based on area of production (Australian Bureau of Statistics 1996). Farmers have responded quickly to incorporate this le¬ gume into their farming systems. In doing so, they have improved their economic viability and the long-term sta¬ bility of agriculture. Lupins provide a direct benefit for the stability of ag¬ riculture in WA, mostly through improvements in crop rotations with cereals. Wheat following lupins is higher yielding, has lower disease levels, fewer weeds and higher grain protein than wheat following another cereal or poor pasture (Nelson 1994). Consequently, the crop rotation is more profitable and more stable in the long term when lupins are included in the system. Lupins also improve soil structure and fertility, are a cash crop in their own right, and the stubbles and seed provide valuable on-farm feed for livestock over sum- 1.5 1 0.5 0 ]D 0) > < Figure 5. Average yield (solid bars; tonne ha1) and area of pro¬ duction (stippled area; x 1000 ha) of narrow-leafed lupin in Western Australia from 1977 to 1995 (growing seasons). Source: Australian Bureau of Statistics (1996) and earlier publications of the same series. mer. Lupins lift whole-farm profitability and help to pro¬ vide farmers with the extra cash needed to help them modify their farms for long-term stability. The total biomass production of a lupinrwheat rota¬ tion exceeds that of continuous cereal cropping or cereal crop following poor grassy pasture; as a result, the lupin:wheat rotation should use more of the available soil moisture and there should be less recharge to the water table. The introduction of lupins into the rotation should reduce potential problems from rising water tables and salinity (George et al. 1996). Recently, yellow lupins (L. luteus) have been found to be very well adapted to the very acidic soils of the east¬ ern wheatbelt of WA (Sweetingham et al 1994). Yellow lupins are resistant to brown spot (Yang et al. 1996) and Pleiochaeta root rot (Sweetingham et al. 1994) and yield more than narrow-leafed lupins on these very acidic soils. This is the first legume crop for very acid soils, and as with the narrow-leafed lupin, yellow lupins will increase productivity of crop rotations on these soils. Lupins, as with any legume or nitrogen fertiliser, con¬ tribute to soil acidity. This is an increasing problem on WA soils, but fortunately there is a remedy - the applica¬ tion of lime to the soil. The added cost of lime should be funded by the increased profitability of lupin.wheat rotations on acid soils. Lupins and new approaches to plant breeding Lupins are a new crop for modern agriculture and are in an ideal position to test some new approaches to plant breeding. Lupinus is a large genus in the Leguminosea with a great diversity of forms in the Americas, the Mediterra¬ nean basin and northern Africa. Perennial lupins are found above the snow-line in Alaska and along the Cali¬ fornia coast, annual types on the fringe of the Mediterra¬ nean sea and in the highlands of equatorial Africa, and simple-leafed types on the coastal plains of Paraguay, Argentina and in southern Florida (Allen & Allen 1981; Monteiro & Gibbs 1986; Planchuelo 1994; Gladstones 1974). Lupins provide an interesting example of the parallel domestication of plants by two geographically and cul¬ turally isolated human civilisations. Lupins were devel¬ oped independently as a food crop by Greek /Roman civilisations in the Mediterranean region and by native American civilisations in the high Andes mountains of South America (Hondelmann 1984). There are frequent references to lupin cultivation in early Greek and Roman literature (Hondelmann 1984). Lupini beans (L. albus) are consumed to the present day after debittering by traditional methods in Italy, Greece, Spain, Portugal, and in the Arabic cultures of the eastern and southern Mediterranean and northern Africa. Cultivated types of the Andean lupin (L. mutabilis, known as "tarwi") were selected independently by Indian civilisations in South America. One of the early Spanish conquerors noted the similarity between the lupin eaten by the Incas and those eaten in Spain (Hondelmann 1984), and it is certain that the technology to debitter lupins was developed by both cultures independently of one another. 187 Journal of the Royal Society of Western Australia, 79(3), September 1996 As a modern crop plant, lupins have a relatively short history. Researchers in Germany in the 1920's and 1930's were convinced that low alkaloid (sweet) forms of lupins could be found. They proceeded to select sweet forms of L. angustifolius, L. albus and L. hiteus and were able to fully domesticate the latter two species. Sweet narrow- leafed lupins were not fully domesticated (with sweet and soft seeds and non-shattering pods) until the 1960's by Dr John Gladstones in WA (Gladstones 1982). The history of narrow-leafed lupin production in WA since the 1960's is a story of successful collaboration among lupin breeders, researchers, advisers, marketers and growers in WA (Nelson 1994). There are few other breeding programs for L. angustifolius in the world, and most of those are based on varieties released in WA, so the domesticated gene pool is relatively small. All improvements in narrow- leafed lupins of relevance to southern Australia for the near future will depend on breeding material developed in WA (Cowling 1994). Wild lupins are an important source of new genetic variability, and a significant proportion of the lupin breeding effort in WA is allocated to conserving and evaluating the genetic resource in the "International Lupin Collection" at Agriculture WA in South Perth (Cowling 1994). It is necessary to cross advanced breed¬ ing lines with wild plants to expand the domesticated gene pool. Crosses with wild narrow-leafed lupins from Spain, Morocco, Portugal, Israel, Italy, and Greece have formed the backbone of genetic improvements in lupins in WA and such crosses will continue to be important for some time (Gladstones 1994). It takes several years to reselect fully domesticated progeny from crosses with wild lupins, and these "first- cross" lines may not be suitable for release due to flaws in yield, quality or agronomic characteristics. Neverthe¬ less, they may have particular features (such as improve¬ ments in disease resistance) that may contribute to crop improvement. I have used recurrent selection to simulta¬ neously improve disease resistance, yield and quality in such breeding material (Cowling 1994). In 1984 when I began breeding for resistance to brown spot (caused by the fungus Pleiochaeta setosa ) in narrow-leafed lupins, there was no known "strong source" of resistance. Re¬ current selection seemed the best choice of methods for this crop and this disease (see below; "Brown Spot Resis¬ tance in Lupins: a Case History of Recurrent Selection"). Recurrent selection is a form of population breeding that is the accepted standard method in cross-pollinating forage plants (Casler & Pederson 1996), but has yet to be fully exploited in self-pollinating plants. Durable disease and pest resistance should result from the application of population breeding methods to resistance (Robinson 1987, 1996). The lupin recurrent selection program in WA is one of few applied breeding programs of self- pollinating plants to use population breeding methods. Population breeding methods Population breeding methods , although not new, offer long¬ term solutions to some serious problems with modern plant breeding. I discussed earlier the seriousness of the problems of narrow gene pools, slow selection cycles and the unreliability of Mendelian genes for resistance in mod¬ em plant breeding. To overcome some of these problems, it is necessary to use population breeding methods. Such methods should improve long-term progress in plant breeding by (i) increasing genetic variance (diversifying the gene pool in order to increase the potential for long¬ term genetic gain), and (ii) accelerating cycles of selection (decreasing the time between crossing of parents and crossing of progeny). Success depends on high selection pressure which results from accurate and reliable yield measurements, uniform disease pressure, and relevant quality measurements. Population breeding methods are designed to improve traits under polygenic control, such as yield, quality and disease resistance. Genes are mixed and re-arranged to form so-called "adapted gene complexes", and the frequency of favourable alleles in the population is increased by selection (Frey 1983). Transgressive segre¬ gation is said to have occurred when, after some cycles of crossing and selection, many lines in the final genera¬ tion are higher yielding or more disease resistant than the original parents. In recurrent selection, recurring cycles of crossing and selection allow favourable genes or gene combinations to be accumulated in the population. Recurrent selection has been used extensively (Frey 1983) in cross-pollinat¬ ing crops (such as maize) and pasture plants (for ex¬ ample sweetclover and alfalfa). More recently, it has been applied to several self-pollinating crops such as wheat, oats, barley, soybean, sorghum, bean, tobacco, and cotton (Kervella et al. 1991; Goldringer & Brabant 1993). In most references in these two review articles, recurrent selection was used to improve polygenically-controlled traits such as yield, seed size, protein, and oil. There are relatively few references in the literature to improvements in disease resistance by recurrent selection, mostly because the research required for publication has not been done. Recurrent selection in self-pollinating plants begins with a partial diallel cross among the selected parent lines, and selection normally takes place on F2-derived lines. The best progeny are selected by the F2, F3 or F4 for inter-crossing to begin the next cycle. If the parents them¬ selves are F2-derived and quite heterogeneous, it is possible to begin selecting in the Ft (called SQ in breeding of cross-pollinating plants). The population may be kept "closed", that is, without introducing new genetic lines to the population after the initial diallel cross, or new lines may be added to crossing in the following cycles. Carver & Bruns (1993) in their review found that re¬ current selection in self-pollinating crops has resulted in yield improvements of 3-4% per year, and quality im¬ provements of 5.3% per year. This compares very favourably with the breeding progress for yield in tradi¬ tional breeding programs around the world referred to earlier (1-2% per year). Pedigree plant breeding may be regarded as a form of recurrent selection, but the long-term genetic gain is limited by the slow cycles, the high number of selfing generations before intercrossing, and the narrow genetic base (Kervella et al. 1991). The chances of improving polygenic resistance are very low using traditional pedi¬ gree breeding in self-pollinating crops. 188 Journal of the Royal Society of Western Australia, 79(3), September 1996 Breeding for polygenic disease resistance by recurrent selection Recurrent selection has been shown to increase disease resistance (presumed to be polygenic) by transgressive segregation. Polygenic resistance developed by these tech¬ niques is durable. Horizontal and vertical resistance are descriptive terms to describe (i) polygenic resistance that is normally durable (horizontal) and (ii) major gene resistance that belongs to the "gene-for-gene" system of plant disease resistance (vertical). In 1954, a small but significant article appeared in the American Potato Journal. It was one of the first reports of the deliberate use of population breeding methods for disease resistance, carried out by John Niederhauser and coworkers in Mexico against late blight on potatoes. Niederhauser et al. (1954) saw that major gene resistance to late blight was of little use in Mexico, and proceeded to select partial resistance by population breeding meth¬ ods - deliberately eliminating simple major gene resis¬ tance. The partially resistant varieties remained greener in field trials than plants with "broken down" major gene resistance, which were completely dead. In addition, the partial resistance developed by population breeding was not specific to one race of the pathogen; "in the field [these varieties] show a degree of resistance that is exhibited equally toward all races of the pathogen" (Niederhauser et al. 1954). Niederhauser produced several cultivars with high levels of partial resistance, accumulated by population breeding methods. His culti- vai " itzimba" is the standard against which other culti- va; are measured in Mexico (Robinson 1987). A similar approach was used by Robinson in Kenya in the 1950's and 1960's against late blight and bacterial wilt of potato, and Robinson's potato varieties remain resistant and are grown on a wide scale in Kenya to this day without the need for expensive seed potato schemes (Robinson 1987, 1996). Several studies have demonstrated improvements in disease resistance beyond that seen in the parents, achieved by transgressive segregation. Transgressive seg¬ regation was shown for resistance to yellow rust in wheat (Krupinsky & Sharp 1979; Wallwork & Johnson 1984), leaf rust in wheat (Lee & Shaner 1985) and net blotch barley (Cherif & Harrabi 1993). Parlevliet & van Ommeren (1988) used recurrent selection to increase re¬ sistance to leaf rust and powdery mildew in barley. By the third cycle of recurrent selection, leaf rust resistance had improved substantially from very susceptible to "sufficiently resistant" (Parlevliet & van Ommeren 1988). All the original parents were considered to be very susceptible. No major genes for resistance were present. The third cycle selections were not immune, but had adequate resistance for protection of barley from leaf rust. Recurrent selection has been used to improve resis¬ tance to several diseases of wheat in Brazil (Beek 1983), purple leaf spot in orchardgrass (Zeiders et al. 1984), Phytophthora rot in soybean (Walker & Schmitthenner 1984), barley yellow dwarf virus in oat (Baltenberger et al. 1988), scab in wheat (Jiang et al. 1994) and powdery mildew in rye (Lind & Ziichner 1985). It is likely that such resistance is oligogenic or polygenic in nature, and should be durable in agriculture. In these examples, the breeders were not using the traditional "good sources" of resistance. They used re¬ current selection to improve resistance by transgressive segregation. I am not aware of reports of "break down" of polygenic resistance. Worthwhile disease resistance can be achieved without demanding immunity in plants. It is unfortunate that decades of breeding for immunity has resulted in the pessimistic attitude that resistance always breaks down and that breeders will always be one step behind the pathogen. It is difficult for population geneticists to understand why plant breeders and pathologists have argued emo¬ tionally for four decades about the value of this general type of resistance, that Niederhauser et al. (1954) termed "partial" and Vanderplank (1963) later called "horizon¬ tal". The words horizontal and vertical resistance have invoked hostile reactions in many quarters (Robinson 1996). Even today, there are very few applied breeding programs where population breeding methods are used to increase disease resistance and other important at¬ tributes in the production of new cultivars. In order to adopt population breeding methods for polygenic resistance, it is necessary for breeders to over¬ come 90 years of bias against polygenic resistance. They must accept that; • it is not necessary to locate a "good source" of dis¬ ease resistance; • disease resistance can be improved to levels far in excess of the parents by population breeding meth¬ ods, simultaneously with improvements in yield and quality; • it is rarely necessary for a crop to be immune to disease (in fact, mild disease resistance may be pref¬ erable to immunity due to lower selection pressure on the pathogen, and may prevent economic loss in combination with other disease control measures); • disease resistance does not always break down (in fact, polygenic resistance is likely to be very stable); and • there is no yield penalty associated with polygenic disease resistance. The terms vertical and horizontal resistance have been defined and refined for 30 years since Vanderplank (1963) first introduced them. There are many other terms that have been used, but since Vanderplank was the originator of the concept that differentiated the two forms of resistance, Robinson (1996) argues that his terms should take precedence. Vertical resistance genes belong to the "gene-for-gene" system of plant-pathogen interaction (Robinson 1996). Vertical genes often completely protect plants from disease; however, the protection offered by vertical resistance genes may not be durable. New races of the pathogen, with matching virulence genes, cause vertical resistance to "break down". The resistance gene inside the plant has not changed - it simply is no longer effec¬ tive as a resistance gene. The gene-for-gene system is a highly evolved and complex system of disease resistance, almost as complex as the immune system in mammals. In the gene-for-gene system, for every resistance gene in the plant population, 189 Journal of the Royal Society of Western Australia, 79(3), September 1996 there is a matching virulence gene in the pathogen popu¬ lation. Sooner or later, the pathogen population will re¬ spond to the introduction of a vertical resistance gene in a crop variety by developing (under strong selection pressure) high frequencies of its matching virulence gene to overcome the resistance. Vertical resistance is well adapted to wild ecosystems where the host tissue is discontinuous in space and time (for example, in wild cross-pollinating annual grasses), but it is not well adapted to modern agriculture where there are large ar¬ eas of genetically uniform self-pollinating crops. Plant breeders normally react to the break down of resistance by introducing new vertical resistance genes. Breeding for vertical resistance is well suited to tradi¬ tional pedigree and backcross methods of breeding. In many developed countries, such as Australia, the wheat crop continues to be protected from rust diseases by a complex combination of vertical resistance genes. Many developing countries simply cannot afford such costly disease resistance schemes (Robinson 1996). Horizontal resistance, on the other hand, is presumed to be effective against all races or strains of the pathogen. In practice this is difficult to prove, but the main feature of horizontal resistance is that it is normally polygenic and durable, although it usually provides incomplete protection and not immunity. Horizontal resistance is not the "good source" of resis¬ tance that plant breeders and pathologists traditionally seek. Horizontal resistance may be increased to useful levels by population breeding methods within an adapted gene pool. Breeding for improvements in hori¬ zontal resistance is cumulative. Small additive improve¬ ments are accumulated over several cycles of crossing and selection. In many cases, horizontal resistance is moderately to highly heritable. Resistance to brown spot in lupins was not strong in its effect, but was expressed very consis¬ tently across sites and years with high broad sense heri- tability (Cowling et al. 1997). It is possible to breed for resistance, quality and yield at the same time (Cowling 1994). However, it is not possible to breed for horizontal resistance in the presence of vertical genes - horizontal RECURRENT SELECTION CYCLES YR 2 SUMMER (IRRIGATION) Figure 6. Recurrent selection cycles for improving yield, quality and resistance to brown spot in narrow-leafed lupins in Western Australia. Cycles last 4 years from crossing to selection of F2- derived progeny for the next cycle of crossing, during which time two years of field testing is carried out in replicated yield trials. Reselection at the F5 allows superior lines to be tested for potential release as new varieties. 190 Journal of the Royal Society of Western Australia, 79(3), September 1996 resistance is masked by vertical resistance genes. Breed¬ ing for horizontal resistance must occur in the absence of vertical genes, as John Niederhauser discovered in 1954. Robinson (1987) discussed procedures for breeding for horizontal resistance against disease. These include: look for moderate, but not strong, resistance in parents; nullify vertical resistance genes (use virulent races of the pathogen); use population breeding methods (recurrent selection); breeding must be "holistic" - against all im¬ portant diseases, in the target environments; care must be taken to detect "escapes" (susceptible types); select the most resistant off-spring - it is the relative difference in disease levels that is important in early generations. Stable agriculture may depend on the development of stable disease resistance in crop plants. Population breed¬ ing methods may provide the means to develop stable resistance. gradually from cycle to cycle, with some 1992 parents out-yielding the best of the 1984 parents (Fig 7 centre). Improvements were made to both disease resistance and yield at the same time, in the same breeding material, without adding any new genes. There was no yield pen¬ alty for improving disease resistance. After completing two cycles of recurrent selection in lupins, the following improvements were evident; com¬ pared with control cv Danja, brown spot resistance in¬ creased 10-14% per cycle, yield increased 5-6% per cycle, Phomopsis resistance increased 30-50% per cycle (this was an added bonus, and unexpected), and average seed alkaloid levels also dropped markedly (Fig 7 bottom). New varieties of lupins in Australia must be lower than cv Danja in seed alkaloids. As a result of the first cycle of recurrent selection, a new lupin variety with a moderate level of resistance to Brown spot resistance in lupins: a case his¬ tory of recurrent selection Recurrent selection is described in lupins in YJA, where closed populations of lupins have been bred with simultaneous improvements in yield, quality and disease resistance. I have adopted recurrent selection as one of my breed¬ ing strategies in lupins. Closed populations undergo early generation selection (as F2-derived lines) for yield, disease resistance and quality. Parents are crossed in all possible combinations (a diallel cross) to begin cycle 1 (Fig 6). After quickly proceeding through early generations, including over-summer generations where possible, F,-derived lines are tested for yield, disease re¬ sistance and quality in replicated field trials in the sec¬ ond and third year. The best of these are selected as parents to begin the next cycle of recurrent selection, and are reselected for further testing and possible release as new varieties. I have bred lupins for resistance to brown spot by recurrent selection. Brown leaf lesions cause the leaves to fall off prematurely, and if seedlings are attacked early in the season they may be killed. Yield is often reduced by brown spot disease, and it is present in nearly all lupin crops. Improvements in resistance to brown spot have oc¬ curred in a closed population during three cycles of re¬ current selection (Fig 7). The vertical axis represents de¬ foliation due to brown spot, measured relative to a con¬ trol variety Danja which is set at 100%. Along the hori¬ zontal axis are the parents used to begin the first, second and third cycle of recurrent selection. All parents were grown together in the same experiment repeated at sev¬ eral sites in WA from 1990 to 1992. There was a significant decrease in defoliation due to brown spot in the parents of cycles 1 (crossed in 1984), 2 (1988) and 3 (1992) (Fig 7 top). Even more importantly, the best parents in 1992 were more resistant than any of the parents crossed in 1984. Without adding any new resistance genes, resistance is now stronger than in the original parents. The lines were also selected for yield and other char¬ acters during the selection process. Yield increased 120 o .2 I 0) Q 100 80 60 40 20 0 (1984) (1988) (1992) Parents Figure 7. Improvements in resistance to brown spot (top), grain yield (centre) and seed alkaloid concentration (bottom) in par¬ ents of the first three cycles of recurrent selection in a closed population of narrow-leafed lupins, compared with control va¬ riety Danja (bars represent the range from highest to lowest parent; letters represent significance of difference between cycles at P = 0.05). Source: Cowling (1994). 191 Journal of the Royal Society of Western Australia, 79(3), September 1996 brown spot (cv. Myallie) was released in 1995 (Anon. 1996). Myallie also has strong Phomopsis resistance, competitive yield (especially in low rainfall areas of WA where brown spot is most damaging to lupins) and low seed alkaloids. These results are significant on a global scale, because very few breeding programs of self-polli¬ nating crops have released commercial varieties from recurrent selection programs. Until recently, recurrent selection has been restricted to specific research projects (usually PhD research projects) or to breeding of cross- pollinating forage crops. The future of plant breeding and stable agriculture There are many strategies that plant breeders can adopt to improve their chances of contributing to the long-term stability of agriculture. What strategies should plant breeders adopt in order to contribute to stable agriculture in the future? 1. Diversify the gene pool: Invest in the future - create diverse populations in addition to the elite material in the program. Add new genes from new varieties bred elsewhere, or use wild types or landraces from genetic resource collections. Cross into elite lines, and reselect for the quality or adaptation that is required. 2. Introduce exotic genes: interspecific crosses, mutation or genetic engineering. Any source of germplasm is potentially valuable. However, for disease resistance, be¬ ware of major genes that may not be durable, or that prevent progress in breeding for polygenic resistance. 3. Use population breeding methods: As demonstrated here, population breeding methods have contributed to increases in disease resistance, yield and quality simulta¬ neously. Disease resistance developed by these methods is usually polygenic and durable. Crop plants have been evolving since the beginnings of agriculture 10,000 years ago. We may be witnessing a period of "punctuated equilibrium" in the evolution of crop plants, as proposed for evolution in nature (Gould & Eldredge 1993). Wheat and maize have evolved away from their wild relatives in the past 10,000 years due to human intervention (Duvick 1996). Lupins are just be¬ ginning to do so. In the Mediterranean region, L. albus var. albus has been cultivated for thousands of years and is relatively remote from its wild ancestor L. albus var. graecus in morphology, colour and seed size (Gladstones 1974). Plant breeders therefore have a great responsibility to lay strong foundations for the future evolution of crop species. The genetic diversity that is now created will become the future gene pool of several new man-made species, which, like wheat or maize, are so remote from their wild ancestors that little introgression from wild relatives will be possible. I believe that stable agriculture may be achieved if crop rotations are profitable and adapted crop species and varieties are available for growing on each soil and in each climate. Land conservation measures will be an essential component of stability, and another essential component will be the breeding of crop plants with du¬ rable disease resistance, low pesticide requirement, high yield and quality, and adaptation to regional environ¬ ments. Regional adaptation will become more and more important, and the universal crop variety will become a thing of the past. Barley in WA is bred in different gene pools for high and low rainfall environments - a similar situation is likely to develop for lupins over the next several decades. Breeding programs must have short-term and long¬ term goals. Given the current push towards privatisation at all levels of Government, it is important to ask: who will take responsibility for the long-term maintenance of genetic diversity in crop breeding? I have indicated that diversity is essential for long-term breeding progress. In Europe and the USA, very few private breeding pro¬ grams have taken on this role in the case of the common bean, and the gap between the identification of useful characters in exotic germplasm and the transfer to culti- vars has widened (Silbemagel 10 m s'1) caused frequent toppling of the instrument station and it was not possible to proceed with the measure¬ ments during the height of the sea breeze. The hydrody¬ namic measurements were made in the surf zone under the influence of breaking waves. The second field survey was carried out over six days in March 1995 and in¬ cluded three moderately-strong sea breezes (wind veloci¬ ties 5-7 m s'1). Hydrodynamic data were collected under shoaling waves just prior to breaking. Results First field survey Time series of wind speed and direction measured during the first field survey are typical of a sea breeze cycle characterized by weak offshore winds in the morn¬ ing and early afternoon, and a strong sea breeze starting in the afternoon and continuing into the evening (Fig 2). Wind velocities associated with the land breeze were less than 5 m s'1, whereas they frequently exceeded 10 m s'1 during the sea breeze. Such wind speeds are above average, but commonly occur in the summer months on the Perth Metropolitan coastline. The sea breeze started at 14:45 hrs and the change in wind speed and direction was almost instantaneous. The direction of the sea breeze was consistently from the south (180°) and hence the sea breeze was blowing parallel to the shoreline. The change in the wind climate is reflected in the incident wave field, nearshore currents and suspended sediment concentrations (Fig 2). Prior to the sea breeze, small-amplitude swell with significant wave heights of 0.4 m and zero-upcrossing periods of 7-8 s prevailed. Mean cross-shore currents were negligible (< 0.05 m s1) and mean longshore currents flowed in the northward direction with velocities less than 0.1 m s1. The mean suspended sediment concentration at a distance of 0.275 m above the bed was about 1 g L 1 and sediment was only suspended during the passage of large waves in wave groups. The onset of the sea breeze induced almost immediate changes in the nearshore hydrody¬ namics. The wave height increased, reaching 0.7 m at the end of the field survey. The wave period decreased and assumed a constant value of 4 s within one hour of the start of the sea breeze. Offshore-directed cross-shore currents in the surf zone rapidly increased in strength to 0.16ms1 and then fluctuated around 0.12ms1. The northerly longshore current progressively increased in strength up to 1 m s 1 and was still increasing when the survey was abandoned. Sediment was continuously sus¬ pended by the waves and the amount of suspended sedi¬ ment in the water column increased seven- fold to 7 g L L A crude estimate of the longshore suspended sedi¬ ment transport rate was obtained by multiplying the longshore current velocity by the suspended sediment concentration and integrating the product over the water column and across the surf zone. Before the start of the sea breeze, the northward transport rate was approxi- 200 Journal of the Royal Society of Western Australia, 79(3), September 1996 Wind speed (m/s) Mean cross-shore current (m/s) Suspended sediment concentration (g/l) Wind direction (degrees from N) 200 150 100 50 10 5 0 1 0.5 0 200 100 0 1 Mean longshore current (m/s) Longshore transport rate (kg/s) 1 12 13 14 15 16 17 Time in hours Figure 2. Wind speed, wind direction, significant wave height, zero-crossing wave period, cross-shore current velocity, longshore current velocity, suspended sediment concentration measured 0.275 m above the bed and longshore suspended sediment transport averaged across the surf zone measured on City Beach (23/01/92). The start of the sea breeze is indicated by the dotted line. The sea breeze induces almost instantaneous changes in the nearshore hydrodynamic conditions Figure 3. Beachface profile of City Beach (23/01/92) before (solid line; 14:00 hrs) and at the height of the sea breeze (dashed line; 19:00 hrs). During the sea breeze, erosion takes place on the upper part of the beach, whereas accretion occurs low on the beach. The elevation is measured relative to approximately 1 m above mean sea level. mately 1 kg s1, but it increased during the sea breeze by two orders of magnitude to 100-200 kg s'1 (Fig 2). The changes in the hydrodynamic conditions caused by the sea breeze induced an adjustment of the beach morphology; up to 0.4 m of erosion occurred on the up¬ per part of the beach, whereas deposition took place on the lower part (Fig 3). The cusp morphology that was present on the beach prior to the sea breeze was com¬ pletely eradicated. Field observations indicated that the beach cusps reformed overnight, after the sea breeze had subsided. 201 Journal of the Royal Society of Western Australia, 79(3), September 1996 10 Wind speed (m/s) — ; - 1 - n - 1 - I — ; - 1 - ‘ sea breeze sea breeze ' sea breeze Wind direction 300 n — - - 1 _ L_j _ i _ i _ 1 _ i - 1 Zero-crossing wave period (s) Date Figure 4. Wind speed, wind direction, zero-crossing wave period and longshore current velocity measured on City Beach (from 12:00 hrs on 06/03/95 to 12:00 hrs on 09/03/95). The start of the sea breezes is indicated by the dotted lines. The effect of the sea breeze on the nearshore hydrodynamic conditions extend a considerable duration after the cessation of the sea breeze. Second field survey The second field survey showed essentially the same features as the first, but due to the smaller wind veloci¬ ties, the hydrodynamic changes induced by the sea breeze were less extreme. Around-the-clock measure¬ ments of three successive sea breeze cycles were collected and these enabled the investigation of the recovery period of the wave conditions after the cessation of the sea breeze (Fig 4). Prior to the sea breezes, offshore winds prevailed with speeds of 2-3 ms1. The start of the sea breezes is indicated by an abrupt change in the wind direction from east to south and an concomitant increase in the wind velocity. All three sea breezes persisted for approximately 7 hours and a relatively constant wind velocity of 5-7 m s'1 was maintained throughout. After the onset of the sea breeze, almost immediate changes occurred to the incoming wave field as indicated by a decrease in the wave period from 10 to 5 s and an increase in the longshore current velocity from insignificant to 0.15-0.2 ms1. Around 20:45 hrs, the sea breezes stopped. The wind direction gradually shifted back to the east and the wind velocities dropped below 4 m s'1. The wave period started increasing and the longshore current velocity started decreasing as soon as the sea breeze had subsided. Around 6:00 hrs, nine hours after the end of the sea breeze, the wave period and the longshore current velocity had reached pre-breeze levels. Energy spectra of the cross-shore current were com¬ puted to construct a three-dimensional time-frequency plot to illustrate the change in spectral signature over the second sea breeze cycle (Fig 5). The cross-shore current data were used, rather than the water surface elevation because they show more clearly the short-period wave energy caused by the sea breeze (frequency > 0.15 Hz). The long-period background swell was present in the time-frequency plot in the form of a linear ridge at frequency 0.07-0.1 Hz. The peak period associated with the swell energy was 12 s and remained relatively constant. After the onset of the sea breeze, wind wave energy started to emerge at the high-frequency end of the spectra, indicating peak wave periods of 2.5 s. During the sea breeze, the frequency associated with the wind waves decreased progressively, forming a curving ridge in the frequency-time plot. At the end of the sea breeze (21:00 hrs), the wind wave energy was concentrated around a frequency of 0.25 Hz, indicating a peak period of 4 s. After the sea breeze had subsided, the wind wave energy gradually decreased, but remained significant. 202 Journal of the Royal Society of Western Australia, 79(3), September 1996 Figure 5. Frequency- time plot of spectra of the cross-shore current velocity measured on City Beach (from 12:00 hrs on 07/03/95 to 12:00 hrs on 08/03/95). The background swell is represented by the linear ridge around 0.08 Hz, whereas the wind waves are indicated by the curving ridge at 0.15-0.4 Hz. The frequency associated with the wind wave energy decreased to 0.15 Hz, merging with the swell energy. Fifteen hours after the cessation of the sea breeze (12:00 hrs 08/03/93), there was still some wave energy present that was generated by the sea breeze in the nearshore zone. This implies that the effect of the sea breeze on the nearshore hydrodynamics may be continuous during the sea breeze season (summer). The three sea breezes monitored during the second field survey induced changes to the beach morphology (Fig 6). Pronounced beach cusp morphology remained present on the beach, but during the sea breeze, erosion took place on the cusp horn and accretion occurred in the cusp embayment. Consequently, the beach cusp mor¬ phology became less pronounced. No sediment exchange was observed between the upper part of the beach and the nearshore zone, since morphological changes involved a shore-parallel redistribution of sediment over the beach cusp system. Discussion The findings of two field surveys, aimed at investigat¬ ing the impact of sea breeze activity on nearshore processes, are similar to those of Sonu et al (1973), the only other study into sea breeze effects. The sea breeze results in; (1) an increase of the wave height, (2) a de¬ crease in the wave period, (3) an intensification of the nearshore currents, (4) an increase in suspended sedi¬ ment levels and suspended sediment transport, and (5) a modification of the nearshore morphology* Due to the predominantly longshore component of the sea breeze, the nearshore hydrodynamics are affected long after the sea breeze has ceased to blow. Sonu et al. (1973) refer to this as the "afterglow effect". Both the strength and con¬ sistency of the sea breeze, and the afterglow effect con¬ tribute to the important role that the sea breeze plays on the nearshore processes along the Perth Metropolitan coastline. The role of sea breeze activity in this region is particu¬ larly important because the sea breeze blows predomi¬ nantly in a shore-parallel direction. The seaward extent of the sea breeze is usually 50-100 km (Hsu 1988; Simpson 1995) and consequently, the generation of wind waves by an shore-normal sea breeze is limited by the fetch length, restricting both the height of the wind waves and the duration over which the waves influence 203 Journal of the Royal Society of Western Australia, 79(3), September 1996 Cross-shore distance (m) Cross-shore distance (m) Figure 6. Beachface profile of the cusp embayment and the cusp horn on City Beach (08/03/95) before (solid line; 14:00 hrs) and after (dashed line; 21:00 hrs) the sea breeze. During the sea breeze, erosion of the cusp horn is accompanied by accretion in the cusp embayment. The elevation is measured relative to approximate mean sea level. nearshore processes. For a shore-parallel blowing sea breeze, the fetch is virtually unlimited and given suffi¬ cient duration of the sea breeze, the wind wave field can attain its maximum energy level and develop into a " fully-developed" sea. In addition, wind waves may con¬ tinue to arrive at the coastline long after the sea breeze has subsided. Finally, the obliquely-incident wind waves and the longshore wind stress may induce strong longshore currents that affect littoral transport. The sea breeze induces a diurnal cycle of beach change by causing erosion of the upper part beach and/or planation of beach cusp morphology. These changes are reversible in that the beach is usually restored to its pre¬ breeze state after the cessation of the sea breeze. On the larger temporal and spatial scale, the dramatic increase in the longshore sediment transport caused by the sea breeze is important. Masselink & Pattiaratchi (in press) and Pattiaratchi et al. (in press) estimate that along the Perth Metropolitan coastline, the annual longshore transport driven by the sea breeze is approximately 100,000 m3. This estimate corresponds to calculations of sediment accumulation at the southern side of Trigg Is¬ land (Perth) at the end of the summer. It is apparent that the sea breeze must play a dominant role in the sediment budget of the coastline around Perth. For example, a significant proportion of the sediment required for shore¬ line progradation of cuspate forelands such as Becher Point and Rockingham (Woods & Searle 1983), Whitfords Cusp (Semeniuk & Searle 1986), Jurien (Woods 1983), and other types of coastal landforms such as tombolas and salients (Sanderson & Eliot 1996) is probably contributed by the littoral drift induced by sea breeze activity. Acknowledgments: The author would like to thank all the people that have assisted in the collection of the field data reported here, in particular I Eliot, B Hegge and C Pattiaratchi. Comments by the two anonymous reviewers are also acknowledged, This paper has Centre for Water Research reference ED1115GM. References Clarke D J & Eliot I G 1983 Mean sea-level and beach-width variation at Scarborough, Western Australia. Marine Geology 51:251-267. Davies J L 1980 Geographical Variation in Coastal Development (2"d Edition). Longmans, London. Department of Defence 1990 Australian National Tide Table 1990. Australian Hydrographic Publication 11, Canberra. Eliot I G, Clarke D J & Rhodes A 1983 Beach width variation at Scarborough, Western Australia. Journal of the Royal Society of Western Australia 65:153-158. Gentilli J 1972 Australian Climate Patterns. Nelson's Australasian Paperbacks, Sydney. 204 Journal of the Royal Society of Western Australia, 79(3), September 1996 Hegge B J, Eliot I G & Hsu J 1996 Sheltered sandy beaches of southwestern Australia. Journal of Coastal Research, 12, 748- 760. Hounam C E 1945 The sea breeze in Perth. Weather Develop¬ ments and Research Bulletin 3:20-55. Hsu S A 1988 Coastal Meteorology. Academic Press, San Diego. Inman D L & Filloux J 1960 Beach cycles related to tide and local wind regime. Journal of Geology 68:225-231. Kempin E T 1953 Beach sand movements at Cottlesloe, Western Australia. Journal of the Royal Society of Western Australia 37:35-60. Masselink G & Pattiaratchi C in press The effect of sea breeze on beach morphology, surf zone hydrodynamics and sediment resuspension. Marine Geology. Pattiaratchi C, Hegge B J, Gould J & Eliot I G in press Impact of sea breeze activity on nearshore and foreshore processes in Southwestern Australia. Continental Shelf Research. Riedel H P & Trajer F L 1978 Analysis of five years of wave data, Cockburn Sound. Proceedings of the 4th Australian Conference on Coastal and Ocean Engineering, Institution of Engineers Australia, Adelaide, 143-167. Sanderson P G & Eliot I G 1996 Shoreline salients, cuspate forelands and tombolos on the coast of Western Australia. Journal of Coastal Research, 12, 761-773. Searle D J & Semeniuk V 1985 The natural sectors of the inner Rottnest Shelf coast adjoining the Swan Coastal Plain. Journal of the Royal Society of Western Australia 67:116-136. Semeniuk V & Searle D J 1986 The Whitfords Cusp - its geomor¬ phology, stratigraphy and age structure. Journal of the Royal Society of Western Australia 68:29-36. Silvester R 1961 Beach erosion at Cottlesloe, W.A.. Civil Engi¬ neering Transactions, March:27-33. Simpson J E 1994 Sea Breeze and Local Wind. Cambridge Uni¬ versity Press, Cambridge. Sonu C J, Murray S P, Hsu S A, Suhayda J N & Waddell E 1973 Sea breeze and coastal processes. EOS, Transactions Ameri¬ can Geophysical Union 54:820-833. Steedman R K 1977 Mullaloo Marina: Environmental investiga¬ tions and physical oceanographical studies for the Shire of Wanneroo. Job No. 048, Steedman & Associates, Perth, un¬ published report. Woods P J 1983 Selecting a harbour site based on studies of coastal evolution at Jurien, Western Australia. Proceedings of the 6th Australian Conference on Coastal and Ocean Engi¬ neering, Institution of Engineers Australia, Gold Coast, 59- 63. Woods P & Searle D J 1983 Radiocarbon dating and Holocene history of the Beacher/Rockingham beach ridge plain, West Coast, Western Australia. Search 14:44-46. 205 Journal of the Royal Society of Western Australia, 79:207-210, 1996 A genetic perspective on the specific status of the Western Rock Lobster along the coast of Western Australia - Panulirus cygnus George 1962 or P. longipes A Milne-Edwards 1868? A P Thompson Department of Zoology, University of Western Australia, Nedlands WA 6907 Manuscript received July 1996 ; accepted October 1996 Abstract Panulirus longipes is widespread throughout the the Indo-west Pacific and P. cygnus is restricted to the coast of Western Australia. Controversy regarding their discreteness has continued since the early 1900's based on the lack of morphological differences and the potential for gene- flow between these species. Allozyme variation indicates a discreteness of P. cygnus and P. longipes along the coast of Western Australia. Unique alleles were found at the PGM locus in P. longipes. At 3 other loci (ESP, GPI, and MDH-2) the probabilities that these individuals came from a population which was genetically common to P. cygnus were extremely low (1.4 1CU, 4.9 1C U, and 2.2 10-6 respectively). Nei's (1978) unbiased measure of genetic identity separated P. cygnus from P. longipes at 0.759. The evidence supports genetic discreteness of the two species along the coast of Western Australia and thus supports the views of George (1962) that P. cygnus is a valid species. Introduction The Western Rock Lobster, Panulirus cygnus, inhabits the west coast of Australia approximately from Cape Leeuwin (34° 22’S) to North West Cape (21° 45’S). Seven other species of rock lobster also inhabit the west coast of Australia, but this zone is mainly occupied by P. cygnus with only minimal overlap with other species. ]asus novaehollandiae (Holthuis 1963, cited by Holthuis 1991) is sympatric with P. cygnus at the southern end of the range, while at the northern end there is an overlap with several tropical species, P. versicolor (Latreille 1804, cited by Holthuis 1991), P. ornatus (Fabricius 1798, cited by Holthuis 1991), and P. penicillatus (Olivier 1791, cited by Holthuis, 1991). Occasional specimens of P. longipes have been recorded in this zone. The two other species are P. polyphagus (Herbst 1793, cited by Holthuis 1991) which occurs in the muddy water near Derby, and P. homarus (Linnaeus 1758, cited by Holthuis 1991) which inhabits turbid waters off the Pilbara coast (see Gray 1992). Conjecture regarding the discreteness of P. cygnus has continued since 1912 (see Gray 1992). Some scientists believed that there were insufficient morphological differences (Glauert 1936; Sheard 1949; Chittleborough & Thomas 1969) as well as a lack of evidence of genetic isolation to warrant ranking P. cygnus as a distinct species. Rather, it was suggested that P. cygnus was part of a more widespread population of P. longipes, which is an Indo-west Pacific species (see Holthuis 1991). The subspecies P. longipes longipes is the western form, occurring from East Africa to Thailand, the Philippines and Indonesia. Occasional specimens of P. I longipes are found along the coast of Western Australia within the zone of P. cygnus. The eastern subspecies P.l. femoristriga inhabits Japan, the Moluccas, New Guinea, Eastern © Royal Society of Western Australia 1996 Australia, New Caledonia, and Polynesia (Holthuis 1991). Recognition of the western rock lobster as a distinct species has significant implications for the fishing industry. If it were part of the more widespread species, P. longipes, as suggested by Glauert (1936), Sheard (1949), and Chittleborough & Thomas (1969), then lobsters caught on the West Australian coast may have come from larvae drifting with ocean currents from elsewhere in the Indian and Pacific Oceans. Thus, protecting the breeding population on the coast of Western Australia may not be such a crucial management strategy. However, recognition of the western rock lobster as a discrete species P. cygnus, restricted to the waters off Western Australia, dictates that the fishery be managed accordingly. Allozyme electrophoresis is a useful tool for distinguishing populations and species. One of the implications of a discrete population is that there is no gene-flow between it and other populations. Absence of gene-flow allows genetic differences to accumulate over generations through natural selection, random genetic drift, and mutation (Wallace 1981). This can result in speciation. Usually, distinct populations and/or species will show genetic differences (Altukov 1981; Johnson et al 1993; Shaklee 1983). Intraspecific populations are distinguished in the most part by different frequencies of shared alleles, while species are distinguished by fixed allelic differences (Altukov 1981). Thompson et al. (1996) found no evidence of latitudinal genetic variation among populations of the western rock lobster from Garden Is¬ land to Shark Bay. The average among 5 sites over the 9 polymorphic loci studied was a very low 0.0002, which is consistent with current interpretations that the western rock lobster is a single panmictic population (see Thomp¬ son et al 1996). The aim of this study was to investigate the genetic discreteness of P. cygnus and P. longipes along the coast of Western Australia. 207 Journal of the Royal Society of Western Australia, 79(3), September 1996 Table 1 Enzymes and electrophoretic buffers used to study genetic variation in P. cygnus and P. longipes. Buffer recipes are given in Hillis & Moritz (1990). Enzyme E.C. number Locus abbreviation Buffer Arginine phosphokinase 2. 7.3.3 APK TC8 Esterase 3.1.1.- EST TC8 Glucose phosphate isomerase 5. 3.1. 9 GPl LiOH Leucylproline peptidase 3.4.-.- LPP TC8 Leucylglycylglycine peptidase 3.4.1.3 LGG TC8 Malate dehydrogenase 1.1.1.37 MDH-1, MDH-2 TM Mannosephosphate isomerase 6-Phosphoglucona te 5.3.1.8 MPI TC8 dehydrogenase 1.1.1.44 6PGD TM Phosphoglucomutase 2.7.5. 1 PGM TM Valylleucine peptidase 3.4.-.- VIP TC8 Materials and Methods A leg from each of 455 adult rock lobsters (P. cygnus) was collected during 1994. Rock lobsters were collected at commercial depots at 5 locations; Garden Island (n=84), Two Rocks (n=81), Geraldton (n=97). Southern Abrolhos (n=99) and Shark Bay (n=94). These samples were frozen immediately in liquid nitrogen before being transferred to a freezer at -70°C. Tissue samples (muscle) from three P. longipes longipes were collected during 1995, also from a commercial depot. The P. longipes specimens were characterised by the presence of pale spots on the legs. These specimens were caught off Shark Bay, within the zone of P. cygnus. They were frozen immediately in liquid nitrogen and later stored at -70°C. Ten enzyme systems representing 11 loci were scored in both P. cygnus and P. longipes (Table 1). Electrophoretic 1 2 3 4 5 6 7 8 Figure 1. Isozyme formation of PGM showing the fixed allelic differences of the P. longipes specimens (positions 2-A) compared to P. cygnus (positions 1, 5-8). Individual genotypes from left to right are: 22, 34, 34, 33, 22, 22, 22, 22. procedures are described in more detail by Thompson et al (1996). Expected genotypic frequencies at each locus were calculated using BIOSYS-1.7 (Swofford & Selander 1981). The genotypic frequencies at each locus from the pooled P. cygnus population (since Thompson et al. 1996 found no evidence of genetic subdivision) were used as prob¬ ability estimates for the observed genotypic frequencies of the P. longipes individuals. The expected probability of a P. longipes individual having a particular genotype was equal to the expected frequency of that genotype occur¬ ring in the P. cygnus population. The genetic differentiation among populations was examined using BIOSYS-1.7 (Swofford & Selander 1981). Clustering was performed based on the methods of Nei's (1978) unbiased genetic identity. Results One locus, PGM, showed unique alleles and genotypes between the 2 groups (Figure 1, Table 2). At ten loci, the 3 P. longipes specimens had alleles common genetic similarity .67 .73 .80 .87 .93 1.00 I - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 — P. cygnus (Shark Bay) — P. cygnus (Abrolhos) — P. cygnus (Dongara) P. cygnus (Two Rocks) — P. cygnus (Garden Island) - P. longipes (Shark Bay) l - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 .37 .30 -22 .15 .07 0 genetic distance Figure 2. Dendrogram based on Nei's (1978) unbiased genetic identities of P. cygnus and P. longipes caught off the coast of Western Australia. 208 Journal of the Royal Society of Western Australia, 79(3), September 1996 Table 2 Genotypic classes of P. cygnus with their associated observed frequencies, and the individual genotypes scored for each P. longipes specimen with the associated probabilities that these animals came from a population that was genetically common to P. cygnus. N is sample size. P. cygnus (N = 455) P. longipes (N = 3) Locus Genotypes Frequency Locus Genotypes Probability APK 2-2 1 APK 2-2, 2-2, 2-2 1 EST 1-1 0.024 EST 1-1, 1-1, 1-1 1.4 10-4 1-2 0.229 1-3 0.747 GPI 1-1 2.198 10-3 GPI 2-2, 2-2, 2-2 4.9 10-3 1-2 4.396 10-3 1-3 8.791 10-3 2-2 0.1692 2-3 0.4132 3-3 0.4022 IDH 2-2 1 IDH 2-2, 2-2, 2-2 1 LGG 2-2 1 LGG 2-2, 2-2, 2-2 1 LPP 1-2 0.0202 LPP 2-2, 2-2, 2-2 0.9415 2-2 0.9801 MDH-1 1-2 6.593 10-3 MDH-1 2-2, 2-2, 2-2 0.9609 2-2 0.9868 2-3 6.593 10'3 MDH-2 1-1 1.32 10-2 MDH-2 1-1, 1-1, 1-1 2.2 104 1-2 0.1165 2-2 0.8615 2-3 8.80 10-3 MPI 2-2 0.9956 MPI 2-2, 2-2, 2-2 0.9869 2-3 4.3956 10-3 PGM 1-2 1.538 10-2 PGM 3-3, 3^, 3-4 0 2-2 0.9846 VLP 1-2 1.32 x Id2 VLP 2-2, 2-2, 2-2 0.8683 2-2 0.9540 2-3 3.330 10'3 to both species. However, the probabilities of them com¬ ing from a population which was genetically linked to P. cygnus were extremely small at the EST, GPI, and MDH-2 loci (Table 3). At the other 7 loci ( APK , IDH, LGG, LPP, MDH-1, MPI, and VLP) the 3 P. longipes specimens shared alleles, with no evidence of differences between the 2 groups. Nei's (1978) genetic identity cluster analysis clearly separted the P. longipes specimens from all P. cygnus populations at 0.759 (Figure 2). Discussion Even though only 3 individual P. longipes were used in this study, it is clear that these animals came from a population that is not genetically linked (i.e. there is no gene-flow) to P. cygnus . PGM was the definitive locus for this conclusion with completely unique alleles and genotypes observed in all 3 P. longipes specimens. Even if there was only a small amount of gene-flow between the 2 groups, one would expect rare genotypes (or at least alleles) to occur in 455 animals. This was not the case. Furthermore, Nei's (1978) unbiased estimate of genetic identity separated the 2 species at 0.759. In a study of the freshwater crayfish genus Cherax, 6 species were separated with genetic identities ranging from 0.92 to 0.53 (Austin & Knott 1996). The electrophoretic results here provide the evidence of genetic isolation between P. cygnus and P. longipes along the coast of Western Australia that was lacking at the time of Chittleborough & Thomas (1969). Phillips (1981) reported that larvae of P. cygnus are transported offshore through a combination of behavioural adaptations resulting in the early phyllosoma being at the surface at nights, and surface currents (summer pattern of offshore wind drift) during the early stage of their life-cycle. Mid to late stages are entrained within mesoscale eddies and gyres up to 1000 km offshore. Final stages are returned by a change in behaviour (spending more time at greater depth) in conjunction with the deeper onshore flow. Pollock (1992) presumed from oceanographic characteristics that the larval pool of P. cygnus off the coast of Western Australia resulted in separation and ultimately in speciation of this group. The genetic evidence supports the notion that P. cygnus is genetically isolated from P. longipes along the coast of Western Australia. 209 Journal of the Royal Society of Western Australia, 79(3), September 1996 However, larval transport processes occasionally result in larvae being moved long distances. The P. longipes collected for this study presumably came to the coast of Western Australia via larval transport rather than by juvenile migration. This shows the possibility for gene-flow between P. cygnus and P. longipes. Phillips et al. (1980) proposed that near-adult specimens of P. cygnus with pale spots on their legs (characteristic of specimens of P. longipes from type locality of Zanzibar) were in fact hybrids due to gene-flow from the north. More likely, these individuals were foreign larvae which managed to metamorphose to become juveniles. The foreign P. longipes specimens negate the hypothesis of Pollock (1992) that larvae from each species only metamorphose into the settling puerulus stage when they experience physical, and/or chemical cues native to endemic environments. It is more probable that species are separated due to physical isolation (i.e. entrainment of larval pools) rather than due to the failure of foreign larvae to metamorphose. Larvae of P. longipes may occasionally be transported and even settle in the waters off Western Australia, but the genetic evidence supports a lack of significant transfer of genes between these groups. Acknowledgements: Thanks to the Fisheries department of Western Australia for facilitating this project, and to M Johnson and K Whitaker for valuable comments on the manuscript. This project was funded by the Zoology Department of the University of Western Australia. References Altukov YP 1981 The stock concept from the viewpoint of population genetics. Canadian Journal of Fisheries and Aquatic Sciences 38:1523-38. Austin CM and Knott B 1996 Systematics of the freshwater crayfish genus Cherax Erichson (Decapoda: Parastacidae) in south-western Australia: electrophoretic, morphological and habitat variation. Australian Journal of Zoology 44:223-58. Chittleborough RG & Thomas LR 1969 Larval ecology of the Western Australian marine crayfish, with notes upon other panulirid larvae from the Eastern Indian Ocean. Australian Journal of Marine and Freshwater Research 20:199-223. George RW 1962 Description of Panulirus cygnus sp. nov., the commercial crayfish (or spiny lobster) of Western Australia. Proceedings of the Royal Society of Western Australia 45:100-110. Glauert L 1936 Exhibits at the Royal Society of Western Australia 8/6/1936, Panulirus. Journal of the Royal Society of Western Australia 22:xxx. Gray HS 1992 The Western Rock Lobster, Panulirus cygnus Book 1: A Natural History. Westralian Books, Geraldton, Western Australia. Hillis DM and Moritz C 1990 Molecular Systematics. Sinauer Associates, Sunderland, Massachusetts. Holthuis LB, 1991 FAO Species Catalogue. Vol. 13 Marine Lobsters of the World. FAO Fisheries Synopsis 125:292pp. Johnson MS, Hebbert DR & Moran MJ 1993 Genetic analysis of populations of North-western Australian fish species. Australian Journal of Marine and Freshwater Research 44:673-685. Nei M 1978 Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Phillips BF 1981 The circulation of the southeastern Indian Ocean and the planktonic life of the western rock lobster. Ocean Marine Annual Review 19:11-39. Phillips BF, Morgan GR & Austin CM 1989 Synopsis of biological data on the Western Rock Lobster Panulirus cygnus (George, 1962). FAO Fisheries Synopsis 128:64pp. Pollock DE 1992 Paleoceanography and speciation in the spiny lobster genus Panulirus in the Indo-pacific. Bulletin of Marine Sciences 51:135-146. Shaklee JG 1983 The utilization of isozymes as gene markers in fisheries management and conservation. Isozymes. Current Topics in Biological and Medical Research 11:213-247. Sheard K 1949 Marine crayfishes of Western Australia. CSIRO Bulletin (Australia) 247:1-45. Swofford DL & Selander R B 1981 BIOSYS-1: a FORTRAN program for the comprehensive analysis of electrophoretic data in population genetics and systematics. Journal of Heredity 72:281-283. Thompson AP, Hanley JR & Johnson MS 1996 The genetic structure of the Western Rock Lobster Panulirus cygnus with the benefit of hindsight. Australian Journal of Marine and Freshwater Research 47: 889-896. Wallace B 1981 Basic Population Genetics. University Press, New York. 210 Journal of the Royal Society of Western Australia, 79:211-216, 1996 Roost selection by the lesser long-eared bat, Nyctophilus geoffroyi, and the greater long-eared bat, N. major (Chiroptera: Vespertilionidae) in Banksia woodlands D J Hosken Department of Zoology, University of Western Australia, Nedlands WA 6907 Manuscript received April 1996 ; accepted July 1996 Abstract Radio telemetry was used to track eight Nyctophilus geoffroyi and three N. major to roosts in Banksia woodland, on the Swan Coastal Plain south of Perth, intermittently between January and August 1995. All roosts were found in large or dead trees and bats were never captured more than 1.2 km from their roosts. Twenty two roosts were identified in six species of trees. Both species of bat changed roosts regularly, and were always found to roost alone. N. geoffroyi showed a strong preference for roosts in dead Banksia trees, although they also roosted in Melaleuca trees during storms. The differential use of the two tree species by N. geoffroyi may relate to water harvesting differences between the two types of tree, and to temperature differences between roosts found in each tree species; roosts in Melaleuca trees stay dry but are much colder than roosts in Banksia trees. N. major were tracked to Eucalyptus rudis and Melaleuca rhaphiophylla trees, but only one roost was actually located. Stands of forest containing dead trees may be neccesary for the persistence of both species. Introduction Roost sites are a critical resource for bats, providing shelter, protection, mating and hibernation sites (Kunz 1982). It has been argued that roost availability is, or may become, a limiting resource for many bat species (Taylor & Savva 1988). While a number of studies have investigated the use of roosts by eastern Australian bats ( e.g . Taylor & Savva 1988; Lunney et al 1988, 1995), little is known about the roosts used by forest bats in Western Australia. This is particularly true for Nyctophilus geoffroyi , for which no detailed account of roost use has been published. The lesser long-eared bat, N. geoffroyi is a small (5.5- 8.0 g) vespertilionid bat. Its diet includes lepidopterans, coleopterans, hymenopterans, dipterans and orthopterans (Vestjens & Hall 1977). N. geoffroyi is known to forage by gleaning and is able to exploit prey generated noise, including acoustic signals, to locate prey (Grant 1991; Hosken et al. 1994). These bats are found throughout Australia, with the exception of Cape York Peninsula (Hall & Richards 1979), although their taxonomy may be more complex than is currently recognised (N L McKenzie pers. comm.; H Pamaby pers. comm.). N. geoffroyi is reported to roost in trees, in hollows and under bark, and also roosts in buildings (Lumsden & Bennett 1995; Reardon & Flavel 1991). These bats usually roost alone (Lumsden & Bennett 1995; L Lumsden pers. comm.) or in small maternity colonies of generally less than 30 individuals, although one colony of about 200 N. geoffroyi has been reported (Reardon & Flavel 1987). Less is known about the biology of N. major (also known as N. timoriensis). It is widely distributed but © Royal Society of Western Australia 1996 uncommon throughout southern Australia, although the presence of distinct geographic forms indicate that a species complex may be present (Pamaby 1995). It is about twice the size of N. geoffroyi, weighing between about 11 to 20 grams. N. major is thought to roost alone or in pairs in tree hollows, but even this is uncertain (Richards 1991). This study primarily aimed to investigate roost selection by N. geoffroyi in Banksia woodlands on the Swan Coastal Plain south of Perth. In addition, the fortuitous capture of three N. major provided an opportunity to investigate roosting in this species. Methods This study was carried out at the Harry Waring Marsupial Reserve, Wattleup (approximately 32° 15’ S, 115° 50' E) from January to August 1995. The reserve is small, approximately 250 hectares, and is predominantly low open woodland on Bibra Sands. It includes a mixture of Eucalyptus rudis, E . gomphocephala, E. marginata and Melaleuca preissiana and M. rhaphiophylla, but is dominated by Banksia woodlands ( Banksia attenuata and B. menziesii), with a variable understory (for further description see Hosken & O'Shea 1994). Bats were captured in mist-nets set in woodlands and were fitted with small radio-transmitters (Titley Electronics) with 8-12 cm flexible wire antennas and an expected battery life of about 8 days. Transmitters weighed between 0.7 and 1.1 g, which represents 11-17% of the mean body weight of the N. geoffroyi captured during this study (less than 9% of the body weight of N. major). This is less than the weight of the two foetuses that female N. geoffroyi carry during late pregnancy (unpublished data), and is proportionally less than the mass of transmitters carried by bats in other studies (e.g. Lunney et al. 1995; 12-19% of body mass). In trials with 211 Journal of the Royal Society of Western Australia, 79(3), September 1996 three captive N. geoffroyi, transmitters did not appear to adversly affect the bats behaviour or mobility, and transmitters were shed in 6-15 days. Bats fitted with transmitters in this study did not lose weight over the 4— 6 days that they carried transmitters, which also indicates that transmitters had no obvious adverse effects. Transmitters were attached to the dorsal fur between the shoulder blades using rapid-set cyanoacrylate glue, with the antenna projecting posteriorly. Diurnal roosts were then located by radiotelemetry, with radio-signals from transmitters received using a receiver and direc¬ tional 'h-frame' antenna (Biotelemetry, SA). Roost loca¬ tion was usually confirmed by sighting the bat or the transmitter antenna. The following roost characteristics were recorded; tree species and whether it was alive or dead, the diameter at breast height (DBH), the roost height, the direction it faced and distance from last sighting of the bat. Distances were either measured directly or calculated using a map marked with 100 x 100m grids. Later, when roosts were vacant, temperature fluctuations inside and outside each roost were recorded during the course of a day. Each roost was visited once each hour from 9am till 5pm and the temperature in the roost (Tr) and the ambient temperature (Ta) just outside the roost were recorded using a Radio Spares type-K thermocouple meter and thermocouple. Statistics were mainly performed using the Statview + SE statistical package, and data are presented as means with ± standard error, unless stated otherwise. Results Eight (four male, four female) N. geoffroyi and three (two male, one female) N. major were tracked over a total of 42 days during 1995. N. major were tracked in January and February, while the N. geoffroyi were tracked intermittently from April to August. During this time N. geoffroyi begins mating; the sperm is stored until about October, when pregnancy is initiated (Hosken unpub¬ lished data). The N. major were each tracked for four days and nights, and six roost trees were identified. Two N. geoffroyi lost transmitters and one was killed by an owl on the first night. The other five were tracked for Table 1 The tree species in which bats were found to roost. (Dead or alive refers to the tree) Tree spp N. major N. geoffroyi Banksia attenuata 0 6 (all dead)’ Banksia menzesii 0 2 (all dead)’ Eucalyptus rudis 3 (2 alive but burnt, 1 dead) 1 (dead) Eucalyptus marginata 0 ' 1 (dead) Melaleuca pressiani 0 4 (all alive) Melaleuca rhaphiophylla 3 (all alive) 0 Two other roosts were located in dead banksia; however, the species could not be determined. Table 2 The diameter at breast height (DBH) and roost height of trees in which N. geoffroyi roosts were found (mean ± SE). Tree DBH roost height Banksia attenuata 0.38m (±0.1) 1.85m (±0.55) Banksia menzesii 0.25m (±0.13) 0.85m(±0.15) Eucalyptus rudis 0.38m 2.9m Eucalyptus marginata 1.19m 5.1m Melaleuca pressiani 0.9m(±0.1) 2.13m (±0.52) four to six days each, allowing 15 roosts to be identified. An additional N. geoffroyi roost was found while search¬ ing for a N. major. N. geoffroyi changed roosts frequently (mean number of days that each roost was occupied was 1.13 ±0.15) and were predominantly found roosting in dead Banksia trees, under bark that had come away from the tree trunk to form a loose fitting sleeve. (Fig 1, Table 1). N. major were found to roost in Paperbark trees or Flooded gums and occupied each tree for 1.83 ±0.48 days. N. geoffroyi tended to move roosts on a daily basis, except for one bat which was found in the same roost on 3 consecutive days during a storm; this roost was in a Paperbark tree (Melaleuca preissiana). The only times these trees were used as roosts by N. geoffroyi was during storms or when it was raining (4 bats on 6 days; 6 out of 6 occasions; sign test P = 0.016). Roosts were always un¬ der bark. Interestingly, these were the only live trees in which N. geoffroyi roosts were found, and no bats were found to be roosting in Banksia trees when there had been rain overnight. N. geoffroyi were located in roosts on 16 occasions and were always alone. Roosts tended to be close to the ground (1.93 ±0.36m), but average roost height varied with the species of tree as did the DBH of trees in which bats roosted (Table 2). The majority of roosts were either on the north or west face of trees or were in direct afternoon sun (9 of 14; note that the roost used on three consecutive days was only counted once) Only one of six N. major roosts was located. This was in a fissure within a branch of a burnt-out, dead E. rudis. This branch was shared with an unmarked N. geoffroyi , although these bats were not in the same fissure. Other roosts, in large Melaleuca rhaphiophylla trees, were inacessable, but two N. major were located in the same trees for three and four consecutive days respectively. While N. geoffroyi were sometimes captured a substantial distance from where they were subsequently found to roost (850-1200 m maximum), the roosts used by an individual were generally much closer to each other (mean distance roost to roost = 194 ±57 m) suggesting some area fidelity (Fig 2). A Student's t-test comparison of the mean distance between the point of capture and the roost location on the day after capture, and the mean distances moved between roosts, revealed that the difference was statistically significant (unpaired t^ value = 3.33, P = 0.0046). There was no significant difference between the sexes in the distances moved between roosts or in the distances between point of capture and subsequent roost location (two tailed unpaired Student's t-test comparison, P > 0.32 for each comparison). 212 Journal of the Royal Society of Western Australia, 79(3), September 1996 Figure 1. Typical N. geoffroyi roost found in dead Banksia tree. Roost (marked with arrow) was 0.8 m above ground. Width of central branch was 0.32 m. There was a significant and positive relationship be¬ tween Ta and Tr in both Banksia and Melaleuca trees (r2=0.78, f = 199.6, p = 0.0001 and r2 = 0.81, f213 = 55.9, P = 0.0001 respectively). However, the slope o^ the line describing the relationship between Ta and Tr for Banksia roosts was significantly greater than that for roosts in Melaleuca trees (test of slopes: t67 = 5.6, P < 0.001) and at Tas above about 16° C the temperatures in Banksia roosts was always greater than those in Melaleuca. In addition, in Banksia trees Tr typically approached Ta by about 1200 and by the time final temperature measurements were taken (between 1600 and 1700) five of six roosts in Bank¬ sia had temperatures that exceeded ambient by about 1.0° C (range 0.3-1. 7° C; Fig 3). The temperature of roosts in Melaleuca trees never exceeded Ta during the measure¬ ment periods (Fig 3). Discussion Nyctophilus geoffroyi roost in trees and change roost regularly. This habit has been recorded in a number of other Australian forest bats (L Lumsden, pers. comm.; Lunney et al. 1988, 1995) and is consistant with the proposal that roost fidelity is directly related to roost permanence (Lewis 1995). The same may also be true of N. major. In a study of roost use by bats in Tasmania, Taylor & Savva (1988) noted that N. geoffroyi change roosts fre¬ quently. They located two roosts under bark, one in a narrow cavity in a tree bole, and one in a fissure. How¬ ever, unlike this study, N. geoffroyi were only found roosting in a dead tree once. In remnant vegetation around farmland in Victoria, N. geoffroyi were found to roost disproportionately in dead trees, a finding similar to that reported here (L Lumsden, pers. comm.). The preference that N. geoffroyi displayed for dead Banksia trees over live Melaleuca trees, during this study, probably relates to different winter thermal characteristics of roosts found in the two tree species. By choosing Banksia roosts, N. geoffroyi are exposed to temperatures above ambient during the late afternoon and this would reduce the costs of arousal from torpor prior to foraging. It was noted that many roosts in Banksia trees were facing the afternoon sun and this may explain why Tr was higher than Ta in the afternoon in Banksia tree roosts. In addition, dead Ba?iksia trees were dark coloured, which presumably aids heating further. If the same thermal characteristics are found during summer, it is reasonable to expect Melaleuca trees to be the prefered roost. With their apparently superior insulation, Melaleuca would be cooler than Banksia roosts; low temperatures enable bats to lower their body temperature, leading to water and energy savings ( e.g . Hosken & Withers in press). That N. major were found to 213 Journal of the Royal Society of Western Australia, 79(3), September 1996 N I THE HARRY WARING MARSUPIAL RESERVE Figure 2. Map of the Harry Waring Marsupial Reserve showing the roost movement patterns of a female N. geoffroyi. x = site of capture, o - roost locations, dotted lines show bat movement between roosts. frequently use these trees during summer appears to support this reasoning, but further investigation is re¬ quired. No bats were found roosting in live Banksia trees. This is probably due to the fact that the bark is not loose on live trees and further indicates that N. geoffroyi are selecting specific roost sites. The use of Melaleuca roosts during rainy periods probably relates to the fact that these trees are not water harvesting, and roosts under the multilayered bark insulation stay dry, while roosts in dead Ba7iksia trees were often damp during and after rain. In addition, storms appear to exact a heavier toll on dead Banksia trees when compared with Meleleuca trees; four dead Banksia trees were found across tracks after storms during the course of the study. This also indicates that the Banksia roosts are relatively ephemeral, which makes the reliance on one roost unprofitable and possibly prompts frequent roost movement. All trees that were found to contain roosts during this study were tagged and longer term observation will reveal the longevity of each roost. In this study, N. geoffroyi were found to roost alone. This is consistent with other published reports (Lumsden &l Bennett 1995, Taylor & Savva 1988). However, Taylor &c Savva (1988) also found three colonies containing three, 12 and 23 N. geoffroyi. The two largest groups were maternity colonies. Since this study was carried out during autumn and winter, no maternity colonies were encountered and it appears that these bats, at least in Ba?iksia woodland, are solitary during the mating period which extends from about April to September (Hosken, unpublished data). As with N. gouldi (Lunney et al. 1988), N. geoffroyi appear to display fidelity to an area and the distances between successive roosts reported here are similar to those reported for other nyctophilines (Lunney et al. 1988, 1995). The largest distance between capture site and roost site for N. geoffroyi in this study was about 1200m. This is similar to the distances moved by male N. geoffroyi in Victoria (L Lumsden pers. comm.) but less than the 4800m reported by Taylor and Savva (1988) and the 6- 12km reported for female N. geoffroyi (L Lumsden pers. comm.). However, the comparatively small distances between capture site and roost site reported here are consistent with the flight morphology of N. geoffroyi which indicates that this species is not suited to long distance flight (Fullard et al. 1991). A similar finding is reported here for N. major. The only individual which regularly changed roosts during this study was found to move about the same distance as the N. geoffroyi. This N. major shed its transmitter at its initial capture site five days after capture. This was approximately 1200m from its last roost tree. N. major is reported to have flight char¬ acteristics similar to N. geoffroyi (Hall & Richards 1979), which suggests that long distance flight would also be energetically expensive for this species. 214 Journal of the Royal Society of Western Australia, 79(3), September 1996 TIME Figure 3. Hourly changes in the roost temperature of a typical Banksia roost (top) and a typical Melaleuca roost (bottom) plotted with Ta measured outside but adjacent to the roost. The use of only large mature or dead trees by both the bat species tracked during this study indicates that a ma¬ ture forest is essential for them. The finding that dead trees were the predominant roost used by N. geoffroyi and that these trees appear to be the main victims of winter storm indicates that the continual tree death is required to maintain the roosts needed by these small bats. Unfortunately, continued clearing on the Swan Coastal Plain may eventually threaten this continuity. Acknowledgments: I am grateful to numerous people who helped at various stages of this study, particularly A F Stucki and B Cooper. My thanks are extended to N McKenzie (Department of Conservation and Land Management, Perth, WA), H Pamaby (The Australian Museum, Sydney, NSW) and L Lumsden (Department of Conservation and Natural Resources, Heidelberg, VIC) for advice, comments and access to unpub¬ lished information. I also thank A Thompson, P C Withers, J E O'Shea and two anonymous referees who commented on earlier drafts of this paper. Bats were captured under permit numbers BB685 and SF1700 is¬ sued by the Department of Conservation and Land Managment. References Fullard J H, Koehler C, Surlykke A & McKenzie N L 1991 Echolocation ecology and flight morphology of insectivorous bats (Chiroptera) in south-western Australia. Australian Journal of Zoology 39:427-438. Grant J D A 1991 Prey location by two Australian long-eared bats, Nyctophilus gouldi and N. geoffroyi. Australian Journal of Zoology 39:45-56. Hall L S & Richards G C 1979 Bats of eastern Australia. Queensland Museum, Brisbane, Booklet No. 12. Hosken D J & O'Shea ] E 1995 Falsistrellus mackenziei at Jandakot. The Western Australian Naturalist 19:351. Hosken D J &: Withers P C in press Temperature regulation and metabolism of an Australian bat, Chalinolobus gouldii (Chiroptera: Vespertilionidae) when euthermic and torpid. Journal of Comparative Physiology B: in press. Hosken D J, Bailey W J, O'Shea J E & Roberts J D 1994 Localisation of insect calls by the bat Nyctophilus geoffroyi (Chiroptera: Vespertilionidae): a laboratory study. Australian Journal of Zoology 42:177-184. 215 Journal of the Royal Society of Western Australia, 79(3), September 1996 Kunz T H 1982 The roosting ecology of bats. In: Ecology of Bats (ed TH Kunz) Plenum Press, New York, 1-55. Lewis S E 1995 Roost fidelity of bats: a review. Journal of Mammalogy 76:481^196. Lumsden L F & Bennett A F 1995 Lesser long-eared bat. In: Mammals of Victoria: distribution, ecology and conservation (ed PW Monkhorst) Oxford University Press, Melbourne, 184-186. Lunney D, Barker J, Priddel D & O'Connell M 1988 Roost selection by Gould's long-eared bat, Nyctophilus gouldi Tomes (Chiroptera: Vespertilionidae), in logged forest on the south coast of New South Wales. Australian Wildlife Research 15:375-384. Lunney D, Barker J, Leary T, Priddel D, Wheeler R, O'Connor P & Law B 1995 Roost selection by the north Queensland long¬ eared bat Nyctophilus bifax in littoral rainforest in the Iluka World Heritage Area, New South Wales. Australian Journal of Ecology 20:532-537. Pamaby, H 1995 Greater long-eared bat. In: The Mammals of Australia 2nd edition (ed R Strahan) Reed Books, Chatswood, 507-508. Reardon T B & Flavel S C 1987 A guide to the bats of South Australia. South Australian Museum, Adelaide. Richards G C 1991 Greater long-eared bat. In: The Australian Museum Complete Book of Australian Mammals 3rd edition (ed R Strahan) Cornstalk Publishing, North Ryde, 328. Taylor R J & Savva N M 1988 Use of roost sites by four species of bats in State Forest in south-eastern Tasmania. Australian Wildlife Research 15:637-645. Vestjen W & Hall L 1977 Stomach contents of 42 species of bats from the Australasian region. Australian Wildlife Research 4:25-35. journal of the Royal Society of Western Australia, 79:217-224, 1996 Waterbirds and aquatic invertebrates of swamps on the Victoria-Bonaparte mudflat, northern Western Australia S A Halse1, R J Shiel2 & G B Pearson1 1 Department of Conservation and Land Management, Wildlife Research Centre, PO Box 51 Wanneroo, WA 6065 2 Murray-Darling Freshwater Research Centre, PO Box 921 Albury, NSW 2640 Manuscript received May 1996; accepted October 1996 Abstract The aquatic invertebrate and waterbird faunas of the western part of the Victoria-Bonaparte mudflat, located in the Victoria-Bonaparte biogeographic region of the Kimberley, Western Austra¬ lia, were surveyed in early 1993. Sixty-two species of waterbird and at least 131 species of aquatic invertebrate were recorded. The mudflat has national significance for some shorebird species, especially Redshanks, and has a significant Asian element in the invertebrate fauna. The survey contributes to growing evidence that the micro-invertebrate fauna of Western Australia is distinctive, but the conservation significance of the area for invertebrates cannot be assessed properly without more information concerning other areas of northern Australia. Introduction Australia has been divided into 80 biogeographic re¬ gions, based on climate, geology, landform, flora, fauna and land use, to provide a framework for assessing the adequacy of Australia's system of conservation reserves (Thackway & Cresswell 1995). The Victoria-Bonaparte mudflat, an extensive area that is occasionally inundated by fresh or saline water, is within the Victoria-Bonaparte biogeographic region in the extreme north-eastern Kimberley area of Western Australia. Approximately 1,000,000 palaearctic shorebirds occur annually in coastal habitats of northern Australia, with major concentrations at Eighty-Mile Beach and Roebuck Bay, in Western Australia, and the Gulf of Carpentaria, in Queensland (Lane 1987). These areas are staging points for shorebirds as they migrate to and from Aus¬ tralia. Many shorebirds remain in northern Australia over the austral winter, rather than returning to the Northern Hemisphere to breed, but Minton & Martindale (1982) recorded only 7120 shorebirds along the coast between Darwin and Kununurra, including the Victoria- Bonaparte mudflat, during an aerial survey in late August 1981. As a result, subsequent shorebird studies in north-western Australia have been concentrated further south in the Kimberley and in the Pilbara, where more shorebirds had been recorded (e.g. Minton & Jessop 1994). Most information about aquatic invertebrates in lentic waters of northern Australia comes from studies in the Alligator Rivers region of the Northern Territory (e.g. Tait et al. 1984; Julli 1986; Hawking 1992) and extensive col¬ lecting by B V Timms (Timms 1988; Timms & Morton 1988). Information for the Kimberley is virtually re¬ stricted to the work of Timms, and what can be extracted from taxonomic and biogeographic papers that are mostly based on the collections of Williams (1979) and Timms (e.g. McKenzie 1966; Watts 1987). Recent work © Royal Society of Western Australia 1996 has shown that the stream fauna of the Kimberley is comparatively rich (M J Smith, W R Kay & S A Halse, unpublished observations) in contrast to the tentative suggestion of Williams (1979) that it was depauperate. The only comprehensively studied lentic site in the Kimberley, Lake Gregory on the edge of the Great Sandy Desert, also supported a rich fauna (Halse et al. in press). McKenzie (1966) commented on the apparent affinities of the ostracod faunas of the Kimberley and Indonesian islands. Lansbury (1984) suggested that the hemipteran fauna of coastal Kimberley waterbodies has similar Indonesian links, unlike the remainder of Australia. The biology of the Victoria-Bonaparte biogeographic region, like much of the Kimberley area, is poorly known, and less than 10 per cent of the Western Austra¬ lian portion is reserved. In a review of the conservation reserve system of the Kimberley, Burbidge et al. (1991) recommended that the conservation value of the Victoria- Bonaparte mudflat be investigated, and McKenzie et al. (1991) examined two rainforest patches (EK05, EK06). The aim of the present study was to gather information on waterbirds and aquatic invertebrates using the mudflat. The conservation value of the mudflat and the biogeographic affinities of some invertebrate taxa are discussed. Study area The Victoria-Bonaparte mudflat is a band approxi¬ mately 10-20 km wide along the coast of Joseph Bonaparte Gulf, between Cambridge Gulf in Western Australia and the mouth of the Victoria River in the Northern Territory (Fig 1). Tidal flats up to 2 km wide and a narrow sandy beach occur on the northern side of the mudflat, south of which is an elevated narrow band of low dunes between the beach and a hinterland of extensive bare mudflat. Several creeks, lined by man¬ groves, drain the mudflat. Along the landward boundary there are many small freshwater pools and some larger areas of freshwater marsh. 217 Journal of the Royal Society of Western Australia, 79(3), September 1996 Figure 1. Victoria-Bonaparte mudflat, showing the sampling sites for aquatic invertebrates and other features mentioned in text (the boundary of the mudflat in the Northern Territory is drawn approximately). 1, Grassed Pool; 2, Brolga Spring; 3, Samphire Pool; 4, Mudflat Pool; 5, Rainforest Swamp; 6, Edge Swamp Several shallow, seasonally inundated freshwater pools covered with emergent grass Vetiveria pauciflora and sedges Cyperus conicus and C. javanicus occur at the north-western end of the mudflat (Fig 1). The northern¬ most of the pools is referred to as Grassed Pool. At the south-western end is a small permanent swamp, Brolga Spring, which contains Melaleuca viridiflora trees and Sesbama cannabina on the southern side. There are two shallow pools east of the mouth of the second largest creek on the mudflat, near the coastal rainforest site, EK05 (McKenzie et al. 1991). The freshwater, seasonal Samphire Pool is located in an interdunal swale and sup¬ ports emergent samphire Halosarcia halocnemoides tenuis and sedge Cyperus acjuatilis, while the brackish, tidally filled Mudflat Pool lies in a depression on the southern side of the dune area with a shallow drainage line connecting it to bare mudflat; it has some emergent grass. Rainforest Swamp, a small semi-permanent freshwater swamp, containing tall Melaleuca argentea and Nauclea orientalis trees and small herbaceous ferns such as Cyclosorus interruptus and Blechnum orientale, was adja¬ cent to the rainforest site, EK06, near Long Spring (McKenzie et al. 1991). There is also a small seasonal freshwater swamp at Long Spring between the fringing woodland and open mudflat, referred to herein as Edge Swamp. Around Snake Spring, on the southern edge of the central part of the mudflat, extensive areas are ephemerally flooded with fresh water via small creeks from the south and covered with dense stands of grass and sedge. Methods Waterbirds were counted from 17-19 February by SAH and from 4-6 April by SAH and GBP in 1993 to document the numbers and species of waterbirds using the area in late summer, during the middle of the wet season, and in autumn when migratory shorebirds are concentrated in northern Australia prior to leaving for the Northern Hemisphere. A partial count of the Western Australian portion of the Victoria-Bonaparte mudflat was made from a Cessna 182 flying at a height of 30-40 m and speed of 100 knots on 17 February. On 18-19 February, several marine and freshwater sections of the mudflat were counted from a Robinson helicopter flying at a height of 15 m and speed of 50 knots. In addition, ground counts were made at Snake Spring, Grassed Pool and on 4 km of tidal flat north of EK05. On 4 April, ground counts were made among mangroves and on tidal flat west of the creek mouth at EK05, and at the second-most north-western grassed pool (Fig 1). On 5-6 April, a complete count of the Western Australian portion of the mudflat was made from a Cessna 182 flying at a height of 20 m and speed of 60 knots. Two samples of aquatic invertebrates were collected from Grassed Pool, Samphire Pool, Mudflat Pool, Rainforest Swamp and Edge Swamp by SAH on 18-19 February by sweeping approximately 50 m of water, and as many microhabitats as possible, using FBA-type pondnets with 50 and 110 pm mesh. The 50 pm mesh sample was preserved in 5 per cent buffered formalin, the 110 pm mesh sample was preserved in 70 per cent alcohol for sorting and identification in the laboratory. Water samples were taken from four of the sites for measurement of conductivity (mS cm1; TPS LC81), which was converted to parts per thousand total dissolved sol¬ ids (ppt TDS) as TDS = 0.6 mS cm1. Results Altogether, 62 species of waterbird were recorded on the Western Australian portion of the Victoria-Bonaparte mudflat, including 28 species of shorebird (Table 1). More than 4,000 birds were counted in both February and April, with almost 2000 shorebirds present each oc¬ casion. Notable records in February were 1701 Magpie Geese, mostly near Snake Spring, 126 Marsh Sandpipers near Snake Spring and Grassed Pool, and 5 Redshanks on the creek near EK05. Notable records in April were 218 Journal of the Royal Society of Western Australia, 79(3), September 1996 Table 1 Waterbirds recorded on the Victoria-Bonaparte mudflat in Feb¬ ruary and April 1993 Species Feb April Magpie Goose Anseranas semipalmata 1701 10 Plumed Whistling-Duck Dendrocygna eytoni 20 6 Wandering Whistling-Duck Dendrocygna arcuata 12 Radjah Shelduck Tadorna radjah 11 5 Green Pigmy-Goose Nettapus pulchellus 11 1 Pacific Black Duck Anas superciliosa 39 37 Grey Teal Anas gracilis 32 Hardhead Ay thy a australis 1 Darter Anhinga melanogaster 3 Little Pied Cormorant Plmlacrocorax melanoleucos 30 Little Black Cormorant Plmlacrocorax sulcirostris 11 Australian Pelican Pelecanus conspicillatus 23 17 Little Egret Egretta garzetta 42 54 Eastern Reef Egret Egretta sacra 6 1 Pied Heron Ardea pictata 4 Great Egret Egretta alba 5 6 Intermediate Egret Egretta intermedia 2 2 Unidentified egret 48 99 Striated Heron Butorides striatus 2 Nankeen Night Heron Nycticorax caledonicus 18 1 Glossy Ibis Plegadis falcinellus 59 Australian White Ibis Threskionis aethiopica 183 281 Straw-necked Ibis Threskionis spinicollis 30 Black-necked Stork Xenorhynchus asiaticus 14 6 Sarus Crane Grus antigone 2 Brolga Grus rubicundis 18 81 Black-tailed God wit Limosa lirnosa 2 2 Bar-tailed Godwit Limosa lapponica 30 171 Whimbrel Numenicus phaeopus 21 53 Eastern Curlew Numenius madagascariensis 6 5 Little Curlew Numenius minutus 1 Common Redshank Tringa totanus 5 Marsh Sandpiper Tringa stagnatilis 126 83 Common Greenshank Tringa nebularia 13 65 Wood Sandpiper Tringa glareola 3 Terek Sandpiper Tringa terek 29 Common Sandpiper Tringa hypoleucos 1 Grey-tailed Tatler Tringa brevipes 15 Ruddy Turnstone Arenaria interpres 5 Great Knot Calidris tenuirostris 10 74 Red Knot Calidris canutus 5 Red-necked Stint Calidris ruficollis 34 Sharp-tailed Sandpiper Calidris acuminata 1 Curlew Sandpiper Calidris ferruginea 2 2 Comb-crested Jacana Jrediparra gallinacea 2 Beach Stone-Curlew Burhinus neglectus 2 1 Pied Oystercatcher Haemotopus longirostris 36 21 Black-winged Stilt Himantopus himantopus 101 261 Red-necked Avocet Recurvirostra novaehollandiae 1 Pacific Golden Plover Pluvialis dominica 1 1 Grey Plover Pluvialis sqatarola 1 14 Red-capped Plover Charadrius ruficapillus 11 2 Greater Sand Plover Charadrius leschenaultii 3 43 Red-kneed Dotterel Erylhrogonys cinctus 1 Masked Lapwing Vanellus miles 37 120 Unidentified wader 1559 903 Silver Gull Larus novaehollandiae 5 3 Gull-billed Tern Gelochelidon nilotica 7 133 Caspian Tern Hydroprogne caspia 3 6 Crested Tern Sterna bergii 1 Little Tern Sterna albifrons 1 32 Whiskered Tern Chlidonias hydrida 37 White-winged Black Tem Chlidonias leucoptera 1123 Unidentified tern 162 354 Clamorous Reed-Warbler Acrocephalus stentoreus 2 Total 4384 4304 1123 White-winged Black Terns and 53 Whimbrels. Counts for some species, especially shorebirds and terns, were under-estimates because not all birds could not be identified to species level during the aerial surveys and categories such as 'unidentified tern' were used. Esti¬ mates of total waterbird numbers were probably reason¬ ably accurate, however (see Halse et al. 1996). All sites sampled for invertebrates on the Victoria- Bonaparte mudflat, except Mudflat Pool (11.8 ppt TDS), were fresh. The salinity of Samphire Pool was 0.84 ppt, Brolga Spring was 0.51 ppt and Edge Swamp was 0.20 ppt. At least 131 species of invertebrate were collected, with the number of species per site ranging from 13 at the brackish Mudflat Pool to 60 at Edge Swamp (Appen¬ dix). Species-rich orders were Cladocera (27 species), Ploimida (24), Coleoptera (20), Copepoda (10) and Ostracoda (10). A significant number of species were undescribed. Apart from Metacy clops sp. EK1, these included the ostra- cods Bennelongia sp. 363, Cyprinotus sp. 362 and Ilyodromus sp. 365, cladocerans Macrothrix sp. A and Macrothrix sp. B and rotifers Lecane sp. A and Lecane sp. B (Appendix). The ostracod, Strandesia/Chamydotheca sp., appeared to be an undescribed genus (P De Deckker, pers. comm.). Many of the insect larvae, especially dipter- ans, could not be identified to species because the taxonomy of the larvae is poorly known. Discussion Fewer waterbird species (62) occurred on the Western Australian portion of the Victoria Bonaparte mudflat dur¬ ing the two surveys than have been recorded on the long¬ term list for Parry Lagoons (77), which is the best known site for waterbirds in the Victoria-Bonaparte biogeo¬ graphic region (Jaensch & Lane 1993). Total numbers were substantially lower than the highest counts at Parry Lagoons and 4304 birds of 50 species were recorded on the mudflat in April, compared with 8145 birds of 33 species at Parry Lagoons on the same day (aerial survey, S A Halse & G B Pearson, unpublished observations). Nevertheless, the mudflat is more important for waterbirds, especially shorebirds, than previously thought. Because of the greater range of habitat for shore- birds, especially tidal flats, the two surveys of the mudflat yielded 28 species of shorebird compared with the long-term list of 24 for Parry Lagoons. In April, 1910 shorebirds were counted on the West¬ ern Australian portion of the Victoria-Bonaparte mudflat, with a further 6642 counted between the Northern Territory border and the Fitzmaurice River, an overall distance of approximately 250 km (Fig 1). In February, 1973 shorebirds were counted on the mudflat. These counts represent higher densities than observed by Minton & Martindale (1982), who recorded 7120 shore- birds on the 630 km between Darwin and Kununurra. The mudflat also has national significance for several shorebird species (see Watkins 1993). The 5 Redshanks seen in February comprised as large a flock as has been recorded in Australia, equalled only by 5 birds at Broome in April 1985 (Lane & Jessop 1985; Hewish 1987). It is possible that other Redshanks were present on the creek near EK05, or on other parts of the mudflat. Australian 219 Journal of the Royal Society of Western Australia, 79(3), September 1996 records are mostly restricted to north-western Australia and the Victoria-Bonaparte mudflat must be regarded as one of the more important Australian sites for the species. Although the count of 53 Whimbrels at the mouth of the creek near EK05 in April under-estimated the number on the whole mudflat, when combined with 235 Whimbrels seen on the same day in the Northern Territory between the Fitzmaurice River and Western Australian border (S A Halse & G B Pearson, unpub¬ lished observations), the total count for the 250 km length of coast was the seventh highest for the species in Aus¬ tralia (Watkins 1993). The count of 126 Marsh Sandpip¬ ers in February, although an under-estimate of the num¬ ber on the whole mudflat, was the fifteenth highest count in Australia (Watkins 1993). The conservation value of the Victoria-Bonaparte mudflat for aquatic invertebrates is difficult to evaluate because there have been few studies in similar areas. The floodplains of the Alligator Rivers region appear to support more species; Shiel & Koste (1983) recorded an average of 34 rotifer species per sample and Julli (1986) recorded approximately 20-30 species of cladocerans compared with 2-15 rotifers and 2-13 cladocerans at sites on the mudflat. On the other hand, with 28-60 species at freshwater sites the mudflat compared well with Lake Gregory, in the southern Kimberley which, when fresh, averaged 35 species per site (Halse et al. in press; S A Halse, unpublished observations). The distributions of most species of aquatic inverte¬ brates in northern Australia are poorly known and, therefore, the significance of unusual records on the mudflat is difficult to assess. For example, the only pre¬ vious record of the harpacticoid copepod Cletocamptus confluens in Australia was from Peron Peninsula, Shark Bay, 2000 km south-west of the mudflat (Lang 1948), so its occurrence at Mudflat Pool has conservation implica¬ tions, except there have been so few invertebrate surveys that the species may well be widespread. Likewise, at present many aquatic habitats in the Kimberley support significant numbers of undescribed species and this, by itself, cannot be used as evidence of special conservation value. A recent study of land snails in rainforest patches of the Kimberley found that at least 36 per cent, and perhaps as many as 50 per cent, of species collected were undescribed (Solem 1991). All 26 pseudoscorpion species collected in the same study were undescribed (Harvey 1991), as were the 11 indigenous earthworm species (McKenzie & Dyne 1991). It is worth noting that, in addition to obviously undescribed taxa, several taxa collected on the mudflat differed only slightly from descriptions of known spe¬ cies. Without collections from other localities and de¬ tailed analyses, it is impossible to determine whether the small differences represented geographic variation or new species. We have listed these taxa as having affinity to known species, including the rotifer Scandium elegans and the cladoceran Moina weismanni, both of which have not been recorded in Australia previously. The rotifers of northern Australia have strong Indo- Malaysian affinities (Shiel & Williams 1990) and this seems to be the case for much of the crustacean fauna on the Victoria-Bonaparte mudflat. The cladoceran fauna of northern Australia has a strong circumtropical element (Timms 1988) and several genera with Indo-Malaysian affinities, such as Pseudosida, Grimaldina and Moinadaphnia, occurred on the mudflat. The undescribed cyclopoid copepod, Metacyclops sp EK1, is unlike other Australian Metacyclops and has affinities with species from north of Australia (D W Morton, personal commu¬ nication). Similarly the ostracod, Cypris subglogosa, which was recorded in Australia for the first time, is wide¬ spread in ricefields of equatorial regions, Asia and east¬ ern Europe (Okubo 1972a). Cyprinotus kimberleyensis, which appears to be widespread in the Kimberley, is also found in Asian ricefields (Okubo 1974) and the ostracod genus, Strandesia , which was well represented in the mudflat samples, is common throughout Asia ( e.g . Okubo 1972b; Victor et al. 1980). In conclusion, the Victoria-Bonaparte mudflat is a na¬ tionally significant area for waterbirds, especially shore- birds, although the numbers of birds found there is not as high as in some other areas of the Kimberley. The use of the area by Redshanks, a species that mostly stops in southern Asia during the austral summer, is consistent with the occurrence of invertebrates on the mudflat that have Asian affinities. A more comprehensive inventory of invertebrate species and their distributions is required before the relative significance of the Victoria-Bonaparte mudflat for the conservation of invertebrates can be as¬ sessed. However, the survey supported Lansbury's (1984) suggestion that there is a strong Asian element in the invertebrate fauna of the coastal Kimberley and added to the growing evidence that Western Australia has a distinctive micro-invertebrate fauna (Frey 1991; Maly & Bayly 1991; Storey et al. 1993). Acknowledgments: We thank G J Keighery for floristic information and A Clarke for sorting the invertebrate samples. P De Deckker (ostracods), R Hamond (harpacticoids), M S Harvey (water mites), J Hawking (odo- nates), W R Kay and I Lansbury (hemipterans), D W Morton (cyclopoids), S Slack-Smith (gastropods), M J Smith (chironomids) and CHS Watts (beetles) assisted with identifications. The February sampling trip was funded by the Department of Environment, Sport and Territories, and the March trip was partly funded by the Australian Nature Conservation Agency under the States Cooperative Assistance Program. References Burbidge A A, McKenzie N L & Kenneally K F 1991 Nature Conservation Reserves in the Kimberley Western Australia. Department of Conservation and Land Management, Perth. Frey D G 1991 The species of Pleuroxus and of three related genera (Anomopoda, Chydoridae) in southern Australia and New Zealand. Records of the Australian Museum 43:291- 372. Halse S A, Pearson G B, Jaensch R J, Kulmoi P, Gregory P, Kay W R & Storey A W 1996 Waterbird surveys of the middle Fly River floodplain, Papua New Guinea. Wildlife Research 23:557-569. Halse S A, Shiel R J & Williams W D in press Aquatic inverte¬ brates of Lake Gregory, north-western Australia, in relation to salinity and ionic composition. Hydrobiologia. Harvey M S 1991 The Pseudoscorpionida and Schizomida of the Kimberley rainforests. In: Kimberley Rainforests of Australia (eds N L McKenzie, R B Johnson & P G Kendrick). Surrey Beatty, Sydney, 265-268. Hawking J H 1992 A preliminary guide to the identification of dragonfly larvae (Odonata) from the Alligator Rivers region of the Northern Territory. Office of the Supervising Scientist, Jabiru. 220 Journal of the Royal Society of Western Australia, 79(3), September 1996 Hewish M 1987 The summer 1987 population monitoring count: rarities and the wader counts. Stilt 11:23-31. Jaensch R P & Lane J A K 1993 Western Australia. In: A Direc¬ tory of Important Wetlands in Australia (eds S Usback & R James). Australian Nature Conservation Agency, Canberra, 10-1-10-178. Julli M E 1986 The taxonomy and seasonal population dynamics of some Magela Creek floodplain micro-crustaceans (Cladocera and Copepoda). Technical Memorandum 18. Of¬ fice of the Supervising Scientist, Canberra. Lane B A 1987 Shorebirds in Australia. Nelson, Melbourne. Lane B & Jessop A 1985 Report on the 1985 north-west Austra¬ lia wader studies expedition. Stilt 6:2-14. Lang K 1948 Monographic der Harpacticoiden, 2 vols. Hakan Ohlsson, Lund. Lansbury I 1984 Some Nepomorpha (Corixidae, Notonectidae and Nepidae) (Hemiptera-Heteroptera) of north-west Aus¬ tralia. Transactions of the Royal Society of South Australia 108:35-49, Maly E J & Bayly I A E 1991 Factors influencing biogeographic patterns of Australasian centropagid copepods. Journal of Biogeography 18:455-461. McKenzie K G 1966 Freshwater Ostracoda from north-western Australia. Australian Journal of Marine and Freshwater Re¬ search 17:259 279. McKenzie N L & Dyne G R 1991 Eartworms of rainforest soils in the Kimberley, Western Australia. In: Kimberley Rainforests of Australia (eds N L McKenzie, R B Johnson & P G Kendrick). Surrey Beatty, Sydney, 133-144. McKenzie N L, Johnston R B & Kendrick P G 1991 Kimberley Rainforests of Australia. Surrey Beatty, Sydney. Minton C & Jessop R 1994 The 1994 north-west wader expedi¬ tion. Stilt 25:8-11. Minton C & Martindale J 1982 Report on wader expedition to north-west Australia in August/September 1981. Stilt 2:14- 26. Okubo I 1972a Freshwater Ostracoda from Japan, II. Cypris subglobosa Sowerby, 1840. Research Bulletin 1. Shujitsu Junior College, Okayama, 61-72. Okubo I 1972b Strandesia camaguinensis Tressler, 1937, from Japan (Ostracoda, Cyprididae). Proceedings of the Japanese Society of Systematic Zoology 8:9-14. Okubo I 1974 Two species of the genus Cyprinotus (Ostracoda, Cyprididae) from Japan. Researches on Crustacea 6:58-70. Shiel R J & Koste W 1983 Rotifer communities of billabongs in northern and south-eastern Australia. Hydrobiologia 104:41- 47. Shiel R J & Williams W D 1990 Species richness in tropical fresh waters of Australia. Hydrobiologia 202:175-183. Solem A 1991 Land snails of Kimberley rainforest patches and biogeography of all Kimberley land snails. In: Kimberley Rainforests of Australia (eds N L McKenzie, R B Johnson & P G Kendrick). Surrey Beatty, Sydney, 145-246. Storey A W, Halse S A & Shiel R J 1993 Aquatic invertebrate fauna of the Two Peoples Bay area, south-western Australia. Journal of the Royal Society of Western Australia 76:25-32. Tait R D, Shiel R J & Koste W 1984 Structure and dynamics of zooplankton communities, Alligator Rivers region, N T, Aus¬ tralia. Hydrobiologia 113:1-13. Thackway R & Cresswell I D 1995 An Interim Biogeographic Regionalisation for Australia. Australian Nature Conserva¬ tion Agency, Canberra. Timms B V 1988 The biogeography of Cladocera (Crustacea) in tropical Australia. Internationale Revue der Gesampten Hydrobiologie 73:337-356. Timms B V & Morton D W 1988 Crustacean zooplankton as¬ semblages in freshwaters of tropical Australia. Hydrobiologia 164:161-169. Victor R, Paul M A & Fernando C H 1980 Two new species of the genus Strandesia Vavra, 1985 (Ostracoda - Crustacea) from Kerala, southern India. Canadian Journal of Zoology 58:727-734. Watkins D 1993 A national action plan for shorebird conserva¬ tion in Australia. Report 90. Royal Australasian Ornitholo¬ gists Union, Melbourne. Watts CHS 1987 Revision of Australian Berosus Leach (Co- leoptera: Hydrophilidae). Records of the South Australian Museum 21:1-28. Williams W D 1979 Notes on the freshwater fauna of north¬ western Australia, especially the Kimberleys. Records of the Western Australian Museum 7:213-227. 221 Journal of the Royal Society of Western Australia, 79(3), September 1996 Appendix Taxa of aquatic invertebrates collected at six sites on the Victoria-Bonaparte mudflat in February 1993. 1, present; la, adult beetles; lb, larval beetles only Taxon PROTISTA LOBOSEA ARCELLINIDA Arcellidae Centropyxis aculeata (Ehrenberg) ROTIFERA DIGONONTA BDELLOIDEA Bdelloidea sp. MONOGONONTA FLOSCULARIACEA Hexarthridae Hexarthra intermedia Wiszniewski Filinidae Filinia longiseta (Ehrenburg) Testudinellidae Testudinella patina (Hermann) PLOIMIDA Asplanchnidae Asplanchna brightwelli (Gosse) Brachionidae Anuraeopsis fissa (Gosse) Brachionus angularis Gosse Brachionus quadridentatus Hermann Brachionus sp. C Platyias quadricomis Ehrenberg Notomamatidae Cephalodella sp. Scandium aff. elegans Segers & De Meester Synchaetidae Polyarthra dolichoptera Nelson Lecanidae Lecane ( Monostyla ) bulla Gosse Lecane {Monostyla) hamata Stokes Lecane (s. str.) curvicomis (Murray) Lecane (s. str.) luna (Muller) Lecane (s. str.) leontina (Turner) Lecane (s. str.) ludwigi (Eckstein) Lecane (s. str.) ungulata (Gosse) Lecane (s. str.) sp. nov. A Lecane (s. str.) sp. nov. B Lecane (s. str.) sp. Euchlanidae Dipleuchlanis propatula (Hudson & Gosse) Dipleuchlanis sp. B Euchlanis dilatata Ehrenberg Euchlanis sp. B Mytilinidae Mytilina acanthophora Hauer Trichotriidae Trichotria sp. MOLLUSCA GASTROPODA BASOMMATOPHORA Planorbidae Physastra sp. Gyraulus sp. Lymnaeidae Lymnaea sp. ARTHROPODA ARACHNIDA HYDRACARINA Pionidae Fiona australica K O Viets Hydrachnidae Hydrachna sp. CRUSTACEA NOTOSTRACA Triopsidae Triops australiensis Spencer & Hall CLADOCERA Sididae Mudflat Samphire Edge Rainforest Grassed Brolga Pool Pool Swamp Swamp Pool Spring 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 222 Journal of the Royal Society of Western Australia, 79(3), September 1996 Mudflat Samphire Edge Rainforest Grassed Brolga Pool Pool Swamp Swamp Pool Spring Diaphanosoma aff. unguiculatum Gurney Latonopsis sp. Pseudosida szalayi Daday Sarsilatona papuana (Daday) Chydoridae Alona aff. rectangula Sars Alona rectangula novaezealandiae (Sars) Alona aff. s etuloides Smirnov & Timms Alona aff. costata Sars Alona sp. A Alona ? sp. nov. B Biapertura aff. karua (King) Biapertura aff. setigera (Brehm) Dunhevedia crassa King Kurzia longirostris (Daday) Lcydigia acantliocercoides (Fischer) Lcydigia aff. ciliata Gauthier Macrothricidae Grimaldina brazzai Richard Macrothrix aff. capensis (Sars) Macrothrix aff. timmsi Smirnov Macrothrix triserialis Brady Macrothrix sp. nov. A Macrothrix sp. nov. B Moinidae Momadaphnia macleayi (King) Moina aff. micrura Kurz Moina aff. weismanni Ishikawa Daphniidae Ceriodaphnta sp. Simocephalus sp. OSTRACODA Ilyocyprididae Ilyocypris australiensis Sars Bennelongia sp. 363 Cyprinotus sp. 362 Cypretta baylyi McKenzie Cyprinotus kimberleyensi McKenzie Cypris subglobosa Sowerby Ilyodromus sp. 365 Strandesia ? camaguinensis Tressler Strandesia sp. 360 Strandesia/Chlamydotheca gen nov. 357 CONCHOSTRACA Cyzicidae Cyzicus sp. A Cyzicus sp. B Cyclestheriidae Cyclestheria sp. COPEPODA Centropagidae Diaptomus sp. Cyclopoidae Apocyclops dengizicus (Lepeschkin) Mesocyclops ? australiensis (Sars) Mesocy clops sp. LS2 Metacy clops sp. EK1 Microcyclops varicans (Sars) Thermocyclops sp. LSI Canthocamptidae Cletocamptus confluens Kiefer Cletocamptus dietersi (Richard) Canthocampidae sp. 267 DECAPODA Hymenosomatidae Holthuisiana transversa (von Martens) INSECTA EPHEMEROPTERA Baetidae Cloeon sp. ZYGOPTERA Coenagrionidae Coenagrionidae sp. ANISOPTERA Anisoptera sp. Aeschnidae Hemianax papuensis (Burmeister) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 111111 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 223 Journal of the Royal Society of Western Australia, 79(3), September 1996 Taxon Mudflat Samphire Edge Rainforest Grassed Brolga Pool Pool Swamp Swamp Pool Spring Gomphidae Hemigomphus sp. Libellulidae Diplacodes bipunctata (Brauer) Pantala flavescens (Fabricius) HEMLPTERA Gerridae ? Limtiogonus sp. Corixidae Micronecta aff. virgata Hale Nepidae Austronepa angustata (Hale) Belostomatidae Lethocerus sp. Notonectidae Anisops malkini Brooks Anisops nasuta Fieber Anisops sp. DIPTERA Culicidae Culex annulirostris Skuse Chironomidae Chironomus sp. Cladotanytarsus sp. K4 Larsia albiceps (Johannsen) Procladius paludicola (Skuse) Tany tarsus aff. K12 Stratiomyidae Stratiomyidae sp. Ephydridae Ephydridae sp. Tabanidae Tabamdae sp. LEPIDOPTERA Pyralidae Pyralidae sp. 1 Pyralidae sp. 2 TR1CHOPTERA Leptoceridae Leptoceridae sp. COLEOPTERA Gyrinidae Macrogyrus sp. Noteridae Canthydrus bovillae Blackburn Dytiscidae Eretes australis (Erichson) Homeodytes sp. Hydroglyphus godeffroyi (Sharp) Hydroglyphus grammopterus (Zimmerman) Hyphydrus sp. Laccophilus clarki Sharp Laccophilus sharpi Regimbart Laccophilus spp Dytiscidae sp. A Dytiscidae sp. B Hydraenidae Paracymus sp. Spercheidae Spercheus sp. Hydrophilidae Berosus debilipennis Blackburn Berosus aff. ralphi Watts Berosus sadie/aquilo complex Enochrus deserticola Blackburn Stemolophus immarginatus d'Orchymont Hydrophilidae sp. Helodidae Helodidae sp. Scirtidae Scirtidae sp. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 11111 111 11 1 1 1 1 1 1 1 1 1 1 1 1 1 la la la lb lb lb lb la lb la la lb lb lb la la la la lb lb lb lb lb lb la la la la la la la la la la lb la lb lb lb Total 13 28 60 42 42 39 224 Journal of the Royal Society of Western Australia, 79(4), December 1996 Symposium on the Design of Reserves for Nature Conservation in South-western Australia Sponsored by The National Biodiversity Council and The Centre for Ecosystem Management, Edith Cowan University June 1995 Edited by P C Withers & P Horwitz for The Royal Society of Western Australia South-West, Western Australia Landsat TM Summer Mosaic. 1994-1996. Bands I. 4. 7 in Blue. Green. Red. Kilometers Produced by: Remote Sensing Services. Department of Land Administration. 1-ccuwin Centre. Perth. W.A. Landsat imagery provided by Australian Centre for Remote Sensing (ACRES), AUSLIG, Canberra, and digitally enhanced and produced by Remote Sensing Services, Department of Land Administration, Perth, Western Australia. l Journal of the Royal Society of Western Australia, 79:iii-iv, 1996 Symposium on the Design of Reserves for Nature Conservation in South-western Australia Preface This issue of the Journal of the Royal Society of Western Australia relates a series of papers, all presented at a symposium on the design of reserves for nature conser¬ vation in south-western Australia. The symposium was held in June 1995 at a time when there was (and still is) increasing public concern about nature conservation and conservation through reserves in the more densely populated, wetter parts of the south-west. The area of concern extends from the Swan Coastal Plain to adjacent forested ecosystems, where intense pressure exists for urban development, agriculture/horticulture, mining and timber extraction. There is a general perception that the scientific community needs to begin an informed debate, involving the general public, and this sympo¬ sium was seen as an opportunity to air the issues. The symposium coincided with a critical point in the forest debate in Australia. The Commonwealth Govern¬ ment, in an attempt to avoid perennial protracted dis¬ putes over woodchip quotas and forest reservation, en¬ gaged several prominent Australian forest ecologists to prepare selection criteria for reserves as a broad bench¬ mark to preserve biodiversity, old growth and wilder¬ ness values of forests according to the aims of the Na¬ tional Forest Policy Statement (1992). A report was pub¬ lished (Commonwealth of Australia, 1995) which rec¬ ommended, amongst many other criteria, that each for¬ est community in Australia should have at least 15% of its pre-1750 extent reserved. Could this figure be suffi¬ cient to ensure the long term security of biodiversity in Australia's forests, or any other ecosystem? The symposium was organised because of concern that any Australia-wide approach might not take into ac¬ count the particular vagaries of south-western Austra¬ lian biogeography. It was sponsored by the National Biodiversity Council and the Centre for Ecosystem Man¬ agement at Edith Cowan University. Speakers were in¬ vited to cover a broad range of topics relevant to reser¬ vation of biodiversity, to ensure that a more comprehen¬ sive scientific knowledge on biodiversity in various landscapes of the south-western part of the continent could be accessed as part of the debate. The Symposium was opened by the Hon Peter Foss, then Minister for the Environment in the Western Aus¬ tralian Government. In his speech the Minister sug¬ gested that scientists concerned with the conservation of biodiversity should direct their priorities to the wheatbelt of Western Australia rather than to the higher rainfall areas. The question arises: are scientific concerns for the management of urban or forested landscapes in Western Australia unfounded, misplaced or of a lower priority? One response of symposium delegates was that the wetter parts of the south-west must never be allowed to become degraded like the wheatbelt region, and that environmental degradation could be averted if early warnings from scientists are promptly considered. Scientists' concerns about the wheatbelt date back to the early part of this century. Warnings about the conse¬ quences of clearing native vegetation for both water quality and nature conservation have been sounded by scientists and naturalists for at least 70 years, but it is only now that such warnings are beginning to be heeded, and only in part. Many scientists take the view that the current management of south-western Austra¬ lian forests and coastal plains requires immediate atten¬ tion, much like the wheatbelt did 70 years ago. Indeed, forest and coastal ecosystems might also provide refugia for wheatbelt biota, as well as the unique biota which persist in these relatively moist environments. The articles in this issue present perspectives on reserva¬ tion of the wetter landscapes of south-western Australia. They include an historical review (Rundle), an overall assessment of the amount of reservation compared to other parts of the state (McKenzie et al), and assess¬ ments of reservation status of biota from the viewpoint of taxonomic groups (Wardell-Johnson; BY Main), habi¬ tats (Trayler et al.) or ecosystem dynamics (Hobbs). An overview of forest reservation is the final paper pre¬ sented (AR Main). The area of focus was that bounded by the 500 mm isohyet, approximating Beard's (1982) Darling Botanical District. Another scheme includes three bioregions equating to Beard's (four) botanical subdistricts, namely the Swan, Jarrah Forest, and Warren Bioregions of the Interim Biogeographic Regionalisation of Australia (Thackway & Cresswell 1995). Alternatively, Hopper (1992) considered that the 800 mm isohyet (High Rain¬ fall Zone) was a meaningful biogeographical boundary, and matching this closely is the Environmental Protec¬ tion Authority's Reserve Systems of Western Australia. All four schemes are mentioned in papers in this issue (see Figure over), and a Landsat image of the south-west is presented as a frontispiece. In 1973 the Royal Society of Western Australia published a special edition which reviewed the current knowledge of the south-western Australian environment. In his Preface to that volume, the Honorary Editor (AJ McComb) referred to the role of this Society and its Journal to provide a forum whereby information could be collated and made accessible to residents of the State. We are pleased to contribute to this tradition. References Beard J S 1982 Vegetation survey of Western Australia. Swan. 1:1 000 000 vegetation series. University of Western Aus¬ tralia, Perth. Commonwealth of Australia 1995 National forest conservation reserves. Commonwealth proposed criteria. Commonwealth of Australia, Canberra. Hopper S D 1992 Patterns of plant diversity at the population and species levels in south-west Australian Mediterranean ecosystems. In: Biodiversity of Mediterranean Ecosystems in Australia (ed R J Hobbs). Surrey Beatty & Sons, Chipping Norton, 27-46. National Forest Policy Statement 1992 Commonwealth of Australia, Canberra. Thackway R & Cresswell I D 1995 An Interim Biogeographic Regionalisation for Australia: A Framework for Setting Priorities in the National Reserves System Cooperative Program. Australian Nature Conservation Agency, Canberra. Dr Pierre Horwitz Symposium Convener, and Honorary Co-editor iii Journal of the Royal Society of Western Australia, 79(4), December 1996 Four biogeographical schemes referred to in papers in this issue. In clockwise order starting from the top left; Beard's (1982) Darling Botanical District (with four subdistricts); the Environmental Protection Authority's reserve systems 1, 2 and 6 which cover the wetter south-west; Hopper's (1992) High Rainfall and Transitional Rainfall Zones; and the three bioregions from the Interim Biogeographic Regionalisation of Australia (Thackway & Cresswell 1995). IV Journal of the Royal Society of Western Australia, 79:225-240, 1996 History of conservation reserves in the south-west of Western Australia G E Rundle WA National Parks and Reserves Association, The Peninsula Community Centre, 219 Railway Parade, Maylands WA 6051 Abstract Focusing on the Darling Botanical District, reservation in the south-west of Western Australia largely involves the forest estate. The remaining natural bushland today is mainly reserves of State forest and so further opportunities to create new national parks or nature reserves of any significance would generally mean converting a State forest reserve to some other sort of conser¬ vation reserve. Thus, the history of Western Australia's State forest reservation is important. The varied origins of some of the region's well-known and popular national parks are of special interest. Their preservation as conservation reserves generally had little to do with scien¬ tific interest and a lot to do with community pleasure in the outdoors and scenery. Their protec¬ tion from early development had little to do with the flora and habitat protection needs that are the focus of these Symposium proceedings. Factors such as lack of shipping access, the discovery of glittering caverns, and the innovation of excursion railways were involved in saving the day. In contrast, the progressive reservation of State Forest was a hard slog by an insular Forests Depart¬ ment against many opponents. The creation of a comprehensive system of conservation reserves in this part of Western Australia is an on-going modem phenomenon with continued wide popular support. While there are controversies concerning forest reservation and the conservation of urban bushland, these largely involve matters of detail over particular pieces of land. However, remaining unresolved is controversy over State forest management which might impact upon strategic wildlife habitat. Management of State forest for multiple uses, including ecologically sustainable timber produc¬ tion and wildlife values, is adopted Government policy. Introduction Geographical context The Darling Botanical District is generally synony¬ mous with several significant natural attributes, includ¬ ing the high rainfall area (over 500 mm per annum) of Western Australia's south-west, and a zone of active streams and rivers, whereas the hinterland beyond is largely characterised by more arid conditions and un¬ coordinated drainage (Beard 1981). Tall forest and tall woodland formations are confined to this part of the State and today comprise the domi¬ nant area of uncleared land. Of particular significance, it forms a large block of generally contiguous reserves of State forest. Created originally to support an export saw milling industry, today's forest management emphasis is on multiple use objectives of which catchment protec¬ tion for water supply is arguably the most important. West of the forest of the Darling Plateau, the fringing narrow Swan Coastal Plain is the earliest part of West¬ ern Australia that was extensively settled by Europeans. The bulk of this land was released into private owner¬ ship within the first hundred years of this settlement, with little consideration given to a need for bushland conservation. Converting uncleared remnants into con¬ servation reserves there now usually requires the buying back of land that is in private ownership. There are situ- Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 ations where there is added justification to do this when key ground water aquifers within the deep sands of the coastal plain need to be protected from the likelihood of pollution arising out of urban and agricultural develop¬ ment. Administratively, the Darling Botanical District is also closely synonymous with several regions. These include the Environmental Protection Authority's (EPA) Systems 1, 2 and 6 collectively, which were designated in the early 1970s as part of a State conservation reserve assess¬ ment project (Ride 1975). State forest and the bulk of Western Australia's conservation reserves are today managed by the Department of Conservation and Land Management (CALM). The contiguous CALM regions of Southern Forest, Central Forest and Swan (formerly Southern Forest and Metropolitan Regions) are also ba¬ sically confined to the Darling Botanical District. Historical context Over the past 120 years of history of Western Australia's conservation and forest reserve systems, there have been a number of Acts of Parliament passed and consolidated that have enabled; • land (and waters) to be reserved, • the provision of degrees of security for reserves against alienation for other purposes, and • the management of reserves by specific agencies. Over that period, the principal reserve management agencies have also had name changes, or have been re¬ placed by another agency. The phases of forest and con¬ servation reserve establishment can be summarised as follows: 225 Journal of the Royal Society of Western Australia, 79(4), December 1996 • Colonial Era : 1829-1889 Little was appreciated about the relationship of wildlife and habitat de¬ pendence, and the need was not seen for conser¬ vation reserves in a region which was still rela¬ tively undeveloped and unpopulated. Little inter¬ est was shown in protecting forest resources. • Early Self-Government: 1890-1929 In this brief 40 year period, gold rushes to the Murchison and Eastern Goldfields led to a rapid population in¬ crease via immigration and the promotion of rail¬ way development. In turn, both events stimulated the creation of recreation-based reserves (e.g. today's John Forrest and Leeuwin-Naturaliste Na¬ tional Parks). A shift away from labour-intensive shallow fossicking to mechanised deep mining on the goldfields diverted development interest to agri¬ cultural expansion, especially in the forested 'south-west' (Darling Botanical District) and ad¬ joining Wheatbelt, aided by a developing railway network. The emphasis on conservation reserve selection re¬ mained oriented toward scenic landscape and rec¬ reation with new national parks being created at Pemberton and Nornalup. The need for strategic reserves for habitat protection was ignored, par¬ ticularly in the Wheatbelt. At the same time, there was rapid expansion of the timber industry, but also a growing awareness that better protection and management was needed for the forest estate. This need was en¬ hanced by the utilisation of some forested catchments for water supply dams. A significant step was the passing of a Forest Act in 1918, the formation of a Forests Department, and the estab¬ lishment of a million ha of State forest reserves by 1929. • Modern Era: 1930-1969 While there was little change over the period of the Great Depression and World War II, post war reconstruction brought major change. Rapid agricultural, mining, industrial and urban expansion brought an aware¬ ness of the need to secure more natural areas for habitat protection and as 'refuges' for passive rec¬ reation. There was also an emerging realisation of the need for professional and well-resourced man¬ agement of reserves, and lobbying for improve¬ ment was a feature of this period. Characteristically, it was an era of land-use con¬ flict between competing developments, and be¬ tween development and conservation. The few at¬ tempts at conflict resolution by Government were not particularly successful. While State forest res¬ ervation made some hard-fought gains against ag¬ ricultural expansion, there was only a slow im¬ provement in the conservation reserve estate. However, the scientific community, mainly through the Australian Academy of Science and the Royal Society of Western Australia, stimulated community interest in the need for the establish¬ ment of an adequate conservation reserve system. Towards the end of this period there was also an emerging change in forest management toward more intensive productivity, both for timber and broader community benefits of catchment man¬ agement, recreation and habitat management. • Present Era: 1970-1995 There was State Govern¬ ment acknowledgment that sound grounds ex¬ isted for expansion of Western Australia's conser¬ vation reserve system, and also the need for better conflict resolution processes. During this recent period the conservation reserve system dramati¬ cally increased. This period has also been notable for its forest management controversy. This situation devel¬ oped as the wider community began to appreciate that production management for timber was fo¬ cussed on the manipulation of the natural struc¬ ture of production forests. Up to this time, there had been general acceptance that forest produc¬ tion management had little impact on wildlife habitat. There has been some rationalisation between con¬ flicting land uses and also rationalisation of re¬ serve management agencies through the creation of the Department of Conservation and Land Management. Overall, reserve management has become better resourced during this current pe¬ riod although not necessarily adequately resourced. Unresolved conflicts remain. Mecha¬ nisms to deal with many of these are absent or do not have broad acceptance. Colonial Years (1829 - 1889) The first 60 years Although the first European settlement in Western Australia was established at King George Sound (Al¬ bany) in 1826, it was an outlier of the New South Wales colonial administration. In practical terms the first settle¬ ment was on the Swan River in 1829, under the charge of Captain James Stirling as Lieutenant Governor. Early land development was based on garden, orchard and agricultural plots close to homesteads, spread along the Swan and Canning estuaries and lower river reaches on the coastal plain near Perth. However, pastoral land use expanded in the 1840s, from north of Perth to the Gingin area; northward from the already settled Avon Valley district to the Victoria Plains and New Norcia-Moore River areas; and southward from the Avon Valley, fol¬ lowing the inland route between Williams and Albany (Cameron 1981). This development avoided the heavily timbered country of the main forest belt. For more than half a century the colonists were sur¬ rounded by plentiful virgin country, and early conserva¬ tion regulations were aimed at protecting fauna rather than the habitat upon which the wildlife depended. For example, the early Game Acts had the intention of pro¬ tecting kangaroos and other fauna to meet the economic and sporting needs of the settlers, and to ensure con¬ tinuance of Aboriginal food sources. The first reserves for conservation purposes were set up under the Land Regulations for the Colony of 1872. Another set of Land Regulations enabling the creation of 226 Journal of the Royal Society of Western Australia, 79(4), December 1996 reserves was promulgated in 1887, and it was not until a new Land Act was passed in 1898 that previous land regulations were consolidated. Perth Park (now Kings Park) was the first of the State's significant bushland reserves to be created under the early Land Regulations. The first part of this Park was established on the sugges¬ tion made in 1871 by Governor Sir Frederick Weld. Fol¬ lowing gazettal of the 1872 Land Regulations, he was able to approve the setting aside of 432 acres (about 200ha) for the purpose of "public park and recreation" on October 1, 1872. Of interest, 1872 was also the year that Yellowstone National Park had been set aside by the US Congress, although the national park idea had an earlier beginning than that of Cornelius Hedges for Yellowstone in 1870. In Europe in 1810, the English poet, William Wordsworth, put forward a notional idea for the Cumberland Lake District as being a 'sort of national property/ Some 40 years before Yellowstone, the Ameri¬ can artist George Catlin in 1832 conceived a similar con¬ cept for the Dakota prairies and its plains Indians pro¬ viding 'a nation's park'. Towards self government In 1880, the colony of Western Australia was granted representative government, a step toward eventual re¬ sponsible government (i.e. self-government). Towards the end of the decade, when the population was about 40 000, the colony was ready to assume responsible gov¬ ernment, and this move was spearheaded by Sir John Forrest, then Surveyor General and Commissioner of Crown Lands, and a Member of the colony's Legislative Council. Land policy and administration was the princi¬ pal issue driving self government and Forrest's report to the Legislative Council in 1889 on land policy was criti¬ cal of development constraints while an Imperial Gov¬ ernment in London still maintained a fair measure of control over the administrative affairs of Western Aus¬ tralia (Forrest 1889b). In the following year, 1890, a Western Australia Con¬ stitution Bill was presented by the Imperial Government to the House of Commons, aimed at granting self gov¬ ernment to the Colony. A Select Committee of the Com¬ mons had been appointed to inquire into the proposal for this remote and little-known colony, and it took evi¬ dence from its Governor, Sir Frederick Napier Broome, who had travelled to London. Forest conservation in the Colony was only briefly covered by questions from the Select Committee and the Governor's recorded answers are a fair assessment of the situation at that time (Anon 1890); • Is there any system of forest conservation; under what rules are the forests allowed to be worked? There has been some attempt at it , but forest conserva¬ tion is still in a very rudimentary state. There was an inspector of forests appointed some time ago, and there were some regulations made under legislation in con¬ nection -with the matter , but they have been insuffi¬ ciently carried out , and forest conservation can hardly be said to be initiated yet. The fact is that the whole of the south-west division is so thickly covered with for¬ ests that the great desire of everybody is to get rid of as many trees as they possibly can. Of course the day will come , and even now is , when the forests ought to be preserved from waste. • Practically , are people allowed to cut wood as they like? No, that is dealt with in the Land Regulations; they must obtain licences and pay fees. • But are they allowed to cut what trees they like? Yes there is venj little restriction upon them; not so much as there ought to be, l think: some system of working the forests carefully and not wastefully ought to be introduced. It is interesting to note that early Timber Regulations under Western Australian colonial land legislation, and later that of its own Parliament, restricted timber felling to extremely large trees, the dimensions of which would now be uncommon, particularly in jarrah forest. A starting point Since late colonial times Western Australia has been divided into several Land Divisions in which varying land release and settlement policies could be applied. It had long been thought that agricultural potential was basically confined to the South West Land Division, which initially extended from the Northampton area in the north through C underdin, to Bremer Bay near Al¬ bany in the south. It was originally intended that broadscale releases of land into private ownership (i.e. freehold) would be restricted to this Division and that land development elsewhere would be covered by leases of land retained in Crown (public) ownership. On the eve of self government in 1890, the South West Land Division was described (Forrest 1889a) as follows; This division contains 67,000 square miles [194 000 km2], and comprises the most temperate part of Western Australia. It is that portion of the Colony first settled, and in which about 39,000 out of the whole population of 42,300 reside. The South-Western corner is heavily timbered, and is fairly well watered and capable of supporting a large population. It is generally an undulating country, and, with the exception of the Darling Range and a fezo other small ranges, has no extensive mountain ranges. Numerous rivers enter the coast within this division, but they are all very short and merely drain the country within 100 miles [160 km] of the coast. The work and expense of clearing the land has proved very laborious and very great, but, when the ground is properly prepared, a fair crop can be depended upon. In its natural state it takes about 10 acres [4 ha] to keep a sheep, but with clearing and improving it will keep a sheep to two acres [0.8 ha], and in choice places a sheep to one acre [0.4 ha]. It is estimated that at the present time there are depasturing, within this division, 866,229 sheep, 52,415 cattle , and 29,786 horses. The climate is very good, and the rainfall varies from 14 inches [350 mm] in the northern and inland portions, to 45 inches [1125 mm] in the southern portion of the divi¬ sion. The area leased from the Crown for pastoral purposes at the present time is 24,514 square miles [63 450 km2]. 227 Journal of the Royal Society of Western Australia, 79(4), December 1996 Early Self-Government (1890 - 1929) Perth Park (Kings Park) Originally created in 1872, Perth Park was enlarged in 1890 at the suggestion of Sir John Forrest, who by that time had become Premier of Western Australia. In 1895, a management committee for the Park was formed and Forrest became its first President. Under his patronage, two important pieces of legislation were passed to both enable management of the Park and to provide security against it being alienated for other purposes. Both pieces of legislation were applicable to other reserves. The Parks and Reserves Act was passed in 1895, en¬ abling the government to create boards of management for reserves and providing such boards with powers to manage the areas vested in them. Under this legislation. Sir John Forrest's management committee became the Perth Park Board. In 1899, the Permanent Reserves Act was passed. It too was initiated by Sir John Forrest to Provide Parliamentary protection against the alienation for other uses of "Public Parks". However, his Parlia¬ mentary' colleague and also a member of the Perth Park Board, Sir Winthrop Hackett, had the original Bill changed in the Legislative Council to provide for the concept of classifying reserves as A, B or C, regardless of their purpose, and this was passed. Under Hackett's concept, class A reserves became the most secure, requiring an Act of Parliament to vary their boundaries, change their purpose or cancel them. B and C class reserves could be changed without the need for Parliament's consent, but Parliament had to be given an explanation in the case of B class reserves. Shortly afterwards, in 1901, the name of Perth Park was changed to Kings Park. In its short history to that point the Park had stimulated the production of the nec¬ essary primary legislative tools to protect and manage reserves - security of purpose, and the appointment of statutory managers with powers to manage (Australian Academy of Science 1962; Christensen 1992). Early forest reservation During the 1890s, forest conservation and manage¬ ment in the southern Australian states followed some¬ what parallel lines. In fact. Western Australia's first Con¬ servator of Forests, John Ednie-Brown, became a com¬ mon link in the efforts of several states to come to grips with forest conservation (Powell 1976). Forest reserva¬ tion and timber cutting initially came under the jurisdic¬ tion of colonial land administration agencies. However, forest management needs conflicted with the primary purpose of these "Land Departments", that of promot¬ ing land settlement and clearing policies for agricultural production. Whilst allowing timber to be cut aided settlers in tak- ing up the cleared land, retaining forest land as a sus¬ tainable resource was seen by Australian land adminis¬ trators as "locking up" the land. In Victoria there were proposals to establish "State forest" as early as 1865, but none of the several Forest Bills presented to Parliament between 1879 and 1892 became law. George Perrin, Victoria s first Conservator of Forests, continued to head a small division within the Lands Department; he died in office in 1900 and was not replaced for some years. Victoria at last proclaimed a Forest Act in 1907, and a new Conservator (A McKay) was armed with new pow¬ ers. Before becoming Western Australia's first Conserva¬ tor of Forests in 1896, John Ednie-Brown had been South Australia's first Conservator of Forests in the late 1870s and subsequently moved to NSW in 1889 to become that state's Director-General of Forests. Like George Perrin in Victoria, when Ednie-Brown came to Western Australia he headed a Woods and Forests "Department" (Le. Divi¬ sion) within the Lands Department. Ednie-Brown's un¬ timely death three years later also prevented any real advance in forest conservation and management in Western Australia, and as in Victoria no trained forester was appointed to replace him for many years. (Forests Department 1969; Underwood 1991). John Ednie-Brown recorded the following (Fraser 1903); & / am, never the-less, pleased to be able to state that the forest's of Western Australia are yet practically unharmed for all purposes of successful conservation and ordinary thinnings and clearings of the matured timber... Tin forests are nature s gift, and should be looked upon and dealt with accordingly , as an inestimable inheritance S^cat commercial and climatic value; besides much of the land upon which the best timber grows is, as a rule, of little or no value for agricultural purposes, and I main¬ tain, without fear of logical contradiction, that what is now upon it is the very best kind of crop that will ever be seen there. To destroy it therefore, for the sake of a few more blades of grass, is suicidal and reprehensible in the extreme. At about this time, a Royal Commission into forestry matters in Western Australia had been held and a Forest Advisory Board established to consider applications for timber cutting concessions. Without specific forestry leg¬ islation under which forested land could be reserved, protection against agricultural selection in Western Aus¬ tralia was afforded by the creation of numerous reserves under land legislation, generally having the purpose of "Timber: Government Requirements". By 1910 a dozen or so of these existed (Dept. Lands and Surveys File 2507/93 Vol. 3), totalling 261 095 ha and were adminis¬ tered by the Woods and Forests "Department" of the Lands Department, headed by an Inspector, using Tim¬ ber Regulations promulgated under land legislation. However, these reserves were Class C and therefore in¬ secure against disposal for farm release. South Dandalup Nature Reserve Focused on the catchment of the South Dandalup River between Pinjarra and Bannister, the former "South Dandalup nature reserve is sometimes also referred to as the "Pinjarra Wildlife Reserve" or "Bannister Flora and Fauna Reserve". It was Western Australia's first na¬ ture reserve, was large at 64 000 ha, had a scientific basis (foi that time) to its creation, and was subsequently can¬ celled to permit timber cutting and orcharding. In the 1880s scientific professionalism in Australia was increasing, and in 1888 the Australasian Association for the Advancement of Science was formed. The Asso¬ ciation held a meeting in Adelaide in 1893, and a con- 228 Journal of the Royal Society of Western Australia, 79(4), December 1996 cept was developed for establishing conservation re¬ serves within Australia. At this time, there was a scatter¬ ing of reserves close to capital cities that were based on scenic landscape and associated recreation interest. However, the Association's concept was oriented to¬ ward habitat protection, with the reserves controlled by local honorary trustees and supported by Government grants. The Association decided to approach the Western Australian Government to have Rottnest Island, near Perth, and the Abrolhos archipelago, near Geraldton set aside for this purpose. These suggested sites were not set aside, but Premier Sir John Forrest invited Bernard Woodward, the Director of the Western Australian Mu¬ seum, to select another site; he chose the South Dandalup forest area. At the time, Bernard Woodward considered this area to be difficult heavily-forested country, and so nobody was seeking it for agricultural development. The timber industry was then focussed further north at Jarrahdale and Canning Mills. The selection of the South Dandalup area was supported by the Premier and the reserve for "Flora and Fauna" was gazetted in 1894. It was 64 000 ha in extent, and was bound in time to cause problems in "locking up" so much land and timber close to settled areas. Unfortu itely, in 1894 the legislation that eventu¬ ally afforded Perth Park its protective management and security of purpose had not been passed. Had it been possible to invoke these provisions immediately, the fate of the "South Dandalup" nature reserve might have been different. Within a few years of its creation, attempts had started to have the timber in the nature reserve made available for commercial exploitation. John Ednie- Brown, the new Conservator of Forests, for example, commented in a memo to the Minister for Lands in 1897 (Lands Dept. File 2507/93); There is some very fine timber upon this reserve but it is simply going to waste and should certainly be utilised. Sir John Forrest apparently withdrew his previous support for the reserve, and agitation for timber cutting and agricultural release for orcharding in valley land continued. In 1901, conditional permission to cut timber on the reserve was granted. Woods and Forest staff sub¬ sequently proposed cancellation of the reserve, causing Woodward to intercede, in 1902 he sought legislation to have the reserve vested in Trustees, but this was rejected by Cabinet, which instead proposed that it be opened for timber cutting. In the following year, 1903, a Royal Commission on forestry recommended that further tim¬ ber cutting in the nature reserve should cease, but this too appeared to have had little affect. By 1907, various attempts to cancel or reduce the reserve's size had been warded off, but a permit to cut timber had been granted to a large milling concern, Whittaker Bros. On May 28 in 1907, Bernard Woodward delivered a lecture to the West Australian Natural His¬ tory Society (forerunner to today's Royal Society of Western Australia) and chose as his topic "National Parks and the Fauna and Flora Reserves in Australasia". In his talk. Woodward gave an account of the move¬ ment for conservation reserves in the various Australian states and New Zealand, and reported on the unsatisfac¬ tory situation of the South Dandalup nature reserve. Arising out of the ensuing discussion, the Society re¬ solved to petition the Governor to have the nature re¬ serve declared a national park, and be vested in trustees to manage. The petition was conveyed by Governor Sir Frederick Bedford to the Premier, J Moore, in August 1907. Pre¬ mier Moore sought a report on the matter from the Lands Department. Officials, including the Surveyor General and Acting Inspector General of Forests, op¬ posed the idea of vesting the reserve in trustees for a national park as it would "lock up" a large tract of valu¬ able jarrah forest for the "mere preservation of flora and fauna". By this time, a number of water catchment re¬ serves in the forest area further north, between the York Road and Jarrahdale, had also been declared under land legislation, totalling over 100 000 ha. It was therefore proposed to set apart the protected Mundaring Catchment Area as a reserve for the protection of "Na¬ tive Flora and Fauna". However this never occurred and the purpose of the South Dandalup nature reserve itself was converted to "Timber - Government Requirements" in 1911/ and subsequently became incorporated into State Forest no. 14 (Dept. Lands and Survey File 2507/ 93, Vols 1 and 2; Woodward 1907; Australian Academy of Science 1962; Conservation Council 1980; Christensen 1992). South-west cave reserves The existence of numerous caves in the Leeuwin- Naturaliste area had been known by early residents since at least the 1870s. As some of these became more widely known and visited, there was some local concern regarding the destruction of some of the glittering for¬ mations they contained. On the urgent representation of Mrs John Brockman, the Minister for Lands, George Throssel, commissioned an exhaustive survey and in¬ ventory to be made of the district's caves. Of particular interest is the approach taken to do this. In 1894, the first step was to declare a reserve of 6 600 ha (Reserve 2565) over vacant Crown land in the Marga¬ ret River-Augusta area. This was designed as a holding action to prevent further farmland releases until the cave survey and inventory had been completed. During the next several years, a Lands Department team under the Chief Inspector of Lands, Charles Erskine May, explored and catalogued the caves. Erskine May presented his report to the Government in 1900, in which numerous caves and the glittering decorations were eloquently described. By this time Western Australia had experienced the main goldrush era and the majority of its population was now residing in major towns in the less-hospitable Eastern and Murchison Goldfields but linked to Perth by an efficient rail system. In his report, May saw the development of the caves as contemporary resort attractions, with the provision of a "sanatorium" on the Margaret River as a 'national' opportunity of fostering health and recreation (Erskine May 1900); " . to the goldfields especially , the Margaret River should be the Blue Mountains of NSW, the Derwent of Tasmania or the Lakes of Nezv Zealand when they are making a holiday". 229 Journal of the Royal Society of Western Australia, 79(4), December 1996 The key to this concept was the existing rail link from Perth to Busselton, providing reliable and comfortable transport from the goldfields as well as from Perth. In 1901 a Caves Committee was set up to recommend on the implementation of Erskine May's report. This re¬ sulted in the cancellation of the original 'holding' re¬ serve and the simultaneous gazettal of a dozen smaller cave reserves in 1902. Unlike the besieged "South Dandalup" nature reserve (which had been gazetted before enactment of the Permanent Reserves Act), all were afforded Class A security and vested in the Committee as a properly constituted board of management under the Parks and Reserves Act. Caves House at Yallingup was developed as a guest house by the Board, and nearby Yallingup Cave was provided with electric lighting to prevent formations from being damaged by other smoke-producing forms of illumination. Caves Road was constructed to link about a dozen "fully stepped and staired" caves to the accommodation centre at Yallingup and the railhead at Busselton. From Busselton, horse-drawn conveyances initially took visitors to their accommodation and, later, motor vehicles were used. A marketing innovation was the issue of "single cost" excursion vouchers that cov¬ ered both the rail and road journeys, accommodation and cave entrance fees. Subsequently, the Caves Board had the caves at Yanchep vested in it and it established camping and tent-hire facilities in that reserve. The Caves Board existed until 1910, after which it ceased to function. Responsibility for the cave reserves was transferred to the State Hotels Department in 1914, which had responsibility for running hotels in wheatbelt "frontier" towns to accommodate the many prospective land seekers, agents and commercial travellers active during that era. When the Department was closed in 1960, administration of the cave reserves reverted back to the Lands Department. Sometime later the reserves were passed on to the State Gardens Board. This Board had been formed in 1920 to manage a number of small parks and gardens within Perth's city precinct but soon found itself responsible for a number of national parks and other unvested bushland reserves, often distant from Perth. The cave reserves eventually became the core of the Leeuwin-Naturaliste National Park (Erskine May 1900, 1903; Australian Academy of Science 1962; Fraser, 1903; Mulcahy et al. 1988). John Forrest National Park Prior to 1895, the Eastern Railway to Northam and its link to the Great Southern Railway to Albany, passed through the Darling Scarp south of Greenmount and was routed via present-day Darlington and Mundaring. A better grade was subsequently located via the Jane Brook valley, north of Greenmount, and this alternative route was opened in 1895. The genesis of John Forrest National Park is directly linked to this new railway, and the reliability and comfort that rail travel provided. Near the point where the railway entered the valley and Jane Brook leaves is the popular Rocky Pool. This was only a short downhill walk from the new Swan View Railway Station. The first part of today's national park was set aside here in 1895 in the form of two re¬ serves, one for "Public Park" and the other for "Public Utility". In 1900 a more extensive reserve for "Parkland" was created upstream of Rocky Pool and excursion sid¬ ings constructed. A year later in 1901, the "Public Util¬ ity" reserve near Rocky Pool was also converted to the purpose of "Parkland". While official correspondence and plans in 1901 re¬ ferred to the area as a "national park", it was not until 1926 that the main reserve had its purpose changed from "Parklands" to "National Park and Native Game". For quite some time afterwards the Park was referred to as "The National Park". At one time, all three reserves were vested in the former Greenmount Road Board, but came across to the newly formed State Gardens Board just prior to the Great Depression. After World War II, the State Garden Board was transformed into the original National Parks Board (and like the Kings Park Board, also operated under the Parks and Reserves Act), to be subsequently replaced by a National Parks Authority in 1976 (operating under its own Act), and in 1971 the Park was officially named "John Forrest National Park" (Aus¬ tralian Academy of Science 1962; Dept. Land Adminis¬ tration reserves database). Other "Hills" attractions close to Perth had similar histories relating to rail accessibility, although not all sites were available for parkland reservation. When con¬ structed, Mundaring Weir was an immense attraction and the construction railway became an excursion rail¬ way to the site. Subsequently motor transport replaced this, and it was the drive to Mundaring Weir via Kalamunda that was later to promote early legislation for protecting wildflowers. The road access from Gosnells railway station to Victoria Reservoir was also promoted as a tourist drive. Further south, the Serpen¬ tine Falls also had an early tourist history with people initially travelling via the railway to Mundijong. In 1911, it was reported that "trainloads of excursionists" were visiting the falls every wildflower season (National Parks Authority 1992). Pemberton National Parks The two principal national parks that were originally created in the karri forest near Pemberton are part of several reserves collectively referred to as the "Pemberton National Parks". Following John Ednie- Brown's death while the State's first Conservator of For¬ ests, Mr V G Richardson, an Inspector of Forests, held the position of Acting Conservator. The first national park in this area was proposed by Richardson in 1901. The Acting Conservator sent a file minute to his superior, the Under Secretary for Lands suggesting that a "fine patch" of karri forest on the Vasse Road "be reserved as a sort of National Park, so as to allow posterity to see what the virgin karri forests were like". Cecil Clifton, the Under Secretary of Lands, was enthusiastically supportive and hoped that the se¬ lected patch would have "some really noble trees on it". Richardson then charged Forest Ranger H S Brockman to select one or two patches of forest, including Beedelup Falls if possible, up to nine or ten thousand acres (about 4000-4500 ha) in area "as it would be inadvisable to lock up too large an area of good land as a Forest Reserve." Within 20 days of Richardson making the original suggestion to Cecil Clifton, Brockman had returned a map showing two blocks of karri forest that should be 230 Journal of the Royal Society of Western Australia, 79(4), December 1996 reserved as National Parks, Beedelup and Warren. They were subsequently gazetted as timber reserves rather than as national parks, probably so that the Acting Con¬ servator of Forests could maintain a protective interest. However, not being 'A' Class, pieces of both reserves were progressively whittled away to provide farm grants and by 1909 parts of the remainder were sought for timber cutting and timber railway routes bv the State Sawmills Department. The Under Secretary for Lands finally settled this insecure situation by having the boundaries of the reserves rationalised to allow passage of the timber railways, having both reserves include re¬ placement areas of forest, and persuading the Govern¬ ment to secure them once and for all as Class 'A' na¬ tional parks to protect the original concept initiated by the Acting Conservator of Forests. Thus, both national parks became a reality in 1909 after being progressively shifted from their original locations (Dept, of Lands and Survey File 74/01). The Forests Department A federation of Australian self-governing States was created on January 1, 1901. Forestry, however, does not seem to have been considered very seriously in the con¬ ventions leading up to federation and forest ownership and management remained with the individual states rather than being transferred to the national govern¬ ment. In general, although the timber industry was an important part of the economics of the States, scientific forestry was not vigorously pursued before 1910 in Western Australia, Tasmania or Queensland. The first Interstate Forestry Conference was held in 1911 and academic forest training facilities were being established in some states. Scientific forestry advanced even further in 1914 when the British Association for the Advancement of Science convened a special conference in Australia (Powell 1976). One of the visitors was D E (later Sir David) Hutchins, a leading forester with wide experience in India and South Africa. Hutchins was sub¬ sequently invited to report on the forests of each state and this report was published in 1916 by the Western Australian Government (Hutchins 1916), indicating the government's deep interest at this time in coming to grips with forest management in Western Australia. Pro¬ fessional forestry was given a boost by Hutchins activi¬ ties and his report became the basis of a new move to¬ wards an Australia-wide management policy. In Western Australia, in 1916, the State Government appointed C E Lane Poole to be the new Conservator of Forests. A graduate of a highly regarded European for¬ est training centre and with subsequent colonial forestry experience, Lane Poole was primarily responsible for drafting Western Australia's Forest Act (Forests Depart¬ ment 1969; Underwood 1991) and drew on parts of Hutchins' report. Passed in December 1918, the Act re¬ ceived the royal assent in January 1919 and a new For¬ ests Department was formed, separate from the Lands Department. Of additional importance, the Act enabled State forest and timber reserves to be created under its own provisions, with 'conveyancing processing con¬ ducted by the Forests Department itself, instead of using Land Act provisions and Lands Department officials. One of the first tasks of the new Forests Department was to survey and classify Crown lands for their timber potential. Although by 1921 close on a million ha of forested land had been classified for retention from agri¬ cultural development, reservation as State forest had be¬ come a slow process. The Government of the day still showed more interest in agricultural land settlement, and a number of farm settlement schemes had come into operation following the close of the Great War in 1918. These had the objective of settling returning sol¬ diers and encouraging immigration from Britain. How¬ ever, a change of government in 1924 resulted in sub¬ stantial increases in State forest reservation. By the end of 1929, the centenary year of the founding of Western Australia, a million ha of State forest had been pro¬ claimed. Amongst the first areas to be covered by de¬ clared State forest under the new legislation were the Mundaring and Perth water supply reserves, that were once considered for the additional purpose of nature conservation in lieu of the cancelled "South Dandalup" nature reserve (Forests Department 1969; Mulcahy et at. 1988; Underwood 1991; Christensen 1992). From its earliest years, the Forests Department and its Conservator set about promoting the principles of forest preservation and management, and 'Arbor Day' was es¬ tablished as an annual institution. In 1920, the Depart¬ ment under Lane Poole published "Notes on the Forests and Forest Products and Industries of Western Austra¬ lia", to present in 'popular form' information on the ob¬ jectives of forest management. A second and enlarged edition was published in 1921, in which the silvicultural concept of "sustained yield" is explained (Forests De¬ partment 1921); Forest Capital and Forest Interest: The "capital" of a forest may be defined as the total amount of marketable timber it contains: the "interest" is the total annual growth of the trees: in other words , the percentage by which they increase in volume . In a forest in which a large proportion of the trees are mature or over-mature , the ratio of annual growth is relatively small; in the case of a forest which contains a large pro¬ portion of young and immature trees , the ratio of annual groivth is relatively large. In a "natural" forest, that is a forest which has received little or no attention at the hands of a skilled forester, the annual growth is much less than is the case in a " cultivated " forest. In a "natural" forest there are to be found many imperfect trees - mis¬ shapen, fire-damaged, or otherwise defective and incapable of developing into trees suitable for milling purposes. In a "cultivated" forest all these imperfect trees would have to be removed, and the space they occupied filled with trees capable of developing into marketable timber. Forest cul¬ tivation, therefore, increases the yield of timber: in other words, increases the annual interest on the forest capital. "A Primer of Forestry" was also produced and issued to schools through the Department of Education. The Second Edition of this work, published in 1925, describes the 'cultivation' of a natural (or wild) forest system to produce the so-called "normal" forest of centuries-old European tradition (Department of Education 1925); The ' cultivated ' forest already referred to is the ideal of modern scientific forestry. In France, Belgium , and Ger¬ many the whole of the forests are cultivated, and in other European countries the process of converting 'wild' or ' uncultivated ' forest to ' cultivated ' ones is proceeding 231 Journal of the Royal Society of Western Australia, 79(4), December 1996 apace. In Australia, with the exception of certain planta¬ tions , mainly of exotic trees, the forests are still unculti¬ vated, but in even/ State the process of conversion is be¬ ing pushed on. The effect of cultivation upon a natural forest is to increase very materially the amount of timber which the forest can yield annually without in any way diminishing its productive power. It was this concept of converting a natural forest eco¬ system into 'younger' and greater wood-yielding pro¬ duction forests that primarily led to the environmental concerns of a future generation. Karri National Parks In the early days of the Swan River Colony, very little was known of the south coast region. Although the settlements of Albany and Augusta are amongst the State's earliest (1827 and 1830 respectively), both were established from the sea and their hinterlands and the coast between them was little known. Coastal landings were difficult due to cliffs, beaches with formidable breaking waves, and inlets with sand-barred mouths. Behind the coastline are lines of swamps and river inlets and estuaries, and a hinterland of heavy forest. In 1830/31, an exploration party led by Captain Ban¬ nister attempted to determine a line that could be fol¬ lowed and developed as an overland route linking Perth to Albany. In the vicinity of "Nornor-up" (Nornalup) the party passed through forests of "blue gum". Bannis¬ ter was amazed at their size (Cross 1833); . if others had not seem them, I should be afraid to speak of their magnitude; I measured one, it was breast high, forty tzvo feet [14m] in circumference; in height, before a branch, 140 or 150 [about 50m] we thought at least, and as straight as the barrel of a gun . Some time afterwards, the existence of such forests of karri trees around Nornalup Inlet was confirmed by former sealers at Albany, and a syndicate of Albany resi¬ dents chartered a boat to examine this potentially lucra¬ tive resource. However, while the forest of enormous trees was confirmed, so was the difficulty of shipping timber out of the shallow and barred estuary. In 1832 a party of sailors were sent in an open whaleboat to ex¬ plore the little known coast inshore. They became stranded on Warren Beach when sheltering from a storm, and walked much of the future D'Entrecasteaux National Park coast to reach Augusta. A decade later, the naturalists James Drummond and John Gilbert at¬ tempted to reach Albany from Augusta (i.e. King George Sound and Cape Leeuwin areas, respectively). However, after spending several days in a bewildering maze of swamps in the Scott River-Gingalup Swamps area, they turned back to Augusta. None of these expeditions pro¬ vided information that the region's timber and land re¬ sources could be readily exploited. Later in the century Ferdinand von Mueller produced a report on Western Australia's forest resources which described the extent of karri forests near the Warren, Shannon, Donnelly, Walpole and Gardiner Rivers. He considered that karri timber would doubtless become important for the timber trade if inlets of the south coast wilderness, between Cape Augusta and Albany, could be developed for shipping. Development headed towards the south coast, in the vicinity of Pemberton and Denmark, in the late 1890s, and it was railways that were used to access forest and to transport milled timber to ports at Bunbury and Albany. However, the forested Frankland and Shannon valleys posed problems for heavy-railway development, which would have assisted forest clearing for agricultural expansion. By the mid 1920s more and more of the forest in the region was being protected by State forest reservations. However, wider knowledge of the south coast region increased as the milling industry expanded. Group Settlement schemes based on dairying were established in some areas cleared of forest, sometimes by the wasteful process of tree ring-barking. Above all, the new popularity of the motor car and their adventurous owners meant that previously unvisited places were be¬ ing promoted. The 1924/25 edition of The Western Australian Tour¬ ists Guide and Hotel and Boarding House Directory, un¬ der the heading 'The Inlets and Rivers of the South Coast - A Great National Asset", stated (Government Tourist Bureau 1925); The zvhole region is a succession of scenic beauties and desirable spots which offer innumerable attractions to the sportsman. Those who wish to make the best of the Inlet Country must camp out , and the only difficulty they are likely to experience is that of deciding where to pitch their tents in a district which presents so many favourable situ¬ ations. Among the rivers and inlets one finds a country that is unhackneyed, for its delights can be enjoyed only by those who are prepared to camp out and to put aside for a time the formalities of city life. But none can go there without benefit from a health point. It is proposed that a big area adjoining Nornalup shall be reserved for all time as a National Park and tourist resort, and the proposal is a wise one. Meantime, in that marvellous country Nature is still primitive, and that fact alone is a powerful lure to many. The original national park was referred to as the Nornalup National Park, after the name of the inlet on which it was focussed, and adjacent town. A series of contiguous reserves for "National Park" were first set aside here in August 1924, with some small additions in 1926. While the total area was over 12 500 ha, most of the reserved land was Class "C". The beginnings of this Park are more adequatelv told, however, by Fernie & Femie (1989). The Nornalup Reserves Board was set up under the Parks and Reserves Act, on the December 5, 1924 and the national park at Nornalup Inlet was placed under its control. The Board consisted of the Under Secretary for Lands, the Surveyor General and the Conservator of For¬ ests. On the October 27 1939 the Town Planning Com¬ missioner was added to the Board. The Director of the Tourist Bureau became a member of the Board on the 22nd March 1946. Although Land Department records in¬ dicate very little activity by the Board in controlling and improving its reserves, by-laws were laid down by the Board to control the park. The Board was cancelled in 1947 and all its reserves were vested in the State Gar¬ dens Board. The members of the cancelled Board were already members of the State Gardens Board. By the late 1920s a significant town had developed at nearby Pemberton, focused on timber milling and ser- 232 Journal of the Royal Society of Western Australia, 79(4), December 1996 vicing the district's farm settlers. This led to the forma¬ tion of another board of management that became re¬ sponsible for the “Pemberton National Parks". During 1928, moves were made by the Pemberton Parents and Citizens Association to have the hillside opposite the town on Big Brook reserved for scenic purposes. The Minister for Lands agreed and on May 16 the Pemberton National Parks Board was constituted under the Parks and Reserves Act. Comprising basically local people, the Board subsequently had the Warren and Beedelup Na¬ tional Parks and other reserves vested in it. Local Forest Department staff provided significant assistance with management of these Parks and the Board remained ac¬ tive until 1976, when it was dissolved and the "Pemberton National Parks" were transferred to the then newly-formed National Parks Authority (Austra¬ lian Academy of Science 1962; Fernie & Fernie 1989). Conservation reserves elsewhere in WA A few flora reserves were created in the period up to 1929, but mainly to protect "novelties" like the Albany pitcher plant, boronia, and the red flowering gum. Large unvested National Parks were also created over the Stirling and Porongorup Ranges. However, of particular significance, an "A" Class flora and fauna reserve was gazetted over Barrow Island in the State's remote north¬ west in 1908, and in 1929 the Abrolhos Islands off Gerald ton were declared a reserve for the "Protection of Flora, Public Recreation and Tourist Resort" and vested in a Board of Control (Australian Academy of Science 1962). The period during and after the Great War saw an extensive expansion of the agricultural development in the wheatbelt, adjacent to the Darling Botanical District. This was aided by an extension of the railway network and a decline of employment opportunities in the gold¬ fields. However, as more and more of the wheatbelt was being allocated for farming, little thought was given to creating strategic conservation reserves in that region. During the State's centenary year of 1929, the Govern¬ ment published a book in celebration, "A Story of a Hundred Years", with chapters on pertinent topics be¬ ing written by specialists in those fields. Naturalist and noted wild flower artist, Emily Pelloe, contributed a chapter on the State's flora and made the following plea (Pelloe 1929); The Eastern Wheat Belt sand-plains , for instance , sole habitat for hundreds of precious species, are fast becoming devastated without reserve in the interests of agriculture. It may be that in 2029, regret will be expressed that so little effort was made as far back as 1929 to ensure the preservation of the rare and beautiful flora. To deny fu¬ ture generations the right to enjoy its wonders is to de¬ serve the censure of the unborn. That little was done to set aside conservation reserves in the Wheatbelt even after 1929 was regretted by 1979, only fifty years later, as one of the World s most exten¬ sive land clearing programs was rapidly effected, rival¬ ling current rainforest clearing in the Amazon basin. However, the extensive forest reserve system declared by 1929 had protected much of the adjoining Darling Botanical District from a similar fate. The Next 40 Years (1930 - 1969) Reserve management institutions In 1933, various separate statutes and regulations dealing with land settlement and administration, includ¬ ing the Permanent Reserves Act, had been consolidated into a single Land Act. However, the Parks and Reserves Act remained separate. A State Gardens Board had existed since 1920, con¬ trolling ten small park and garden reserves in and around the city of Perth. Other reserve management boards existing around 1930 which had also been estab¬ lished under the Parks and Reserves Act of 1895 were the Kings Park Board, Rottnest Island Control Board, Nornalup Reserves Board, Pemberton National Parks Board, and outside the Darling Botanical District the Abrolhos Islands Board of Control. The State Gardens Board was seen by the State Gov¬ ernment as an instrument in which to vest a variety of reserves with which nobody else was prepared to be burdened. For example, when the former Greenmount Road Board wished to relinquish its responsibility for the reserves comprising John Forrest National Park, they were vested in the State Gardens Board. The Board also progressively picked up small Darling Range reserves that had become popular weekend excursion and picnic sites, such as Serpentine Falls and along the Canning River in the Araluen area, and in 1947 reserves making up the Walpole - Nornalup National Park (Australian Academy of Science 1962). It was not until 1956 that the State Gardens Board was transformed into the National Parks Board of West¬ ern Australia, still operating under the provisions of the Parks and Reserves Act. By this time, the Board also had the enormous Stirling Range National Park (created in 1913) and the Porongorup Range National Park (estab¬ lished in 1925) vested in it. However, while Yanchep "Park" and the large camping reserve at Hamelin Bay near Augusta came under the Board's control, the cave reserves in the Yallingup-Augusta area remained under the management of the State Hotels Department (Aus¬ tralian Academy of Science 1962). Up to this time very little thought had been given by the State Government to protecting habitat for native fauna. Fauna protection was largely a matter of hunting control over particular species via Game Acts. For many years these had been administered by the Fisheries De¬ partment, while "vermin" regulations had been admin¬ istered by agricultural-based agencies, including the former Rabbit Department, and local vermin boards. A significant change commenced with the passing of the Native Fauna Protection Act of 1950. Prior to this, in 1944 a Fauna Advisory Committee had been set up to advise the Minister responsible for administering the Game Act of 1912. Members of the Advisory Committee included the Chief Inspector of Fisheries, Curator of the Western Australian Museum, and Dr D L Serventy, then a fisher¬ ies research officer with the CSIRO and noted naturalist. Under the new Act, the Fauna Protection Advisory Committee became a corporate body in which wildlife reserves, created via the Land Act, could be vested (Aus¬ tralian Academy of Science 1962). 233 Journal of the Royal Society of Western Australia, 79(4), December 1996 Thus, as Western Australia entered a new develop¬ ment phase stimulated by post-war reconstruction after World War II, there were three principal reserve man¬ agement bodies seeking to protect forest land, flora and scenic landscape, and wildlife habitat. All were con¬ fronted by the competing demand for land for agricul¬ tural expansion. The members of the National Parks Board tended to have an interest in flora protection, so that new reserves with a flora emphasis tended to be¬ come national parks regardless of their landscape qual¬ ity. Some so-called "National Parks" reserved in this pe¬ riod more closely fit the criteria of nature reserves. Re¬ serves sought by the Fauna Advisory Committee (later to be renamed the WA Wildlife Authority, and its Act changed to the Wildlife Conservation Act) became nature reserves, even though some contained high value land¬ scape. Expansion of the conservation reserve and forest reserve estates After 1929, expansion of State forest reservation tailed off in the 'timber belt' of Western Australia's south-west. Up to 1954 only about 100 000 ha more were added and most of this occurred in 1938. As happened in the pe¬ riod following the Great War, after World War II a con¬ flict of interest again arose between demands for un¬ allocated forested land to be released for post-war agri¬ cultural development. In 1953, at the suggestion of the Forests Department, an inter-departmental Land Utilisation Committee was set up by Cabinet to recom¬ mend on the allocation of remaining uncommitted for¬ est land for forest conservation, water catchment protec¬ tion, and agricultural development. This replaced a simi¬ lar committee initially established in 1943 but which had generally been inactive (Forests Department 1969; Christensen 1992). Largely as a result of the deliberations of the Land Utilisation Committee, up to 1958 a further 200 000 ha were added to State forest reserves in the south-west. Little consideration seems to have been given in this process to separately reserving some of the forest land as specific reserves for nature conservation. At the time (1957), the Forests Department was still producing edu¬ cational material which explained its intention to con¬ vert virgin natural forest to a restructured production forest of rotational [tree] age classes, "ideal" for sus¬ tained commercial harvesting (Forests Department, 1966); Managed and Unmanaged Forests Possibly the idea of the cultivated forest is not entirely clear. One may ask just what advantages has a managed forest over a virgin forest if the latter is able to provide trees in perpetuity , maintain a stable composition and the soil fertility. It is not always realised that the virgin for¬ est is not the most economical forest from man's point of view. Virgin forests have no normal succession of trees of all ages , but by virtue of their great age, usually contain a majority of over-mature trees. Such trees lose more wood by internal decay each year than they are capable of putting on in their condition of poor vigour. Their large crowns overtop and suppress young trees and prevent germination of seed on the forest floor. Managed forests , on the other hand , aim to have the opti¬ mum number of vigorously growing trees per acre. Once a tree slackens off in increment , it is removed to make way for more vigorous young ones coming on. All age classes of trees are represented in the forest so that as trees are cut for milling, others are available to produce a future final crop with a minimum lapse of time. Spacing between the trees is also controlled to permit an adequate area for growth of each member and the minimum of competition from neighbours. Managed Forests , therefore, are cultivated to produce the maximum amount of desir¬ able produce while guaranteeing that there is always a crop ready to replace the one that is removed for utilisation In 1959, the inter-departmental Land Utilisation Com¬ mittee was superseded by another inter-departmental committee established by Cabinet, the Crown Land Tri¬ bunal. Its task was to ascertain use allocation of "sparsely timbered" Crown land and in the ensuing de¬ cade the Tribunal's consideration of some 400 000 ha of Crowm land resulted in the forest reserve estate being increased by only about 100 000 ha, much of which was coastal plain earmarked for conversion to pine planta¬ tion. Thus by 1969, State forests and timber reserves in the South West that were under Forests Department con¬ trol totalled some 1.8 million ha, with additional exten¬ sive forest estate in the Eastern Goldfields (Forests De¬ partment 1969; Christensen 1992). During the same period, relatively little improvement had been made to the conservation estate in the State's forested south-west, in adjoining agricultural regions and Wheatbelt, nor in the State generally. One redeem¬ ing feature was the creation in June 1930 of 18 timber reserves (under provisions of the Land Act) to protect mallet groves in the Great Southern area of the wheatbelt at the request of the Forests Department (former Forests Dept, file, now CALM file 011479 F3003 53). These to¬ talled some 130 000 ha, much of which was later con¬ verted into nature reserves, such as Lake Magenta and Dragon Rocks Nature Reserves. Nature conservation-ori¬ ented groups, like the Royal Australian Ornithologists Union (RAOU) and the Western Australian Naturalists Club, continually lobbied the State Government over the years to retain strategic bushland habitats in areas being opened up for farming. This tended to increase in the 1950s, following the end of World War II, with even the Country Women's Association of WA also making pleas, along with newer conservation groups like the Tree Society and WA Branch of the Society for Growing Australian Plants, now the Wildflower Society of West¬ ern Australia (Dept. Lands and Survey File 2507/93, Vol. 4). On the whole, the coverage of conservation reserves across Australia was poor and not systematic. As a con¬ sequence, in 1958 the Australian Academy of Science appointed a Committee on National Parks and Reserves to inquire into the situation. The committee included Dr W D L Ride, then Director of the Western Australian Museum. In turn, the Academy of Science committee itself formed State sub-committees, and the WA Sub¬ committee on National Parks and Nature Reserves was chaired by Dr Ride, producing a report to the Academy in 1962 on conservation reserve proposals. A very 234 Journal of the Royal Society of Western Australia, 79(4), December 1996 limited number of copies of the WA Sub-committee report had been produced and, at the instigation of the Royal Society of Western Australia, the National Parks Board collaborated with the Australian Academy of Science and published an edited version (in 1965) for popular use. Now that the Academy of Science sub¬ committee's proposals for new conservation reserves in the State was publicly available, there was strong community support for their implementation by the State Government Annual Report of WA Dept. Lands & Survey, WA National Parks Board, 1966; (Australian Academy of Science 1962, 1968; Conservation Through Reserves Committee 1974; Ride 1975). In 1969, the WA Cabinet established another inter¬ departmental committee, the Reserves Advisory Coun¬ cil, to recommend on the Sub-committee's conservation reserve proposals and other similar proposals from a variety of sources that had been put forward for the establishment of conservation reserves. While a number of the Reserve Advisory Committee's recommendations were adopted by Cabinet, a backlog began to develop over proposals where there was opposition from mining interests. A culmination of this deteriorating situation occurred later in 1969 after the Hamersley Range Na¬ tional Park (now Karijini National Park), proposed by the Academy of Science WA Sub-committee, was created on the recommendation of the Reserves Advi¬ sory Council. Reservation of this extensive (630 000 ha) Park apparently by-passed a Mines Department vetting process and received Cabinet approval in the face of a number of iron ore development agreements covering the area. The Reserves Advisory Council then ceased to meet because outstanding recommendations remained , unapproved (Annual Reports of WA Dept. Lands and Survey and WA National Parks Board, 1969, 1970; Conservation Through Reserves Committee 1974; Ride 1 1975). While the conservation reserve estate in the forested south-west had not substantially increased up to 1969, a \ number of conservation reserves had been set aside in the northern part of the Darling Botanical District. This occurred as new land for agricultural expansion was be¬ ing made into sandplain areas, following success with trace-element fertilisers. The allocation of conservation reserves in these more recent times was a reflection of Government response to changing community attitudes. These were outside of the main forest belt and largely comprised sandplain (kwongan) vegetation, and in¬ cluded the Moore River National Park. Elsewhere in the State, the Government Botanist, C A Gardiner, had been successful in persuading the State Government to create a series of large strategic nature reserves along the South West and Eremean Botanical Provinces interface. These comprise today's Kalbarri and Cape Arid National Parks and Jilbadji Nature Reserve, and were gazetted in 1954. Gardiner was also successful in having the Fitzgerald River National Park set aside at the same time, initially as a nature reserve (Australian Academy of Science 1962, Ride 1975). As the State Government had apparently refused to consider any further significant conservation reserve proposals, this helped stimulate the initiation of a public campaign by the Conservation Council of Western Aus¬ tralia in 1969. A Modern Reserve System (1970 - 1995) Environmental Protection Authority and its Conserva¬ tion Through Reserves Committee The influence of the "Conservation Campaign" mounted by the infant Conservation Council in 1970 is often overlooked as one of the stimuli to government action in that era. Formed in 1967 by a consortium of the main conservation groups existing in Western Australia at the time (including the Tree Society, Western Austra¬ lian Naturalists Club, Kings Park and Swan River Soci¬ ety), the Council initially acted as clearing house for in¬ formation and a means of co-ordinating group activities where there was a common interest. These ranged from opposing inappropriate development plans for Kings Park, opposing mining proposals in numerous conser¬ vation areas (such as Jilbadji Nature Reserve) through the Mining Wardens Court, and promoting expansion of the State's conservation reserve system. Matters came to a head in 1969 when serious indus¬ trial pollution was manifesting itself in Cockburn Sound and other coastal areas; with continued mining threats to key conservation reserves, culminating in bauxite mining in State forest and mineral sand pegging in popular resort areas around Geographe Bay to Augusta; and State Government disinclination to consider any more proposals for conservation reserves. At the time, the State Government was also promoting its "million acres a year" farmland release scheme involving virgin bushland in the Wheatbelt. This stimulated the formation of more groups that were fighting local issues; even established Progress As¬ sociations and the Farmers Union were becoming in¬ volved in opposing some industrial development and mining proposals. Many more groups became affiliated with the Conservation Council and in 1969 it began planning a campaign. The campaign focused on three basic actions, wide promotion of a Conservation 'Bill of Rights' outlining the Council's view of Government ac¬ tion that was needed, the gathering of a massive petition for presentation to Parliament, and an organised torch¬ light parade on Parliament after it opened for its last session before the 1971 Parliamentary election (Churchward 1991). Amongst the 'Bill of Rights' requests were the forma¬ tion of a Ministry of Conservation, creation of another 8 million ha of conservation reserves in the State, and a reform of mining legislation applicable to conservation areas. Public and media support for the campaign was high and the government of the day responded with an inquiry into the 1904 Mining Act being conducted and the slow but eventual great improvement for the protec¬ tion of conservation areas, as well as initial environmen¬ tal protection legislation. Flowever, the legislation was not formally proclaimed prior to the 1971 Parliamentary election and the incoming Tonkin Government initiated stronger legislation and the creation of an Environmen¬ tal Protection Authority (EPA). In turn, the Authority immediately set in train a project to promote the estab¬ lishment of a State-wide conservation reserve system. The latter commenced in 1972 and coincided with grow¬ ing public concern over native forest management as the concept of clear-felling forest areas for an export 235 Journal of the Royal Society of Western Australia, 79(4), December 1996 woodchip industry was being planned in several States, including Western Australia. At its first meeting after being established in 1971, the EPA set up a Conservation Through Reserves Commit¬ tee (CTRC) to examine Western Australia's existing con¬ servation reserve system and to report to the EPA on proposals to significantly expand it. Dr W D L Ride was appointed Chairman. In its work, the CTRC was assisted by a Technical Sub-committee of specialists in the fields of demography, zoology, botany and geology, and by an executive officer. The latter group were seconded from government departments to work full-time on the project (Annual Reports of the WA Environmental Protection Authority, 1972, 1973; Conservation Through Reserves Committee 1974; Ride 1975). The CTRC divided the State into 12 fairly homog¬ enous regions, or "Systems", and began producing re¬ ports on them for public comment and EPA consider¬ ation, the first being released in 1974 and the last in 1981 (the so-called "Green Books"). In turn, the EPA issued its own reports (the so-called "Red Books") with recom¬ mendations to the State Government for new reserves and the expansion of some existing ones in the various Systems. Within this project, the Darling Botanical Dis¬ trict is basically covered by Systems 1, 2 and 6. In these regions some public land was still available to extend the reserve system but the majority of the uncleared land had already been included in State forest reserves (Ride 1975; Mulcahv 1991; O'Brien 1991; Christensen 1992). The major new conservation reserves created in the Darling Botanical District through the CTRC process were the adjoining D'Entrecasteaux and Shannon Na¬ tional Parks in System 2. Both were extremely contro¬ versial with the former being initiated by local Forests Department staff and vigorously opposed by various land development interests, and the latter being initiated by the CTRC itself and vigorously opposed by the For¬ ests Department and other agencies. However, the For¬ ests Department did support conservation priority man¬ agement of some parts of the Shannon Basin. The other major controversial CTRC proposal was in System 1, involving the connection of the discontinuous Leeuwin- Naturaliste National Park (based on the original cave reserves once managed by the 'old' Caves Board) through the addition of other miscellaneous reserves and private land purchases. Because of several contentious issues, the EPA arranged for a special inter-departmen¬ tal committee to review the CTRC proposals within Sys¬ tems 1 and 2 and its report (the "Brown Book") was considered by the EPA in conjunction with the CTRC proposals before making recommendations to the State Government. While the EPA did not support the origi¬ nal Shannon "Basin" national park proposal, various conservation groups maintained a continuous campaign for the concept and were successful a decade later in having the area excised from State forest and converted into a national park (Department of Conservation and Land Management 1987a; Mulcahy et al. 1988; Thomas 1988; O'Brien 1991; Christensen 1992). Elsewhere in existing State forest areas, the CTRC and EPA basically supported Forests Department plans to have identified high nature conservation and recreation areas retained as forest reserve under Forests Depart¬ ment management but with priority for those uses. However, the Conservation Council campaigned for a jarrah forest conservation reserve in the Nanga area near Dwellingup (System 6) to be taken out of State forest and managed separately (Conservation Council 1980; Mulcahy et al. 1988; Department of Conservation and Land Management 1990; Christensen 1992). To some ex¬ tent, this was seen as a means of redressing the loss of the nearby original "South Dandalup" nature reserve in 1911 to forest production interests. The State Govern¬ ment eventually agreed to this proposal as well, and today's Lane Poole Reserve is the result. System 6 as a whole was the most complex of the CTRC regions be¬ cause it included most of the State's population and fewer opportunities to significantly increase the conser¬ vation reserve estate. However, out of a number of the proposals involving the Swan Coastal Plain came the concepts of urban bushland protection and regional park focuses which would later be used to channel commu¬ nity and government efforts to put protective measures in place (Environmental Protection Authority 1983; Mulcahy 1991). Reviewing national park management During the 1971-1973 Labor government under Pre¬ mier John Tonkin, there was significant change in State Government policy toward environmental protection. The Environmental Protection Act and the EPA (along with the EPA's 'conservation through reserves' project) became established. Toward the end of its term, the La¬ bor government also examined the operation of the former National Parks Board with a view to improving its structure and management capability. Basically, three options were canvassed; • transferring national park management to an ex¬ isting and better resourced Forests Department (the "Queensland model"), • amalgamating the wildlife management division of the existing Department of Fisheries and Wild¬ life with the National Parks Board, to provide a National Park and Wildlife Service with an ex¬ panded conservation management role (the "NSW model"), and • converting the existing "Board" into an Authority operating under its own Act with improved fund¬ ing and staff resources, and a professional Direc¬ tor. The abysmal national park allocation in Queensland and growing controversy over forest management in Western Australia mitigated against the first option. While there was community support for establishing a National Parks and Wildlife Service similar to the ser¬ vice then recently established in NSW, it was opposed at bureaucratic levels. Key officials associated with the Na¬ tional Parks Board and senior staff in the Department of Fisheries and Wildlife associated with wildlife manage¬ ment argued against an amalgamation. An in-coming conservative coalition government, under Premier Sir Charles Court, subsequently drafted legislation for a new National Parks Authority, which came into being in 1976. While the new Western Australian Government maintained a programme of improved environmental management, controversy continued over its adoption 236 Journal of the Royal Society of Western Australia, 79(4), December 1996 of the EPA recommendation that the Shannon basin re¬ main reserved as State forest, and there was strong op¬ position to the proposed D'Entrecasteaux National Park which would "lock up" potential farmland in areas such as the Pingerup Plains. In 1979 and 1980 the Legislative Council agreed to the appointment of Select Committees to inquire into, and report on national parks in Western Australia. Both were prompted and chaired by the Hon Sandy Lewis MLC, Member for the South West Province. The scope of the 1979 proposal was all-embracing with an impossible two-month deadline for tabling a report. The resultant eight-page report, without conclusions or recommenda¬ tions, did little justice to the time and effort of the many people who made written submissions or appeared be¬ fore the Select Committee (Select Committee 1979). The second attempt was more successful and resulted in a report exceeding 150 pages. However, this time many of the people who participated in the 1979 inquiry didn't participate in the second inquiry. Community participation in the second Select Committee inquiry is also remarkable for the amount of interest shown by local governments and regional development organisations which were pro-land development, and deeply concerned over EPA proposals for expanding the State's conservation reserve system (Select Committee 1981). Very little became of this Parliamentary commit¬ tee report. However, while the concept of a National Parks and Wildlife Service was supported, so was a na¬ ture conservation precinct concept which allowed re¬ source utilisation. It is clear that the idea was based on the English-style "National Park" (where landscape con¬ servation is imposed on utilised rural land), but in re¬ verse where development could be permitted in what started out as fairly pristine land, such as the Pingerup Plains area of the D'Entrecasteaux National Park (per¬ sonal communication with the Select Committee). Forest management priority areas The period 1975 to 1987 was one of great controversy over forest management and forest reserve issues. Close questioning in Parliament by the opposition Member for Warren, the Hon David Evans MLA, had revealed the extent of forest resource over-cutting that had been al¬ lowed for some time (WA Parliament, Legislative As¬ sembly questions on notice No. 19 17/10/1974, and No. 67 28/11/1974), while the Forests Department was still promoting the concept of "sustained yield" in its educa¬ tional material (Forests Department 1971); A well-managed forest may be likened to a bank account in which the forest itself and the forest soil represent the capital invested and held in trust , while the annual growth in timber (the increment of the forest) represents the interest earned. The fundamental idea of Forest Man¬ agement is to harvest this increment only , and to pre¬ serve and improve the forest capital for increased future production. It was probably with the State's first woodchip pro¬ posal in 1973, from the Forests Department itself, that the full significance dawned on the interested commu¬ nity of the long-term change to forest structure that was proposed. Essentially, the concept of natural forest con¬ version to "ideal" or so-called "normal" forest involved restructuring to a regular succession of all age classes to provide regular timber crops of young trees (around 100 years old, with even younger "thinnings") in perpetuity. To do this conversion on a large scale in the mixed karri- marri forest included extensive felling of marri trees to waste, and therefore without recovery of expensive fell¬ ing costs. However, marketing marri woodchips for ex¬ port as a feedstock for paper making would make clear- felling and subsequent forest conversion economical (Annual Reports of the Environmental Protection Authority 1973, 1974; Forests Department 1973). While publicity over the past two decades has tended to focus on the concept of extensive clear-felling as a silvicultural practice, this is just a means to an end, being the conversion of a natural forest ecosystem to a manipulated tree crop system, in which the long-term environmental implications still remain unclear. On top of the Shannon basin and woodchip issues, the Forest Department created another controversy with the proposal to re-afforest Phytophthora-dieback affected native forest with pines in the Donnybrook Sunklands (Blackwood Plateau), between Nannup and Margaret River. The statement of intent, issued in September 1975, proposed a program to convert some 60 000 ha of jarrah forest, considered to be of poor quality for timber utilisation, to pine plantations over a 30 year period (Forests Department 1975). However, like the Shannon basin issue, the matter was determined by commitments in the State Platform of the Australian Labor Party, and the program was discontinued when the Burke Labor Government came into office in 1983. Faced with growing community concern and pres¬ sure within its own ranks (R Underwood, personal com¬ munication), and a realisation also that multiple-use for¬ est management needed to be a significant feature of its operations, the Forests Department embarked upon a system of identifying forest blocks having a variety of appropriate priorities for management (Mulcahy et al. 1988; Christensen 1992). Termed management priority areas (MPA), the concept first appeared in General Working Plan number 86 in 1977. It was also promoted in issues of the Department's publicity journal. Forest Focus (No. 18, August 1977; No. 22, January 1980) as part of its new zoning policy for forest management. The key MPAs in a conservation sense were those iden¬ tified as having wildlife or recreation priorities. The MPA concept also generally comprised a core/outer buffer system, and identified MPAs were well estab¬ lished when the CTRC/EPA System 6 study was under¬ way in iate 1970s/early 1980s. While the EPA's 1983 recommendations for System 6 (Environmental Protec¬ tion Authority 1983) supported the MPA concept and that such areas remain under Forests Department man¬ agement, the Authority also recommended that legisla¬ tion be provided to make the priority purpose secure and that any change from conservation or recreation re¬ quire Parliamentary endorsement. It also advocated that management plans for conservation MPAs be prepared and made public as soon as practicable. These proposals were aimed at maintaining security of purpose for forest conservation areas that remained under Forests Depart¬ ment management. There was still community concern that forest conser¬ vation areas identified by the Forests Department were inadequate in extent and would remain insecure in 237 Journal of the Royal Society of Western Australia, 79(4), December 1996 terms of management while still vested in the Depart¬ ment. Further change was to come within a couple of years after the Forests Department was merged with other agencies following promulgation of the Conserva¬ tion and Land Management Act in 1984. The Conservation and Land Management Act One of the first actions of the incoming Burke Gov¬ ernment in 1983 was to have land resource management in the State's south-west reviewed, because of continu¬ ing controversies over the use of forest and coastal land. The mechanism created for this was the appointment of a Land Resource Management Task Force. In an interim report, the Task Force sought and was granted an exten¬ sion of its terms of reference to take in the whole of Western Australia. The principal outcome of the final report of the Task Force, in January 1984, was the amal¬ gamation of the Forests Department, National Parks Au¬ thority and the wildlife component of the Department of Fisheries and Wildlife to form a new department to manage State forest and timber reserves, national parks and nature reserves, and wildlife generally. At the same time, it advocated that separate advisory bodies for for¬ estry and nature conservation be established to assist the Minister, and that a higher level "Policy Commission" be created above the Department (Task Force 1984). The structure eventually adopted by the Government was that of an amalgamated department, with three ad¬ visory bodies (referred to as 'controlling bodies' in the Act); Lands and Forests Commission (LFC), Timber Pro¬ duction Council (TPC), and National Parks and Nature Conservation Authority (NPNCA). The concept was strongly opposed by conservation groups generally (as well as local government and other sectors of the community), who would have preferred a simpler amalgamation to form a National Parks and Wildlife Service similar to the NSW model. As a com¬ promise, the Government provided a more pro-active role for both the LFC and NPNCA, with forest and con¬ servation reserves being vested in each, respectively, (rather that in the Department), and that these two "con¬ trolling bodies" be responsible for management plan production and monitor the Department's implementa¬ tion of management plans approved by the Minister, and policy formulation. This structure was put in place in 1985 after the passing of the 1984 Conservation and Land Management Act, with the amalgamated depart¬ ment becoming the Department of Conservation and Land Management (CALM). A feature of the new de¬ partment was a decentralising of administration through the establishment of regional and district offices throughout the State. At the same time, the Wildlife Con¬ servation Act was modified to relate to wildlife manage¬ ment generally, and nature reserve provisions were transferred to the Conservation and Land Management Act. In 1987 the former Forests Department's Forest Work¬ ing Plan No 87 was due to expire. CALM, in conjunc¬ tion with the LFC and NPNCA, set about replacing it with three separate region management plans into which the main State forest areas fell, - CALM's Northern (now Swan), Central and Southern Forest Regions. Draft re¬ gion management plans were issued for public comment and extended the forest conservation/recreation MPA system. It also proposed that the purposes of MPAs be given Parliamentary security through amendment of legislation. However, from within the Labor Party, the Government was persuaded to promote conversion of conservation/recreation MPAs to national parks, con¬ servation parks or nature reserves and have them vested in the NPNCA rather than remain as "zones" within State forest still vested in the LFC (Department of Con¬ servation and Land Management 1987b, 1987c, 1987d; Mulcahy et al. 1988; Shea & Underwood 1990; Underwood 1991; Christensen 1992). The approved 1987 Forest Region Management Plans therefore set in train the transfer of over 500 000 ha from forest reserves to conservation reserves. Subsequent amendments to these three region management plans (Department of Conservation and Land Management 1994) increased this transfer proposal by a further 50 000 ha approximately. The mechanics of the transfers are administratively complex and the proposals are being progressively implemented. Under provisions of the Conservation and Land Management Act both the NPNCA and LFC are obliged to monitor the implementation of formal management plans. However, to date (1995) no status report has been publicly released. During the pe¬ riod 1989-1993, the EPA maintained a stance that State forest retained for timber production should also be managed to maintain its wildlife habitat values, and ad¬ vocated the establishment of forest research and moni¬ toring overview mechanisms. In 1994, the State Govern¬ ment adopted as policy the ecological sustainability of timber production from native forests reserved as State forest (Department of Conservation and Land Manage¬ ment 1994). Conclusion In the "design" of conservation reserves, the selection of many of the early reserve sites was a matter of someone's personal fancy but usually focused on an outstanding resource-base. For example. Kings Park (1872) provides panoramic vistas across Perth and Melville Waters; Beedelup and Warren National Parks (1901) contain truly "noble" karri trees, and appealing riparian landscape; the same can be said of John Forrest (1895) and Serpentine National Parks. However, there was more science and "system" to the selection of the former "South Dandalup" nature reserve (1894) and the original cave reserves that formed the nucleus of today's Leeuwin-Naturaliste National Park (1902). It was another fifty years or so after the initial cre¬ ation of conservation reserves around the turn of the century, however, before science and "system" were ap¬ plied to reserve selection in Western Australia. The first were the strategic botanical province interface reserves promoted by Charles Gardiner in the early 1950s, and resulting in today's Kalbarri and Cape Arid National Parks and some other areas. While the Australian Acad¬ emy of Science began raising community interest in the need for a systematic approach to conservation reserve selection, the eventual result in Western Australia was the EPA's more extensive decade-long ^Conservation Through Reserves' program. Reserve selection in that process was dependent upon constraints such as the availability of uncleared land, other competing utilisation interests ( e.g . mineral development) and local 238 Journal of the Royal Society of Western Australia, 79(4), December 1996 community attitudes. Thus, boundary outcomes were influenced by politics and compromise. By the time the EPA's Conservation Through Re¬ serves Committee had completed its first major assess¬ ment of over two-thirds of the State (Conservation Through Reserves Committee 1974), the Academy of Sci¬ ence had produced a further report, outlining the need to preserve genetic diversity, along with other concepts relating to reserve '"design" (Australian Academy of Sci¬ ence 1975). This was supplemented by the "Specht Re¬ port" (Specht et al. 1974) concerning the conservation of major Australian plant communities. However, while these were able to have some influence on new conser¬ vation reserve proposals for Swan Coastal Plain areas in the EPA's subsequent System 6 study, many of the "re¬ gional park" proposals were based on landscape/ recre¬ ation potentials identified through early (1950s/ 1960s) town planning concepts. Nevertheless, both reports in¬ fluenced the Forests Department in its selection of con¬ servation MPAs within State forest reserves (Mulcahy et al. 1988; Christensen 1992). Through several processes, the conservation reserve situation within the Darling Botanical District is under review and doubtless will be subjected to further fine- tuning. Currently there is a heavy focus on the concept of 'urban bush' conservation and regional parks, and the situation was adequately summed up by the EPA (1991): There is much unfinished business in implementation of conservation through reserves. We have done as well as possible for late starters in agricultural areas , we are do¬ ing better than before in forest areas , but continue to lag in pastoral areas, and have reached about halfway in implementation in urban and near urban areas. In 1996 the EPA is scheduled to review aspects of CALM's management of State forest, including forest research and monitoring issues. The content of the Authority's report and its conclusions may indicate if the objectives of multiple-use management are being met satisfactorily or not. 1995 has been taken as the centenary year for Kings Park, as it was 100 years ago that the serious business of management commenced with the appointment of its management board. However, it has taken 100 years for the production of a formal management plan for this Park, effectively the State's first regional park. The story of urban bushland protection and regional park estab¬ lishment and management is one with a "town" plan¬ ning focus and of long-term indecision. However, some key decisions appear about to be made by the present State Government on urban bushland policy and re¬ gional park administration. Compiling a history on this topic requires branching into the realms of town and strategic planning, and is a story in its own right. It is a task best left for a little while longer, until State Govern¬ ment direction becomes clearer. References Anon 1890 Report from the Select Committee on the Western Australia Constitution Bill, together with the proceedings of the Committee, Minutes of evidence and appendix [House of Commons, London]. In: British Parliamentary Papers Relating to Australia 1890 (Vol 32); Irish University Press, Shannon (1970). Australian Academy of Science 1962 Report of WA Sub-commit- tee, Committee on National Parks and Nature Reserves. Aus¬ tralian Academy of Science, Canberra [This mimeographed document was limited to about a dozen copies. The more commonly used reference is an edited version produced in 1965 by the Acad¬ emy and the former National Parks Board of Western Australia]. Australian Academy of Science 1968 National parks and reserves in Australia (Report of Australian Academy of Science Com¬ mittee on National Parks and Reserves). Australian Academy of Science, Canberra. Australian Academy of Science 1975 A natural system of ecologi¬ cal reserves in Australia. Proceedings of a symposium - A national system of ecological reserves in Australia (ed F Fenner). Report No. 19, Canberra. Beard J S 1981 Vegetation survey of Western Australia; Swan. 1:1000 000 Vegetation Series, Explanatory Notes to Sheet 7. University of Western Australia Press, Perth. Cameron J M R 1981 Ambition's fire. The agricultural colonisation of pre-convict Western Australia. University of Western Australia Press, Perth. Christensen P E S 1992 The karri forest. Its conservation signifi¬ cance and management. Department of Conservation and Land Management, Perth. Churchward B 1991 A conservationist's view. In: Reflections on 20 years (ed R Hughes). Environmental Protection Authority, Perth, 33-41. Conservation Council of Western Australia 1980 Jarrah reserve: A proposal for a major reserve in the northern jarrah forest of Western Australia. Conservation Council, Perth. Conservation Through Reserves Committee 1974 Conservation reserves in Western Australia. Report to the Environmental Protection Authority. Department of Conservation and Envi¬ ronment, Perth. Cross J 1833 Journals of several expeditions made in Western Australia during the Years 1829, 1830, 1831 and 1832. J Cross, London. Department of Conservation and Land Management 1987a Shan¬ non Park and D'Entrecasteaux National Park management plan 1987-1997. CALM, Perth. Department of Conservation and Land Management 1987b Northern Forest Region management plan 1987-1997. CALM, Perth. Department of Conservation and Land Management 1987c Cen¬ tral Forest Region management plan 1987-1997. CALM, Perth. Department of Conservation and Land Management 1987d Southern Forest Region management plan 1987-1997. CALM, Perth. Department of Conservation and Land Management 1990 Lane Poole Reserve Management Plan 1990-2000. CALM, Perth. Department of Conservation and Land Management 1994 Forest management plan 1994-2003. CALM, Perth. Department of Education 1925 A primer of forestry with illustra¬ tions of the principal forest trees of Western Australia. Gov¬ ernment Printer, Perth. Environmental Protection Authority 1983 Conservation reserves for Western Australia as recommended by the Environmental Protection Authority - 1983: The Darling System - System 6. Report 13, Department of Conservation and Environment, Perth. Environmental Protection Authority 1991 Foreword. In: Reflec¬ tions on 20 years, (ed R. Hughes). Environmental Protection Authority, Perth. Erskine May C 1900 Report on the limestone caves of Western Australia. In: Department of Lands and Survey Annual Re¬ port 1900, Appendix No. 4. Government Printer, Perth. Erskine May C 1903 Description of the Limestone Caves of West¬ ern Australia. In: Fraser 1903 35-41. 239 Journal of the Royal Society of Western Australia, 79(4), December 1996 Femie L & Femie G 1989 In Praise of a Park. The origins and history of the Walpole-Nornalup National Park. Fernie & Fernie, Walpole. Forests Department 1921 Notes on the forests and forest prod¬ ucts and industries of Western Australia. Government Printer, Perth. Forests Department 1966 Forestry in Western Australia. Govern¬ ment Printer, Perth. Forests Department 1969 Fifty Years of forestry in Western Aus¬ tralia. Supplement to the 1968/69 Annual Report. Govern¬ ment Printer, Perth. Forests Department 1971 Forestry in Western Australia. Govern¬ ment Printer, Perth. Forests Department 1973 Marri wood chip project. Environmen¬ tal impact statement. Forests Department, Perth. Forests Department 1975 Afforestation with pines in the Donny- brook Sunkland. Statement of intent. Forests Department, Perth. Forrest J 1889a Notes and statistics respecting the six Land Divi¬ sions of Western Australia. In: Enel. No 44 British Parliamen¬ tary Papers Relating to Australia 1890 (Vol 32); Irish Univer¬ sity Press, Shannon 1970. Forrest J 1889b Report on the land policy of Western Australia from 1829 to 1888. Government Printer, Perth. [Presented to the Western Australian Legislative Council , third session of 1889. Copy in State Archives , PR1234]. Fraser MAC (1903) Notes on the natural history, etc. of Western Australia, being extracts from the Western Australian Year Book for 1900-1901. Government Printer, Perth. Government Tourist Bureau 1925 The Western Australia tourist's guide and hotel and boarding house directory. Government Printer, Perth. Hutchins D E 1916 A discussion of Australian forestry, with spe¬ cial references to forestry in Western Australia. Government Printer, Perth. Mulcahy M J, White B J & Havel J J 1988 Parks and reserves in the south-west forests. In: Case Studies in Environmental Hope (ed P Newman S Neville & L Duxbury). Environmental Protection Authority, Perth. Mulcahy M 1991 System 6 revisited. In: Reflections on 20 years (ed R Hughes). Environmental Protection Authority, Perth, 65-67. National Parks Authority 1992 Annual Report for 1991/92. Gov¬ ernment Printer, Perth. O'Brien B 1991 The EPA in the beginning. In: Reflections on 20 years (ed R Hughes). Environmental Protection Authority, Perth, 5-18. Powell J M 1976 Environmental management in Australia, 1788 - 1914. Oxford University Press, Melbourne. Pelloe 1929 Floral Glory. In: A story of a hundred years. Western Australia 1829-1929 (ed H Colbatch). Government Printer, Perth, 128-148. Ride W L D 1975 Towards an integrated system. A study of se¬ lection and acquisition of national parks and nature reserves in Western Australia. In: Proceedings of a symposium - A National system of ecological reserves in Australia (ed F Fenner). Australian Academy of Science Report No. 19, Canberra. Select Committee 1979 Report of the Select Committee appointed by the Legislative Council to inquire into and report upon the current position of national parks within the state. Legislative Council of the WA Parliament, Perth. Select Committee 1981 Report of the Select Committee of the Legislative Council on national parks. Legislative Council of the WA Parliament, Perth. Shea S & Underwood R 1990 Forests for the future. Landscope 6(l):35-43. Specht R L, Poe E M & Broughton V H 1974 Conservation of major plant communities in Australia and Papua New Guinea. Australian Journal of Botany (Supplementary Series 7): Task Force 1984 A report to the government of Western Austra¬ lia by the Task Force on Resource Management in Western Australia. Government of Western Australia, Perth. Thomas W 1988 The Shannon - saving a special area. In: Case Studies in Environmental Hope (eds P Newman, S Neville & L Duxbury). Environmental Protection Authority, Perth, 43-59. Underwood R 1991 Keepers of the forest. In: The WA 4W Driver's Guide (Glove Box Edition). Westate Publishers, Perth, 18-23. Woodward B H 1907 National Parks and the fauna and flora reserves in Australia. West Australian Natural History Soci¬ ety Journal 2(4):13-27. 240 Journal of the Royal Society of Western Australia, 79:241-248, 1996 Assessing the conservation reserve system in the Jarrah Forest Bioregion N L McKenzie1, S D Hopper2, G Wardell-Johnson13 & N Gibson1 1 Science and Information Division, Department of Conservation and Land Management, P O Box 51, Wanneroo WA 6065; 2 Kings Park and Botanic Gardens, Fraser Avenue, West Perth WA 6005; 3 present address: Department of Biology, University of Namibia, Private Bag 13301, Windhoek Namibia Abstract Recent reviews have assessed the comprehensiveness of the conservation reserve system over the northern part of the Jarrah Forest Bioregion in terms of vegetation complexes. The complexes were distinguished in terms of geomorphology and dominant vegetation. The least reserved complexes are those of the Darling Scarp, Blackwood Plateau, Collie Coalfields and those with agriculturally desirable soils. Available maps can be used to estimate the reserved area of each of the Bioregion's vegetation complexes or geomorphic units, but there are not enough data on patterns in biodiversity to assess other facets of its adequacy, even in the northern part of the region. A quadrat-based regional survey is necessary if the representativeness of the area's reserve system is to be assessed from an ecosystem perspective. The sampling would need to cover a range of the different components of the biota (perennial floristics, vertebrates and selected invertebrate taxa). Such surveys are time-consuming and expensive. Current studies of rare, restricted and endemic spe¬ cies, of weed, feral animal and pathogen impacts, and of forest management effects, need to continue in parallel. A recent national assessment of the major gaps in Australia's conservation reserve network suggests that other Western Australian regions have higher priority for regional survey than the Jarrah Forest Bioregion. Seventeen of the State's 26 bioregions have a smaller proportion of their area reserved than the Jarrah Forest, and of these, 11 have a more biased reserve system. Should we commit scarce survey resources to the Jarrah Forest Bioregion? Introduction The high rainfall part of south-western Australia is covered by the Darling Botanical District (Beard 1982). This district is made up of the Swan Coastal Plain (Drummond Botanical Sub-district), the northern jarrah forest (Dale Sub-district), the southern jarrah forest (Menzies Sub-district) and the karri forest (Warren Sub¬ district). We provide a brief review of the data that are avail¬ able to assess the coverage of the existing conservation reserve system for the communities indigenous to the Dale and Menzies Sub-districts. Together, these two sub¬ districts form the Jarrah Forest Bioregion (Thackway and Cresswell 1995). We then describe and discuss previous assessments of the reserve system in this area. Current knowledge Beside bioclimatic surface models (Nix & Gillison 1985), various environmental maps at 1:250,000 scale cover all or most of the bioregion; • surface lithology (Wilde & Low 1978, 1980; Wilde & Walker 1982, 1984), • landforms and soils (Churchward & McArthur 1980; Churchward et al 1988; Tille & Lantzke 1990; Churchward 1992). The 1:250 000 coverage Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 is 90% complete whereas the 1:100 000 coverage is about 50% complete; R Harper, pers. comm.), and • vegetation structure (Beard 1982; Heddle, Loneragan & Havel 1980). At such scales, considerable heterogeneity is averaged within each map-unit. There are also 1:25 000 Aerial Photographic Interpre¬ tation (API) maps of tree-canopy floristics, mapped at 1: 15 800 and ground-truthed along tracks and roads (F J Bradshaw et al, unpublished). They were produced by the Forests Department during the 1950s and 1960s, and cover virtually all the forested parts of the Bioregion except; • parts of the far northern end (north of 31° 45’ S, although Julimar State Forest was mapped), al¬ though the mapping has recently been extended, • most of the eastern edge between 32° 00’ S and 34° 30’ S, although some parts, such as Dryandra State Forest, are covered by 1:15 800 scale "Mallet Clas¬ sification" vegetation maps made during the 1930s, and • the far south-eastern corner, east of 117° 40’ E. Most of the areas not covered at this scale are re¬ served for nature conservation and mapped at an equivalent scale, or are private land that is extensively cleared. The mesic bioregions of south-western Australia are remote from their eastern Australian counterparts, being isolated by 2000 km of the semi-arid and arid environ- 241 Journal of the Royal Society of Western Australia, 79(4), December 1996 ments that comprise districts such as the Nullarbor and the Great Victoria Desert. They contain many endemic species that are relicts from previously wetter and less seasonal climatic periods. Examples from the Jarrah For¬ est Bioregion include the frogs Geocrinia alba and G. vitellina, and the eucalypts Yarri (Eucalyptus patens), E. relictua and E. Virginia (Wardell-Johnston et al. in press). Their vegetation and floristics have been reviewed by Havel (1975, 1989), Wardell-Johnson et al (1996) and Wardell-Johnson & Horwitz (1996), and the rare flora by Kelly et al. (1990), while knowledge of the fauna has been reviewed by Wardell-Johnson & Nichols (1991), Abbott & Christensen (1994) and Wardell-Johnson & Horwitz (1996). Several factors, including the isolation, the small area and previous climate changes, have lowered diversity in the larger indigenous vertebrates of the south-western forests, but the invertebrates and small vertebrates are rich in species (Wardell-Johnson & Horwitz, 1996). In situ speciation is real in the south-west, for the fauna as well as the flora; Havel and others have demonstrated a rich mosaic of landforms, soils and floristic variation in the jarrah forest despite the area's great structural ho¬ mogeneity in vegetation, and Wardell-Johnson & Rob¬ erts (1993) concluded that the rich but fine-scale varia¬ tion in soil and landscape properties maintains barriers between closely related frog species. A recently discov¬ ered frog, endemic to the Bioregion, is the most geo¬ graphically restricted frog known from mainland Aus¬ tralia (Roberts et al., 1997). It is clear that many scales should be considered in any discussion of conservation in the Jarrah Forest Bioregion. However, coordinated work on the distribu¬ tion and biology of the naturally rare and of the re¬ stricted and endemic flora of the Bioregion has com¬ menced only in the last decade, and most studies have focused on the central and northern parts of the bioregion. Systematically-gathered data are still too frag¬ mentary to reveal their general patterns of occurrence. Autecological studies of fauna have concentrated on vulnerable species, especially mammals that have be¬ come either extinct or rare outside the forests, wood¬ lands and shrublands of the Darling District since Euro¬ pean settlement (e.g. Dasyurus geoffroii, Myrmecobius fasciatus, Bettongia penicillata , Macropus eugenii and Pseudocheiriis occidentalis; Burbidge & McKenzie 1989). The numbat study reported by Friend & Thomas (1995) is a good example. The first three species are now the subject of detailed management programs (e.g. Orell & Morris 1994; Start et al 1995). Considerable research has also been carried out on Macropus eugenii, Pseudocheiriis occidentalis and Isoodon obesulus (e.g. P de Torres, pers comm.) Detailed information has also been collected on the distribution and habitat requirements of some rare frogs (e.g. Geocrinia alba and G. vitellina), reptiles (e.g. Ctenotus delli) and birds (e.g. Atrichornis clamosus). While the critical requirements of some vulnerable vertebrates are still unknown (e.g. Macropus irma and Setonix brachyurus), research programs are now in-place. Recov¬ ery plans are now being implemented for others such as Geocrinia alba (Driscoll et al. 1995). Most research on the disturbance ecology of flora and fauna in the Darling District has been directed to mining and rehabilitation, logging and regeneration, fire, plant disease and the introduced fox Vulpes vulpes (Wardell- Johnson & Nichols 1991; George et al. 1995, Abbott & Christensen 1996). Rehabilitation following bauxite mining has been studied in considerable detail (e.g. Nichols & Bamford 1985; Nichols & Watkins 1984; Majer 1990; Majer & Nichols, unpublished observations), as have the dynamics and management of dieback disease caused by Phytophthora cinnamomi in the jarrah forest (Shearer & Tippett 1989). Integrated studies have re¬ cently been commenced that will allow predictive-mod¬ elling of fauna responses following logging and burning in the jarrah forest (Wardell-Johnson & Nichols 1991; Friend 1993; K Morris, pers. comm.). Other studies are being undertaken to examine the occurrence of tree hol¬ lows in the forest and to determine the impact of timber harvesting on the availability of these hollows (Faunt 1992; Rhind 1996; K Whitford, pers. comm.). Clearing for agriculture (Beard & Sprenger 1984), exotic animals such as foxes (Kinnear 1993; Morris 1993; Friend et al. 1994), introduced weeds (Pigott & Gray 1993; Burke unpub¬ lished), and plant diseases such as those caused by spe¬ cies of Phytophthora, pose the greatest threats to the plant and animal communities of the Jarrah Forest Bioregion. According to Burbidge & McKenzie (1989) faunal changes have not been as dramatic in the Darling Bo¬ tanical District as in the pastoral and agricultural zones. Seven mammals, fifteen birds, two reptiles and three frogs from this district are currently gazetted as "fauna which is likely to become extinct or is rare". Unfortu¬ nately, data on their distribution within the Jarrah Forest Bioregion at the time of European settlement are scant, vague and localised; even the sub-fossil record is frag¬ mentary (A Baynes, pers. comm.). While some of these species were apparently rare or had restricted ranges in the bioregion originally, available records indicate that the ranges of most have contracted and/or become frag¬ mented (e.g. Bettongia penicillata , Setonix brachyuris. Macropus eugenii, Dasyurus geoffroii, Myrmecobius fasciatus , Pseudocheiriis occidentalis, Trichosurus vulpecula, Leipoa ocellata , Dupetor flavicollis and Ninox connivens). Wardell-Johnson & Nichols (1991) suggested that spe¬ cies with specialised habitat requirements and low dis¬ persal abilities were the first to decline (e.g. Macropus eugenii and Geocrinia alba). Other species persist within the bioregion only at Two Peoples Bay (the bioregion's south-eastern margin), although they were recorded elsewhere in its southern parts earlier this century (Potorous Iridactylus, Psophodes nigrogularis, Dasyornis longirostris and Atrichornis clamosus). The Western Bristlebird (Dasyornis longirostris) and Western Whipbird (Psophodes nigrogularis) also occur in the Esperance Plains Bioregion to the east. The only likely mammalian extinctions have involved species of adjacent bioregions with ranges that extended into the periphery of the Jarrah Forest Bioregion; Potorous platyops and P. tridactylus on the bioregion's south-eastern periphery, and Macrolis lagotis and Bettongia lesueur along its eastern periphery. Leeuwin's Rail (Rallus pect oralis) is the only bird thought to have vanished from the Bioregion; it was only known from peripheral southern areas. In explaining why the forested bioregions of the south-west have maintained more of their conservation 242 Journal of the Royal Society of Western Australia, 79(4), December 1996 values than the arid and semi-arid areas of Western Aus¬ tralia, the resilience of the forest is frequently cited (e.g. Underwood et al. 1991). This should not lead to compla¬ cency for there has been a rapid intensification of land- use demands in the Jarrah Forest Bioregion since the 1960s (Havel 1989). The end-results of more intensive use, including extensive clearing and urbanisation, have been documented in the adjacent Swan Coastal Plain (Drummond Sub-district), where 13 bird species de¬ clined in abundance in Kings Park between 1927 and 1988 (Recher & Serventy 1991), and at least 8 mammals are known to have become extinct (Rettongia lesueur, B. penicillata, Macropus eugenii, Rattus tunneyi, Pseudomys fieldi, P. shortridgei and Notomys mitchclli; Kitchener et al 1978, and Western Australian Museum records). In overview, many specialist projects have been car¬ ried out in the Darling Botanical District, but no attempt has yet been made to define the pattern of any group of biota (e.g. the plants or the vertebrates) throughout any of its four sub-districts. The most extensive quantitative studies so far have been of the vascular plants in the southern and central Drummond Sub-district (Gibson et al. 1994) and in the Tingle Mosaic (which overlaps with parts of the Warren and Jarrah Forest Bioregions; Wardell-Johnson et ah 1996), and of the wetland fauna of the Southern Forest Region (Horwitz 1994). Other¬ wise, current knowledge is summarised by the "Jarrah Book" (Dell et al. 1989) for the Dale Sub-district and by Wardell-Johnson & Nichols (1991) for parts of the Menzies and Warren Sub-districts. There is also consid¬ erable site-specific data for some localities such as the Worsley alumina leases of the eastern jarrah (Anon 1985) and of the alternative water supply sites (Anon 1987). Existing information is sufficient to demonstrate areas of high mammal richness such as the proposed Perup Na¬ ture Reserve (Anon 1987a), and some sites of high ende¬ mism, but does not allow an explicit understanding of the overall patterns of the biota across the study area. Reserve system assessments Existing network. The existing conservation reserve system, developed through the work of Havel (1975), is described by Anon (1987b) and Heddle et al. (1980). It comprises National Parks, Nature Reserves and "conser¬ vation parks". In previous reviews of the forest reserve system, some of the "conservation parks" were Manage¬ ment Priority Areas (MPAs), but were treated as conser¬ vation reserves because the relevant lands are to be Figure 1. The Jarrah Forest Bioregion showing subdistrict boundaries and Havel's "northern jarrah forest". 243 Journal of the Royal Society of Western Australia, 79(4), December 1996 vested in the National Parks and Nature Conservation Authority. Subsequently, these MPAs were enlarged and inter-connected as a series of "proposed Conservation Parks". The formal "Conservation Park" legislation is now in place, so the proposals that have been gazetted (virtually all) are conservation reserves for the purposes of this review. We have also included Lane Poole Reserve, which was declared in 1987 under Section 5g of the Con¬ servation and Land* Management Act. It should be noted that both the Conservation Parks and "5g" reserves in this region retain the optional purpose of bauxite min- ing. Figure 1 shows the Jarrah Forest Bioregion, with the "northern jarrah forest" cross-hatched to show the area that Havel assessed (modified from Havel 1989, and as mapped on page 394 of Dell et al. 1989). Land-classes. Heddle et al. (1980) integrated available climatic, floristic, geomorphic and vegetation structural data to produce a set of three maps showing vegetation complexes at a scale of 1:250 000. Havel (1989) con¬ cluded that maps such as these "... are , in fact , the only basis on which adequacy of coverage [by the conservation reserve system in the forested regions of south-western Australia] can be assessed at piresent". W ardelLJ ohnson & Nichols (1991) reached a similar conclusion. Given this basis, the comprehensiveness of the reserve system was assessed as early as 1980 (Heddle et al. 1980). Havel (1989) reviewed and updated this analysis (Table Table 1 Percent of the total area of the northern jarrah forest's vegetation complexes that are within reserves (from Havel 1989 and Anon 1994) \ Neither list of percentages corresponds exactly to the Jarrah Forest Bioregion (see Fig l)2. Vegetation % Representation in % Representation including Complex reserves (Havel 1989) new reserve proposals (Anon 1994) Helena (L/M) - 25.5 Williams - 0.9 Lowden 0.2 5.8 Darling Scarp 1.5 3.8 Michibin 3.5 9.0 Wilga 3.8 13.4 Muja 4.3 3.0 Cardiff - - Coolakin (L) 7.5 17.4 Dwellingup (H) 6.0 6.5 Yarragil (M/S) 5.7 6.7 Murray (L/M) 8.7 10.6 Murray (M/H) 8.9 18.1 1 In the National Forest Policy terminology (NFPS 1992), this as¬ sessment relates to the reserve system's "comprehensiveness" and, in part, its "adequacy" rather than to its "representative¬ ness"; 2 The figures in column 1 are based on mapped vegetation complexes within the System Six boundary (see Anon. 1993). The figures in column 2 are based on the total area of vegetation complexes depicted in maps published by the former Depart¬ ment of Conservation and the Environment (Heddle et al. 1980). The issue of conservation reserves in the Collie Coal Basin is being addressed by a detailed "Structure Plan". Complexes in the Drummond Botanical Sub-district (Swan Coastal Plain Bioregion) are excluded because they are outside the scope of this review. 1), concluding that there was no representation of those vegetation complexes in the northern jarrah forest that are centred in the adjacent regions and have been desir¬ able acquisitions for agriculture since European settle¬ ment (Helena, Williams River). The Forrestfield vegeta¬ tion complex, which straddles the western edge of the Jarrah Forest Bioregion, and Cardiff (on the Collie Coal measures) were also unrepresented. Havel's analysis also showed that the agriculturally- desirable complexes centred in the northern jarrah forest had less#than 5% representation (Wilga, Lowden, Michibin and Darling Scarp), as did the Muja vegetation complex of the Collie Coal Basin. Furthermore, Collie (the third complex of the Collie Coal Basin) was 5-10% reserved, as were the remaining complexes listed in Table 1, from: • the lateritic uplands and shallow valleys in the western high rainfall zone — Dwellingup and Yarragul respectively. • rocky valleys in the eastern low rainfall zone; Coolakin, and • deep major river valleys; Murray (L/M) and Murray (M/H). Havel (1989, pp 387-89) showed that the rest of the vegetation complexes had higher percentages re¬ served. These included the complexes; • on lateritic uplands of the low to medium rainfall zone, and in the shallow valleys and depressions in the high to moderate rainfall zone (10 - 15%); • on uplands and in the shallow valleys of the east¬ ern low rainfall zone, and • on infertile sands, steep rocky slopes of the monadnocks and swampy headwaters of the streams, as well as certain complexes of eastern dry lateritic uplands and of the deep valleys in the medium to high rainfall zone (20%+). Generally, the coverage by conservation reserves within the forested landscape was found to be greater in the east than in the west of the region, and better on the extreme and less productive parts of the landscape than on the more productive ones. Historically, alienation for agriculture preceded both conservation and forestry, thereby limiting the opportunities for subsequent reser¬ vation. Anon (1994) updated Havel's analysis, although somewhat different study area boundaries were used (Table 1). Since these assessments, the coverage of suit¬ able environmental maps has been extended across the Blackwood Plateau and Manjimup areas in the south¬ west of the Jarrah Forest Bioregion (Tille & Lantzke 1990 and Churchward 1992, respectively), and through south coastal areas from Northcliffe to Manypeaks (Churchward et al. 1988) which includes the south-east¬ ern corner of the Jarrah Forest Bioregion. In broad geo¬ graphical terms, the currently recognised forest, wood¬ land and shrubland communities that are least con¬ served by reserves are those: • on the western margin of the Dale Sub-district (=Darling Scarp), • agriculturally desirable surfaces throughout, • the Collie Coalfields, and 244 Journal of the Royal Society of Western Australia, 79(4), December 1996 • the north-western 30% of the Menzies Sub-district (=Blackwood Plateau) (Havel 1989 p 387; Keighery 1990). Discussion Gradients in community species composition The problems of "land-class" approaches to reserve assessment mainly derive from assumptions of homogeneity and determinism, and have been reviewed by McKenzie et al. (1989). Intuitive regionalisations (land-classes such as land-systems, land-units and structurally based vegetation complexes) allow large areas to be surveyed quickly and are objective (Pressey & Nicholls 1991), but are insufficient from a biodiversity perspective because they usually reflect some mixture of vegetation patterns, topography, lithology and/or climate. This does not necessarily reflect the pattern of the entire biota (or even an "averaged" pattern), and does not quantify or map internal heterogeneity (McKenzie 1984; Margules et al. 1991; Pressey & Bedward 1991; McKenzie et al 1992). In fact, land-classes are unlikely to be internally homogeneous and their boundaries may not be meaningful for many of the species in the region. The approach lacks spatial resolution because the relationships between the four layers (vegetation, topography, lithology and climate) are usually undefined in quantitative ways, and at¬ tributes such as species are not necessarily present at any point at any one time. This means that the generated data-base is intractable for more detailed analyses of eco¬ logical patterns. At fine scales, intuitive regionalisations also assume that compositional patterns are static, which they certainly aren't for mobile vertebrates. Thus, land-class approaches provide only a first ap¬ proximation of biological diversity patterns, and the re¬ serve network should be subsequently refined using the ecological realism offered by quadrat strategies of data acquisition (McKenzie & Sattler 1994). Point data-sets derived from quadrat sampling strategies provide the spatial resolution that is needed to address the continu¬ ous gradients underlying ecological patterns, and allow us to compare sites quantitatively (Margules & Austin 1994). These data allow species with similar responses to environmental attributes to be clustered so that patterns in species composition across landscapes can be mod¬ elled for reserve design, and the predictions can be tested (Margules & Nicholls 1994; McKenzie et al 1989 1991). Although quadrat sampling strategies are slow and expensive, significant cost-savings are offered by "gradsec" approaches to quadrat stratification (Gillison & Brewer 1985; Austin & Heyligers 1989) that systemati¬ cally sample regions along major environmental gradi¬ ents. Quadrat data from surveys in different study areas (e.g. districts) can be combined into a geographically more extensive data-base without loss of spatial resolu¬ tion, and quadrat-data also provide a basis for monitor¬ ing future changes (McKenzie et al 1991a). Jarrah forest context The recent reviews of the conservation reserve system in the northern jarrah forest have concluded that a thor¬ ough assessment of this reserve system's coverage in terms of the study area's biodiversity was not possible with available data (Havel 1989; Wardell-Johnson & Nicholls 1991). Havel (1989, p. 389) pointed out; The ahead y barely adequate representation of Yarragil (M/S) becomes even more critical when it is recognised that some of the component site-vegetation types , such as W and D, occur in topographical positions which render them susceptible to dieback infection , and contain a num¬ ber of species which are highly vulnerable. This is also true of the Muja and Collie vegetation complexes. According to the National Forest Policy Statement (NFPS 1992, p. 49) forest reserve systems should be as¬ sessed using three criteria; 1. comprehensiveness: include the array of commu¬ nity-types, 2. adequacy: ecologically viable and maintain integ¬ rity of populations, species and communities, and 3. representativeness: reasonably reflect the biotic di¬ versity of each community. While contemporary reserve selection algorithms can use land-class data to address "comprehensiveness" (Pressey & Nicholls 1989, 1991), data with more spatial and biological resolution are required to address "ad¬ equacy" and "representativeness" (Margules et al. 1991; Woinarski & Norton 1993; McKenzie & Sattler 1994). Detailed knowledge of species biology and commu¬ nity dynamics and regional context are required to as¬ sess adequacy. Standardised, point-based data-sets that encompass a wide variety of organisms are required to assess representativeness. Wardell-Johnson & Nichols (1991) discuss this subject in a south-western forest con¬ text. While such data sets have been collected in parts of the forest (e.g. at Boddington, Worsley 1985), the prob¬ lem is that they are too localised, and/or limited to a particular sort of organism (e.g. endangered plants, Kelly et al. 1990; birds, Wykes 1983; ants, Majer 1980). Havel (1975) collected the most geographically exten¬ sive point-based data set, but his sampling concentrated on uplands and was confined to a selection of perennial plants. Granite outcrop, alluvial flat and heathland com¬ munities were under-represented, while annuals, peren¬ nials such as orchids, vertebrates, invertebrates, and the rarer components of the flora were all ignored (Havel 1989, p. 384). Conclusion We concur with previous reviews - the reserve sys¬ tem has already been assessed in terms of its compre¬ hensiveness over most of the Jarrah Forest Bioregion. As yet there is no quantitative basis for assessing the re¬ serve system's representativeness in terms of the jarrah forest's biodiversity at a level equivalent to that now possible in remote areas such as the Nullarbor (McKenzie et al. 1989), Western Australia's Eastern Gold¬ fields (e.g. How et al. 1988; Burbidge et al. 1995) and Kimberley rainforests (McKenzie & Belbin 1991), and South Australia's Gawler Ranges (Robinson et al. 1985). A quantitative biogeographical survey would provide 245 Journal of the Royal Society of Western Australia, 79(4), December 1996 a model of the area's biological patterns from an ecosys¬ tem perspective. Such a survey would be quadrat-based and cover the major geological and geographical gradi¬ ents in the area. At each quadrat, detailed sampling would need to cover different components of the biota that could be reasonably expected to be responding to different scalars. Typically such a survey would include vascular flora, small mammals, birds, reptiles and am¬ phibians and selected invertebrate and cryptogram groups for which a good taxonomic basis exists. Such data allow the development of regional models of gradients in the species composition of communities (quantified regional contexts) and allow decisions con¬ cerning land-use and management to be more explicit (McKenzie 1988). Such an approach would allow a more representative regional conservation network to be de¬ signed, one that could be expected to better represent the genetic resources of the widespread elements of the biota. It would not achieve adequate representation of the rare or restricted elements of the biota, for which the program of specific studies will need to continue in par¬ allel. Regional priorities have to be considered in proposals for major biogeographical surveys. There are 25 other biogeographical regions in Western Australia, and large- scale biogeographic surveys are expensive in terms of time, and in human and financial resources (Burbidge 1991) . In the last 15 years it has not been practical to conduct more than one such survey at a time. Each sur¬ vey requires up to 15 people during the fieldwork, and five people during the analysis and write-up phase for a period of 3 to 5 years. Recent analysis of the reservation status within the 26 bioregions recognised in Western Australia (Thackway & Creswell 1995) showed that 17 have a smaller proportion of their area reserved than the Jarrah Forest Bioregion, and 11 of these have a more biased reserve system (this means that entire sub- regions, or many of the extensive ecosystems that characterise a region, are missed). All of these regions can be considered much less resilient than the Jarrah Forest Bioregion. From a biodiversity perspective at a state-wide scale, the Jarrah Forest Bioregion does not have a high priority for ecological survey. Nevertheless, the appraisals re¬ viewed in this paper have identified gaps in the existing reserve system, even at the level of its comprehensive¬ ness. These gaps relate to certain vegetation complexes on the Darling Scarp, the Blackwood Plateau, the Collie coalfields, and on agriculturally desirable soils. They should be addressed as a priority during the implemen¬ tation of the National Forest Policy Statement (NFPS 1992) . Acknowledgements : We thank I Abbott, A Burbidge, A Hopkins, N Marchant and G Stoneman for their helpful comments on the draft. This review was updated from notes compiled in 1991 by N L McKenzie, S D Hopper and G Wardell -Johnson. References Abbott I & Christensen P 1994 Applications of ecological and evolutionary principles to forest management in Western Australia. Australian Forestry 57:109-122. Abbott 1 & Christensen P 1996 Living with logging: the survival of species. Landscope 1 1 (2):2 1-27. Anon. 1985. Worsley Alumina Project flora and fauna studies, phase 2. Worsley Alumina, Perth. Anon. 1987 Next major public water supply for Perth (post-1992). ERMP Stage 1: Evaluation of alternatives, supporting docu¬ ment. Flora and vegetation of four alternative water resource development sites in the northern jarrah forest, Western Aus¬ tralia. Water Authority of Western Australia, Perth. Anon. 1987a Southern Forest Region regional management plan 1987-1997. Management Plan 11. Department of Conservation and Land Management, Perth. Anon. 1987b Northern Forest Region regional management plan 1987-1997. Management Plan 9. Department of Conservation and Land Management, Perth. Anon. 1993. Red Book Status Report orf the implementation of conservation reserves for Western Australia as recommended by the Environmental Protection Authority (1976-1984). Envi¬ ronmental Protection Authority, Perth. Anon. 1994 Forest Management Plan 1994-2003. Department of Conservation and Land Management, Perth. Austin M P & Heyligers P C 1989 Vegetation survey design for conservation: gradsec sampling of forests in north-eastem New South Wales. Biological Conservation 50:13-32. Beard J S 1982 Vegetation survey of Western Australia. Swan. 1:1 000 000 vegetation series. University of Western Australia Press, Perth. Beard J S & Sprenger B S 1984 Geographical data on the vegeta¬ tion survey of Western Australia, with map of aeolian land- forms 1:3,000,000. Vegetation Survey of Western Australia Oc¬ casional Paper No. 2. Vegmap, Applecross. Burbidge A A 1991 Cost constraints of surveys for nature conser¬ vation. In: Nature Conservation: Cost Effective Biological Sur¬ veys and Data Analysis (eds C R Margules & M P Austin). CSIRO Division of Wildlife and Ecology', Canberra, 3-6. Burbidge A A, Hall N J, Keighery G J & McKenzie N L 1995 The biological survey of the Eastern Goldfields of Western Aus¬ tralia. Part 12: Barlee - Menzies study area. Records of the Western Australian Museum Supplement 49:169-312. Burbidge A A & McKenzie N L 1989 Patterns in the modern decline of Western Australia's vertebrate fauna: causes and conservation implications. Biological Conservation 50:143- 198. Churchward H M & McArthur W M 1980 Landforms and soils of the Darling System, Western Australia. In: Atlas of Natural Resources: Darling System, Western Australia. Western Aus¬ tralian Department of Conservation and Environment, Perth, 25-36. Churchward H M, McArthur W M, Sewell P L & Bartle G A 1988 Landforms and soils of the south coast and hinterland, West¬ ern Australia. CSIRO Division of Water Resources, Divisional Report 88/1, Canberra. Churchward H M 1992 Soils and landforms of the Manjimup area, Western Australia. Land Resources Series 10. Western Australian Department of Agriculture, Perth. Dell B, Havel J J & Malajczuk N 1989 The jarrah forest. A com¬ plex Mediterranean ecosystem. Klewer Academic Publishers, Dordrecht. Driscoll D, Wardell-Johnston C. & Roberts J D 1995 Genetic struc¬ turing and distribution patterns in rare southwestern Austra¬ lian frogs: implications for translocation programmes. In: Re- introduction Biology of Australian and New Zealand Fauna (ed M Serena). Surrey Beatty & Sons, New South Wales, 85- 90. Faunt K 1992 Formation, frequency and longevity of hollows in jarrah. Interim Report to Department of Conservation and Land Management, Perth. Friend G R 1993 Impact of fire on small vertebrates in mallee woodlands and heathlands of temperate Australia: a review. Biological Conservation 65:99-144. 246 Journal of the Royal Society of Western Australia, 79(4), December 1996 Friend G R, Morris K D & Thompson C 1994 Bountiful Batalling. Landscope 9(4):16-22. Friend ] A & Thomas N D 1995 Reintroduction and the numbat recovery programme. In: Reintroduction Biology of Austra¬ lian and New Zealand Fauna (ed M Serena). Surrey Beatty & Sons, New South Wales, 189-198. George R J, McFarlane D J & Speed R J 1995 The consequences of a changing hydrologic environment for native vegetation in southwestern Australia. In: Nature Conservation 4: The Role of Networks (eds D A Saunders, ] L Craig & E M Mattiske). Surrey Beatty & Sons, New South Wales, 9-22. Gibson N, Keighery B J, Keighery G J, Burbidge A FI & Lyons M N 1994 A floristic survey of the southern Swan Coastal Plain. Report to the Australian Heritage Commission prepared by the Western Australian Department of Conservation and Land Management and the Conservation Council of Western Australia. Gillison A N & Brewer K R W 1985 The use of gradient directed transects or gradsecs in natural resource surveys. Journal of Environmental Management 20:102-127. Havel J J 1975 Site vegetation mapping in the northern jarrah forest. 1. Definition of site-vegetation types. Forests Depart¬ ment of Western Australia Bulletin 86. Havel ] J 1989 Conservation in the northern jarrah forest. In: The jarrah forest. A complex Mediterranean ecosystem, (eds B Dell, J J Havel, & N Malajczuk). Klewer Academic Publishers, Dordrecht, 379-399. Heddle E M, Havel J J & Loneragan O W 1980 Focus on northern jarrah forest conservation and recreation areas. Forest Focus 22:3-27. Heddle E M, Loneragan O W & Havel J J 1980 Vegetation com¬ plexes of the Darling system. Western Australia. In: Atlas of natural resources: Darling system. Western Australia. Depart¬ ment of Conservation and Environment, Perth, 33-74. Horwitz P 1994 Patterns of endemism in the freshwater fauna of the far southern peatlands and shrublands of south-western Australia. Report to the Australian Heritage Commission and the Heritage Council of Western Australia. How R A, Newbey K R, Dell J, Muir B G & Hnatiuk R J 1988 The biological survey of the Eastern Goldfields of Western Aus¬ tralia. Part 4: Lake Johnston - Hyden study area. Records of the Western Australian Museum Supplement 30: 1-233. Keighery G ] 1990 Floristics of Yelverton State Forest. Depart¬ ment of Conservation and Land Management, Perth. Kelly A E, Coates D J, Herford I, Hopper S D, O'Donoghue M & Robson L 1990 Declared rare flora and other plants in need of special protection in the northern forest region. Western Aus¬ tralian Wildlife Management Program 5. Department Conser¬ vation and Land Management, Perth. Kinnear J E 1993 Masterly marauders, the cat and the fox. Landscope 8(2):20-27. Kitchener D J, Chapman A & Barron G 1978 Mammals of the northern Swan Coastal Plain. In: Faunal Studies of the North¬ ern Swan Coastal Plain. A Consideration of Past and Future Changes. Western Australian Museum, Perth, 54-92. Majer J D 1980 A preliminary ecological survey of the Wagerup ant fauna. Environmental Research Bulletin 3. ALCOA Aus¬ tralia, Perth. Majer J D 1990 Rehabilitation of disturbed land: long-term pros¬ pects for the recolonization of fauna. Proceedings of the Eco¬ logical Society of Australia 16:509-519. Margules C R & Austin M P 1994 Biological models for monitor¬ ing species decline. Philosophical Transactions of the Royal Society of London B 344:69-75. Margules C R & Nicholls A O 1994 Where should nature re¬ serves be located? In: Conservation Biology in Australia and Oceania (eds C Moritz & ] Kikkawa). Surrey Beatty & Sons, New South Wales, 339-346 Margules C R, Pressey R L & Nicholls A O 1991 Selecting nature reserves. In: Nature Conservation: Cost Effective Biological Surveys and Data Analysis (eds C R Margules & M P Austin). CSIRO Division of Wildlife and Ecology, Canberra, 90-97. McKenzie N L 1984 Survey methods for nature conservation by the Western Australian Department of Fisheries and Wildlife - a current view. In: Survey Methods for Nature Conservation (eds K Myers, C R Margules, & 1 Musto). CSIRO Division of Water and Land Resources, Canberra, 88-1 17. McKenzie N L 1988 Biogeography and its use for setting priori¬ ties for management. In: Ecological Theory and Biological Management of Ecosystems (eds D A Saunders & A A Burbidge). Occasional Paper 1/88. Department of Conserva¬ tion and Land Management, Perth, 39-44. McKenzie N L, Belbin L, Margules C R & Keighery G J 1989 Selecting representative reserve systems in remote areas: a case-study in the Nullarbor region, Australia. Biological Con¬ servation 50:239-261. McKenzie N L & Belbin L 1991. Kimberley rainforest communi¬ ties: Reserve recommendations and management consider¬ ations. In: Kimberley Rainforests of Australia (eds N L McKenzie, R B Johnston & P G Kendrick). Surrey Beatty & Sons, New South Wales, 453-468. McKenzie N L, Robinson A C & Belbin L 1991 Biogeographic survey of the Nullarbor District, Australia. In: Nature Conser¬ vation: Cost Effective Biological Surveys and Data Analysis (eds C R Margules & M P Austin). CSIRO Division of Wildlife and Ecology, Canberra, 109-126. McKenzie N L, Beeston C. R, Bowen P J, Burbidge A A, Burbidge A H, Gibson N, Keighery C, J & Hopper S D 1992 Environ¬ mental regionalisation for conservation in Western Australia. In: Environmental Regionalisation - Establishing a Systematic Basis for National and Regional Conservation Assessment and Planning (ed R Thackway). Australian National Parks and Wildlife Sendee, Canberra, 20-25. McKenzie NL & Sattler P S 1994 Biological surveys and reserve design. In: Conservation Biology in Australia and Oceania (eds C Moritz & J Kikkawa). Surrey Beatty & Sons, New South Wales, 379-384. Morris K D 1993 Return of the Chuditch. Landscope 8(2):10-15. NFPS 1992 National Forestry Policy Statement. A New Focus for Australia's Forests. Commonwealth of Australia, Canberra. Nichols O G & Bamford M J 1985 Reptile and frog utilisation of rehabilitated bauxite minesites and dieback affected sites in Western Australia's jarrah ( Eucalyptus marginata) forest. Bio¬ logical Conservation 34:227-249. Nichols O G & Watkins D 1984 Bird utilisation of rehabilitated bauxite minesites in Western Australia. Biological Conserva¬ tion 30:109-131. Nix H A & Gillison A N 1985 Towards an operational frame¬ work for habitat and wildlife management. In: Wildlife Man¬ agement in the Forestry Controlled Lands in the Tropics and Southern Hemisphere (ed J Kikkawa). University of Queensland, Brisbane, 39-55. Orell P & Morris K 1994 Chuditch recovery plan. Western Aus¬ tralian Wildlife Management Program 13. Department of Conservation and Land Management, Perth. Pigott J P & Gray C M 1993 Watsonia workshop. Plant Protection Quarterly 8(3):75-105. Pressey R L & Bed ward M 1991 Mapping the environment at different scales: benefits and costs for nature conservation. In: Nature Conservation: Cost Effective Biological Surveys and Data Analysis (eds C R Margules & M P Austin). CSIRO Divi¬ sion of Wildlife and Ecology, Canberra, 7-13. Pressey R L & Nicholls A O 1989 Application of a numerical algorithm to the selection of reserves in semi-arid New South Wales. Biological Conservation 50:263-278. Pressey R L & Nicholls A O 1991 Reserve selection in the West¬ ern Division of New South Wales: development of a new pro¬ cedure based on land system mapping. In: Nature Conserva¬ tion: Cost Effective Biological Surveys and Data Analysis (eds C R Margules & M P Austin). CSIRO Division of Wildlife and Ecology, Canberra, 98-105. 247 Journal of the Royal Society of Western Australia, 79(4), December 1996 Recher H F & Serventy D L 1991 Long term changes in the rela¬ tive abundance of birds in King's Park, Perth, Western Aus¬ tralia. Conservation Biology 5:90-102. Rhind S G 1996 Habitat tree requirements and the effects of re¬ moval during logging on the marsupial brush-tailed phascogale (Phascogale tapoatafa tapoatafa) in Western Austra¬ lia. Western Australian Naturalist 21: 1-22. Roberts J D, Horwitz P, Wardell-Johnson G, Maxson L R & Mahony M J 1997 Taxonomy, relationships and conservation of a new genus and species of myobatrachid frog from the high rainfall region of south-western Australia. Copeia 1997: 373-381. Robinson A C, Casperson K D, Canty P D & Macdonald C A 1985 A biological survey of the Gawler Ranges South Australia in October 1985. Government Printer, Adelaide. Shearer B L & Tippett J T 1989 Jarrah Dieback: the dynamics and management of Phytophthora cinnamomi in the jarrah ( Eucalyp¬ tus marginata ) forest of South-western Australia. Research Bul¬ letin 3. Department of Conservation and Land Management, Perth. Start A N, Burbidge, A A & Armstrong D 1995 Woylie recovery plan. Western Australian Wildlife Management Program 16. Department of Conservation and Land Management, Perth. Tille P & Lantzke N 1990 Busselton - Margaret River - Augusta land capability study; methodology and results. Vol 2. West¬ ern Australia Department of Agriculture, Perth. Thackway R & Cresswell I D 1995 An interim biogeographic regionalisation for Australia. A framework for setting priori¬ ties in the National Reserve System Cooperative Program. Version 4.0. Australian Nature Conservation Agency, Canberra. Underwood R J, Walker A W, Schuster C, Sands A & Lush A 1991 The Karri Forest, Western Australia. In: Forest Manage¬ ment in Australia (eds F H McKinnell, E R Hopkins & J E D Fox). Surrey Beatty & Sons, New South Wales, 22-36. Wardell-Johnson G, Hearn R, Williams M & Annels A 1996 A floristic survey of the Tingle mosaic. Report to the Heritage Council of Western Australia, Perth. Wardell-Johnson G & Horwitz P 1996 Conserving biodiversity and the recognition of heterogeneity in ancient landscapes: a case study from south-western Australia. Forest Ecology and Management 85:219-238. Wardell-Johnson G & Nichols O 1991 Forest wildlife and habitat management in south-western Australia: knowledge, research and direction. In: Conservation of Australia's Forest Fauna (ed D Lunney). Royal Zoological Society of NSW, Sydney, 161-192. Wardell-Johnson G & Roberts J D 1993 Biogeographic barriers in a subdued landscape: the distribution of Geocrinia rosea (Anura: Myobatrachidae) complex in south-western Austra¬ lia. Journal of Biogeography 20:95-108. Wardell-Johnson G, Williams J, Hill K & Cummings R in press Historical biogeography and current distribution patterns of eucalypts. In: Eucalvpt Ecology: Individuals to Ecosystems (eds J Williams & J Woinarski). Cambridge University Press, Cambridge. Wilde S A & Low G H 1978 Perth, Western Australia, 1:250,000 geological series map sheet S150-14 and explanatory notes. Geological Survey of Western Australia, Department of Mines, Perth. Wilde S A & Low G H 1980 Pinjarra, Western Australia, 1:250,000 geological series map sheet SI50-2 and explanatory notes. Geological Survey of Western Australia, Department of Mines, Perth. Wilde S A & Walker I W 1982 Collie, Western Australia, 1:250,000 geological series map sheet SI50-6 and explanatory notes. Geological Survey of Western Australia, Department of Mines, Perth. Wilde S A & Walker I W 1984 Pemberton - Irwin Inlet, Western Australia, 1:250,000 geological series map sheet SI50-14 and explanatory notes. Geological Survey of Western Australia, Department of Mines, Perth. Woinarski J C Z & Norton T W 1993 Towards a national system of forest reserves: A discussion paper. Department of Envi¬ ronment, Sport and Territories, Canberra. Wykes B J 1983 The jarrah avifauna and its re-establishment after bauxite mining. Western Australian Institute of Technology Bulletin 11, Perth. 248 Journal of the Royal Society of Western Australia, 79:249-276, 1996 A floristic survey of the Tingle Mosaic, south-western Australia: applications in land use planning and management G Wardell-Johnson1 & M Williams- 1 Science and Information Division, Department of Conservation and Land Management, PO Box 51 Wanneroo, WA 6065. present address. Department of Biology, University of Namibia, Private Bag 13301, Windhoek, Namibia. 2 Science and Informa¬ tion Division, Department of Conservation and Land Management, Research Centre, Hayman Road, Como WA 6152 Abstract A floristic survey of the Tingle Mosaic, an area of 3 700 km2 which includes the wettest, least seasonal and southern-most part of Western Australia recorded a total of 857 vascular plant taxa in 441 quadrats (20 m x 20 m). These included 825 indigenous and 32 introduced taxa. Important families included the Papilionaceae (74 species), Proteaceae (73), Myrtaceae (64) and Orchidaceae (63). Cluster analysis and ordination techniques defined five floristic communities supergroups, 12 community groups and 44 community types. The Open-forest, Tall open-forest and Shrubland/ woodland Communities Supergroups included most of the quadrats (356 of 441), and also occu¬ pied the largest areas within the region. There was high alpha diversity for the Woodland and Open- forest communities supergroups, while there was low alpha and gamma diversity for the Tall open-forest Communities Supergroup. Considerable variation in vegetation structure, and high gamma diversity was found for the three non-forest communities supergroups. An ex¬ panded program of survey would be required to target the exceptional variety of sites in the Swamp and outcrop Communities Supergroup. The Tingle Mosaic had high levels of local ende¬ mism, many taxa (both wet and dry country taxa) which have range limits in the area, and several relictual high rainfall taxa whose distributions are centred in the area. A high proportion of the region lies within the conservation reserve network. Nevertheless the conservation signifi¬ cance and complexity of the fine-scale biotic pattern in the area urge increased attention in management and policy for the conservation of biodiversity. Methods to integrate site-based work, to define complexes of community types, and of the mapping of these floristic assemblages are presented. These applications would be invaluable in management for the conservation of biodiversity in the region. Introduction The south-west of Western Australia includes an ex¬ traordinary diversity of vascular plants in a generally subdued landscape (Hopper 1979; Hopper et al. 1992). This diversity has been especially noted for the inland transitional rainfall zone (TRZ sensu Hopper 1992) which is dominated by speciose genera of woody peren¬ nials in families such as the Myrtaceae, Proteaceae, Fabaceae and Epacridaceae (Hopper 1979; 1992) but not the high rainfall zone (HRZ) closer to the coast. Never¬ theless wetland monocotyledonous taxa, including gen¬ era of Cyperaceae, Xyridaceae, J uncaginaceae, Restionaceae and Orchidaceae, are species-rich in the re¬ gion. For example at least 1947 taxa are known from the Warren Botanical Subdistrict alone, despite limited sur¬ vey (Hopper et al. 1992). The HRZ is also notable for its high diversity of eucalypts compared with similar areas elsewhere in the Darling Botanical District (Christensen 1980; Smith et al. 1991, Wardell-Johnson & Smith 1991; Wardell-Johnson & Coates 1996, Wardell-Johnson et al. in press). At least Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 four species of large forest eucalypts, Eucalyptus brevistylis (Rates tingle), E. jacksonii (red tingle), E. guilfoylei (yellow tingle), and E. ficifolia (red-flowering gum) are locally endemic to the south-west between Walpole and Denmark. Each occurs in several allopatric populations in an area of high landscape and vegetation structural diversity. Because it was the tingles that first drew attention to the conservation significance of this landscape mosaic (Fernie & Fernie 1989), the survey area is described as the Tingle Mosaic. The Tingle Mosaic is also notable for its scenic diver¬ sity and has been the subject of intense public interest (Smith et al. 1991; Wardell-Johnson & Smith 1991; Wardell-Johnson & Horwitz 1996). It occurs in close proximity to the towns of Walpole, Denmark and Al¬ bany (Smith et al. 1991) but includes areas that are gen¬ erally remote and have not been explored botanically. The vascular flora has been chosen for study because of its richness (Hopper et al. 1992) and because of its pro¬ pensity to describe landscape pattern (Havel 1981; Wardell-Johnson et al. 1989). The identification and clas¬ sification of plant community types is a useful first step in land-use planning (Havel 1981; Wardell-Johnson et al. 1989). This allows the identification of rare or vulnerable communities at a regional level and allows a manage¬ ment context to be provided. Historically, vegetation classification in Australia has been based on a structural or physiognomic basis (Diels 1906; Speck 1952; Webb et 249 Journal of the Royal Society of Western Australia, 79(4), December 1996 al. 1976; Beard 1981). More recent work has utilised flo- ristic attributes ( e.g . Havel 1975a,b; Webb et al. 1984; Cresswell & Bridgewater 1985; Foran et al 1986). How¬ ever, the final choice of attributes used to classify veg¬ etation remains a function of the aims of the research project at hand (Anderson 1981). Here, we describe the variation in the floristic compo¬ sition of the area within the range limits of four locally endemic forest eucalypts (Tingle Mosaic) and identify sites with similar species composition (community types; terminology after Whittaker 1973). Methods Study area The study area in south-western Australia lies be¬ tween latitudes 34° 45' and 35° 10' and longitudes 116° 30' and 117° 45'. This area of approximately 3 700 km2 (Table 1, Fig 1) within the Warren and Menzies Botani¬ cal Subdistricts of Beard (1980) includes the wettest and least seasonal part of Western Australia, although isohyets decline rapidly eastwards and from the coast. Peter Title (Agriculture WA, pers. comm., 1995) de¬ fined 98 land systems within 16 zones for The Darling Botanical District of Beard (1980), based on geology and recurring landform patterns. This area includes almost all of the jarrah and karri forests and woodlands of south-western Australia. The Tingle Mosaic lies within three zones (the Warren-Denmark southland, the Stirling coastal, and the Stirling Sandplain) and nine Land sys¬ tems. The area includes two geomorphic provinces within the South-west Land Division (the Avon and the Stirling). Climate The region has a mediterranean climate (Gentilli 1971) with dry, generally temperate summers and cool, wet winters. The high rainfall parts of the Tingle Mosaic re¬ ceive more reliable summer rain than elsewhere in the south-west. There is a considerable range in temperature and rainfall with gradients from both south-to-north and west-to-east. Thus in the south-western parts, annual rainfall exceeds 1 400 mm (Fig 1), w'hereas in the north¬ east it is about 750 mm. Geology, physiography and soils The Tingle Mosaic includes the southern fringe of the Great Plateau of Western Australia (Jutson 1914). It is composed of Precambrian granite rocks, partially over- lain by various consolidated and unconsolidated sedi¬ ments. This land surface has been subjected to a long and complex history of weathering and denudation which is expressed as variations in topography, soils and hydrology. These factors, together with the nature of the rock Figure 1. Study area showing the distribution of permanently located floristic quadrats, isohyets and major towns in south-western Australia. 250 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 1 Tenure context of survey area. Reserves are National Parks, Nature Reserves and Conservation Parks. State Forest includes Executive Director lands, Timber Reserves and Other Reserves. Location Reserves km2 (%) State Forest km2 (%) Other km2 (%) Private Land km2 (%) Total Area km2 Darling Botanical District 8 070 (11.3) 16 100 (22.6) 1 100 (1.5) 46 060 (64.6) 71 300 Warren Botanical District 2 870 (27.8) 3 290 (32) 430 (0.9) 3 709 (36) 10 300 Tingle Mosaic 909 (24.6) 686 (18.5) 387 (10.4) 1 717 (46.5) 3 700 types, formed the basis for the recognition of the land- form/soils units defined and described by Churchward et al. (1988). These authors provided a detailed set of five 1:100 000 scale landform and soils maps covering an area between Windy Harbour eastward to Cheyne (Hassell) Beach, 80 km east of Albany, and extending inland to latitude 34° 31' S (as far north as Rocky Gully). Thirty-five units were mapped, based firstly on general geological features (i.e. units developed on granite or unconsolidated sediments, on silts tones and sandstones, on coastal aeolian and fluvial sediments and on drain¬ age lines), and then on landform (plateau elements, hills and ridges, swampy terrain, dune systems, and major and minor valleys). They also provided further subdivi¬ sion into individual units based on soils, local relief, slope and drainage patterns. The maps of Churchward et al. (1988) provided a sound context for stratification in this survey (and also for broad-scale ecosystem manage¬ ment in the area). Most of the Tingle Mosaic has soils which are devel¬ oped on components of laterite profiles either exposed by erosion or as colluvial waste released by this process. Many of these soils have ferrugenous gravels in the sur¬ face horizons. In the unconsolidated sandy sediments, soil morphology is influenced mainly by drainage sta¬ tus, while in the coastal sand dunes, soil variation often relates to the age of parent material. The headwaters of the Frankland River, a large river which rises well to the north of the Tingle Mosaic, are in broad valleys with salt lakes. Hence, there is a natural contribution of salts to the surface water. This is at a maximum following heavy winter rains. The headwa¬ ters of the Deep, Denmark, Kent and Hay Rivers also rise to the north of the Tingle Mosaic, and occur in areas with high soluble salts in the subsoil. Surface incidence of salinity is most evident in the north-east part of the Tingle Mosaic, and is expressed both as seepage on slopes and as saline valley floors. Salinity is not appar¬ ent on the plateau developed on marine sediments even though rainfall is low and the substrata known to be saline (Teakle 1953). The general level of the Great Plateau gradually falls toward sea level in the area and forms part of the Ravensthorpe Ramp (Cope 1975). In the east, the land¬ scape consists of a plateau developed on Pallinup Silt- stone. The plateau surface represented by gently undu¬ lating plain slopes from about 180 m above sea level to about 40 m near the coast. In the remainder of the area, much of the terrain is developed on granitic rocks al¬ though there is a variable incidence of unconsolidated deposits as well as westward extensions of the Eocene sediments. This is at an elevation of from 260 to 180 m above sea level and is part of the Great Plateau under¬ lain by deeply weathered granite. Some broad sandy, often swampy tracts are included. To the west, extensive tracts such as the Pingerup Plains grade in elevation from 70 m inland, to less than 20 m where they abutt the narrow zone of coastal dunes. Granite is sometimes exposed as domes and pinnacles emergent above the sandy tracts (e.g, Mt Pingerup). The terrain is dominated by a pattern of ridges of granitic rocks alternating with broad swampy corridors, having a west-north-west orientation. The crests of many of the ridges are in excess of 100 m above the corridors. On the coastal fringe, both Precambrian and Tertiary rocks are overlain by Tamala Limestone of Pleistocene age (Logan 1968) and/or unconsolidated Holocene aeolian sands. The broad ridges of Tamala Limestone rise to about 100 m and often act as barriers behind which estuaries, such as Wilson Inlet, have developed. A system of Holocene parabolic dunes extend from the coast in a general east-north-east direction, overlying the broad ridges of limestone and of granite as well as allu¬ vial and estuarine deposits. Much of the coastline is characterised by a succession of arcuate bays with gra¬ nitic headlands linked by the barriers of Tamala Lime¬ stone in which steep cliffs have usually been cut. Most of these bays face south-west. The coastal plain in the immediate hinterland of the dunes is mantled by uncon¬ solidated alluvium and/or aeolian sands. Vegetation Each of the 35 landform units defined by Churchward et al. (1988) were described in terms of their physiogra¬ phy, geology, soil morphology and associated native vegetation. The latter was described structurally, and dominants were listed for each major stratum. This pro¬ vided a means of determining quadrat locations to en¬ sure a regional coverage of a complex area. The natural vegetation of the area has been mapped (scale 1:250 000) by Smith (1972) and Beard (1972-80). Hopper et al. (1992) provided a regional perspective for the flora of the Warren Botanical Subdistrict. They listed 1947 taxa including 1628 native and 319 naturalised introduced taxa for the area (roughly equiva¬ lent in size to the Perth Region; 8323 knv extending over 300 km from Yallingup on the Leeuwin-Naturaliste Ridge to Albany on the south coast). Many site-based floristic studies have been carried out, in or nearby to the Tingle Mosaic. For example, 251 Journal of the Royal Society of Western Australia, 79(4), December 1996 Strelein (1988) presented an ordination using over 400 sample sites and 100 indicator species in the southern jarrah forest. He defined seventeen site types from this work using the methods of Havel (1968, 1975a,b) and discussed the regeneration, dieback susceptibility and productivity of each. Inions et ai (1990) derived a floris- tic classification of regenerating karri forest in the Nornalup System of the Warren Subdistrict. They used 204 permanent inventory plots (Campbell et al. 1985) and 105 species were sampled. Thirteen community types were defined by cluster analysis, ordination and discriminant analysis of the 312 m2 quadrats. Wardell- Johnson et ai (1989) developed a floristic classification of the Walpole-Nornalup National Park based on 219 quadrats and 233 species. Twelve community types were derived with clustering and ordination techniques. Sev¬ eral smaller site-based studies have also been carried out in the area (e.g. Hopkins & Griffin 1984). These have been reviewed by Hopper et at. (1992). Other opportu¬ nistic flora surveys have been carried out in the area over many years and a preliminary list of flora for the Warren Botanical Subdistrict is now available. Thus, over 2200 indigenous vascular plant taxa, many yet to be named, are known from this subdistrict alone (Hop¬ per et al. 1992; N Gibson, CALM pers. comm , 1995). Sampling sites This study was based on intensive sampling of the different vegetation types in representative areas. Areas were chosen on the basis of the landform/soils classifi¬ cation of Churchward et al. (1988), although more inten¬ sive survey in the Walpole-Nornalup National Park was completed prior to the availability of this work (but see Wardell-Johnson et al. 1989). Data from 144 sites of quadrat size 10 m radius (312 m2) have been included from this work (see also Wardell-Johnson et al 1989) Hence the survey effort is considerably greater for the WNNP than for the remainder of the Tingle Mosaic. The park does however include a major proportion of the populations of all three tingles as well as E. ficifolia (Smith et al. 1991, Wardell-Johnson & Smith 1991). Hence the concentration of quadrats in this area reflects the concentration of the rare eucalypts in the area. The locations of sampling sites were selected to give as wide a range as possible of vegetation types from throughout the study area. All quadrats were located in undisturbed indigenous vegetation, with few weeds oc¬ curring in the area and no recent history of high inten¬ sity fire. Sampling preference was given to areas which were in existing conservation reserves rather than private property or road reserv es. Detailed studies have recognised that an appropriate sample area is about 400 m2 in forested areas (Burbidge & Boscacci 1987; D Keith, Forests Commission of NSW, pers. comm. 1994). This allows a representative floristic list for the site, and minimises the influences of indi¬ vidual large trees or logs within a quadrat. Quadrats larger than this risk encountering ecotones in the diverse vegetation of the HRZ of south-western Australia. All quadrats in this study were 20 m x 20 m, except those from the Walpole-Nornalup National Park survey men¬ tioned above. All quadrats were established and perma¬ nently marked in the field by metal star pickets at centre points, and droppers at corners. The sites were checked at least twice, including at least once in spring. Species nomenclature follows Green (1985). Analytical techniques All sites (441) and all taxa (857) were analysed for plants presence/ absence, which provides most of the in¬ formation by ordination and classification of site-based data (Anderberg 1973). A matrix of pairwise associa¬ tions between sites was calculated using the Czekanowski (1913) metric. UPGMA (Sneath & Sokal 1973) was used to derive clusters from the dataset; al¬ though this is sometimes prone to minor misclassification, it has the advantage of taking more than one species into account at any fusion. The cluster¬ ing-intensity coefficient beta ((3) was -0.10; under such conditions the clustering strategy is space-dilating and resists the formation of a single large group by forcing the formation of even-sized groups (Booth 1978). The use of clustering techniques assumes that a popu¬ lation is discontinuous; the validity of this assumption depends on the species' response to environmental gra¬ dients and the nature of the gradients themselves (Aus¬ tin & Cunningham 1981). In reality, vegetation is likely to vary from apparently continuous to apparently dis¬ continuous with the nature of boundaries varying in width and level of diffusion throughout, and by how they are defined. The location of sample sites can also have an influence over the degree to which the vegeta¬ tion of a region is considered continuous or discontinu¬ ous. The acceptability of imposed groups was examined by ordinating the sites using semi-strong-hybrid multi¬ dimensional scaling (PATN; Belbin 1993), and examin¬ ing the position of group members in component space. As analysis should not be performed across major data discontinuities (Green 1980), further cluster analysis was performed on each of the five major discontinuity's determined through initial analysis. Results Sites and species A total of 857 vascular plant taxa were recorded in the 441 quadrats of the Tingle Mosaic. Important plant families included the Papilionaceae (74 species), Proteaceae (73), Myrtaceae (64) and Orchidaceae (63). Relative to the number of taxa represented in the War¬ ren Botanical Subdistrict, the Orchidaceae (63), Cyperaceae (34), Restionaceae (34), Poaceae (24) and Asteraceae (19) were relatively poorly represented in quadrats (see Hopper et al. 1992). The largest representa¬ tion of genera included Acacia (23 species), Stylidium (36), Leucopogon (20) and Eucalyptus (22). Caladenia (12) and Drosera (12) were relatively poorly represented in quadrats in comparison with the Warren flora as a whole. A list of the taxa and their constancy (proportion of quadrats in which the species is present) in commu¬ nity groups is presented in Appendix 1. Several name changes occurred during the course of the study, and several taxa were found to have been misidentified once multivariate analysis was complete. These were few, thus providing little likelihood of influencing the overall classification. They are also shown in Appendix 1. 252 Journal of the Royal Society of Western Australia, 79(4), December 1996 Jarrah/sheoak woodland-upland sites Jarrah/sheoak woodland-loamy sand Shrubland-humus podzols/granite Tall shrubland-peaty podzols Shrubland/woodlancf-shallow gritty soils Open shrubland-humus podzols Jarrah/marri open forest and woodland-sands Forest /shrubland ecotone-podzols Forest/shrubland ecotone-yellow duplex soils Shrubland-interdune plains Banksia woodland-interdune plains Coastal herbland-interdune plains Tall shrubland-sumps Shrubland-older dunes Shrubland-recent dunes Shrubland-limestone substrate Shrubland /heathland-coastal headlands Shrubland-peaty interdune plains Shrubland-peat swamp Closed shrubland-peaty sands Blackbutt woodland-clay loams Melaleuca woodland-seasonally inundated Melaleuca woodland-estuarine Open sedgeland-clay Melaleuca woodland-sumps Granite outcrop-coastal Granite outcrop-sands Granite outcrop-loam Shrubland /forest-fertile valley floors Yate woodland-clay Granite outcrop-shallow soils Granite outcrop-shallow soils Granite outcrop-shallow soils Blackbutt wood land-clay/ loam Wandoo woodland-valley slopes Open woodland-clay/granite outcrop Tall open-forest-brown gravels Open Jarrah/Marri forest-brown gravels Open Jarrah/Marri forest-laterite Open Jarrah/Marri forest-gritty/outcrop Low open Jarrah/Marri forest-laterite/outcrop Tall open Karri/Tingle forest-brown gravels Tall open Karri forest Tall open Karri forest 0.7640 1.1032 1.4424 1.7816 2.1208 2.4600 Figure 2. Dendrogram classification of 441 sites, based on 857 vascular plant taxa showing communities supergroups (A-E) and community types. The predominant vegetation and soil features of the 44 community types are shown with numbers of quadrats in brackets. Reprinted from Wardell-Johnson & Horwitz (1996) Forest Ecology and Management, 85(1-3) , Conserving biodiversity and the recognition of heterogeneity in ancient landscapes: a case study from south-western Australia, 219-238, 1996 with kind permission of Elsevier Science - NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands. Cluster analysis and ordination defined five floristic communities supergroups, 12 community groups and 44 community types (Figs 2-3; Table 2). As the five communities supergroups represent a clear discontinuity in both cluster analysis and ordination, it is likely that this will mask differences amongst community types (Green 1980). Thus the five communities supergroups were each analysed separately for higher resolution within the supergroups and to define community sub- types. A clear separation of sites at the 12 group level de¬ fined groups of similar sites according to broad topogra¬ phy, drainage and soils characteristics. These 12 groups are referred to as community groups and the constancy of taxa within these groups are presented as a summary of the floristics of the area (Appendix 1). All but the tall open-forest communities supergroup comprise more than one community group. Shru bland/woodland Communities Supergroup. (A: 131 quadrats and 464 taxa) included four community groups (forest/sandplain ecotone sites, woodland communities in shallow soils, shrubland on sandplains, woodland on sandy crests and valley divides), nine types and ten sub-types (Fig 4, Table 3). These were generally sites in broad sandy terrain. Dune Vegetation Communities Supergroup. (B: 39 quadrats and 268 taxa) included two community groups (Merrup Dunes, Interdune plain and swamp), eight types and ten sub-types in aeolian dunes (Fig 5, Table 4). Swamp and outcrop Communities Supergroup. (C: 46 quadrats and 524 taxa) included three community groups (wandoo woodland and outcrop, saline swamps, and peat swamps) and 19 types (Fig 6, Table 5). These sites were in swampy terrain or areas of impeded drainage or outcrop. The limited sampling and heterogeneity of these community types prevented as¬ sessment below the 19 group level. Open-forest Communities Supergroup. (D: 136 quadrats and 428 taxa) included two community groups (jarrah/marri/tingle open-forest, jarrah/marri open- forest), five types and 14 sub-types (Fig 7, Table 6). 253 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 2 Summary statistics of communities supergroups and community groups of the Tingle Mosaic. Community Community Sub- taxa quadrats Community group (community types) supergroup types community types A: Shrubland /woodland communities 9 10 464 131 Al; Forest/ sandplain ecotone sites (3) A2: Woodland communities in shallow soils (2) A3: Shrubland on sandplains(2) A4: Woodland on sandy crests and valley divides (2) Walpole, Kentdale, King, Broke, King (swamps, plateau, valley divides) B: Dune vegetation 8 10 268 39 Bl: Merrup Dunes (4) B2: Interdune plain and swamp (4) Nullakai (dunes) C: Swamp and outcrop 19 19 524 46 Cl: Wandoo woodland and outcrop (11) C2: Saline swamps (5) C3: Peat swamps (3). All (swampy terrain, plateau and hills). D: Open-forest 5 14 428 136 Dl: Jarrah/marri/tingle tall open forest (2) D2: Jarrah/marri open forest (3) Walpole, Roe, Kentdale, King, Pa rdalup, King, Redmond (hills, ridges, plateau) E: Tall open-forest 3 12 132 89 El: Karri /tingle tall-open forest (3) Walpole (hills and ridges) Figure 3. Centroids of the 44 community types clustered in three dimensions using semi-strong-hybrid multidimensional scaling. 254 Journal of the Royal Society of Western Australia, 79(4), December 1996 Jarrah/sheoak woodland Jarrah/sheoak woodland Jarrah/sheoak open-forest Shrubland Tall shrubland Low shrubland /open woodland Open shrubland Jarrah/marri open forest Forest /shrubland ecotone Forest/ shrubland ecotone 0.7620 0.8296 0.8972 0.9648 1.0324 1.1000 Figure 4. Dendrogram classification of the 131 sites (based on 424 vascular plant taxa) of the Shrubland/woodland Communities Supergroup (A); community type and sub-type are shown in brackets. 0.7240 1 Shrubland (10 ) — Banksia woodland (11 )— Coastal herbland (12a) n Coastal herbland (12b) Tall shrubland m sumps (13 ) — Shrubland on old dunes (14 ) — Shrubland on new dunes (15 ) — Shrubland on limestone (16a) — , Shrubland on limestone (16b) — ' Heathland on headlands (17 ) 0.7240 0.7912 I 0.8584 I 0.9256 I 0.9928 I 1.0600 I 0.7912 I 0.8584 0.9256 I 0.9928 I 1.0600 Figure 5. Dendrogram classification of the 39 sites (based on 241 vascular plant taxa) of the Dune vegetation Communities Supergroup (B); community type and sub-type are shown in brackets. Table 3 Description of community types in Shrubland/woodland Communities Supergroup (Communities Supergroup A). Community- Description typea l(a,b) Jarrah/sheoak 2 Jarrah/sheoak open forest and woodland 3 Shrubland to tall shrubland 4 Tall shrubland 5 Low shrubland to open woodland 6 Open shrubland 7 Jarrah/marri open forest and woodland 8 Forest-shrubland ecotone 9 Forest-shrubland ecotone Number of Species Landform/soilsb quadrats* 2 3 4 5 6 7 8 9 richness3 49 (39, 10) 43.0 (42.9, 43.6) 15 36.3 11 23.4 3 29.3 22 47.2 11 41.9 10 40.7 5 40.8 5 21.0 Freely drained sands in uplands and valley divides (CA, Bwp, Ds, Dc, Q). Fertile sands with some loam (WA, CA, HA). Humus podzols in gently sloping sandy terrain (HA, A). Peaty podzols in lower slopes of sandy terrain (HA, A). Shallow gritty duplex soils/podzols (Lp, Kp, Mtp, Cop, BU). Humus podzols (F, CA). Sands (HA, A). Leached sands and podzols (Gs, Ks). Sandy yellow duplex soils (A, HA) 3 values for sub-communities is parentheses; b units of Churchward et al (1988). 255 Journal of the Royal Society of Western Australia, 79(4), December 1996 0.7640 Shrubland 1 (18) - Shrubland (19 ) Closed shrubland (20) Blackbutt woodland (21 ) - Melaleuca woodland (22) - Melaleuca woodland (23) Open shrubland (24) (25) -1 Melaleuca woodland Granite outcrop (26) - Granite outcrop (27) Granite outcrop (28) - Fertile valley floors (29) - Yate woodland (30) Granite outcrop (31 ) - Granite outcrop (32) - Granite outcrop (33) - Blackbutt woodland (34) - Wandoo woodland (35) - Open woodland (36) - 1 1 0.7640 0.8372 I 0.9104 0.9836 I 1.0568 I 1.1300 I I 0.8372 0.9104 I 0.9836 I 1.0568 I 1.1300 Figure 6. Dendrogram classification of the 46 sites (based on 524 vascular plant taxa) of the Swamp and outcrop Communities Supergroup (C); community type and sub-type are shown in brackets. Tall open karri/marri/tingle forest Tall open karri/marri/tingle forest Open jarrah/marri forest Open jarrah/marri forest Open jarrah/marri forest Open jarrah/marri forest Low open/tall open jarrah/marri forest Low open/tall open jarrah/marri forest Low open/ tall open jarrah/marri forest Low open/tall open jarrah/marri forest Low open/tall open jarrah/marri forest Open jarrah/marri forest Open jarrah/marri forest Low open jarrah/marri forest 0.6280 0.7444 0.8608 0.9772 1.0936 1.2100 Figure 7. Dendrogram classification of the 136 sites (based on 428 vascular plant taxa) of the Open-forest Communities Supergroup (D) community type and sub-type are shown in brackets. v These were in hilly terrain, but with poor soils or moisture holding capacity Tall open-forest Communities Supergroup. Separate community groups were not recognised in this commu¬ nity supergroup (E: 89 quadrats and 136 taxa). How¬ ever, three types and 12 sub-types were defined in this supergroup (Fig 8, Table 7) which included sites occur¬ ring in loamy soils in freely drained upland areas with good moisture retention capabilities. Considerable varia¬ tion in vegetation structure was noted in the three non¬ forest supergroups. Community types and species richness The Tingle Mosaic was rich in species, although quadrats in tall-open forest were species-poor (Table 7). Species richness between quadrats was not normally dis¬ tributed (Fig 9). This distribution reflected the uneven sampling effort within the study area. There was a large number of quadrats in tall open-forest writh low species richness, and in open-forest and woodland with high species richness. Community types are species-rich in all but relatively fertile freely-drained upland sites, and ex¬ treme areas such as outcrop or saline seepage. Species richness is very high in open-forest and woodland envi¬ ronments on shallow or sandy soils. The high richness of the Tingle Mosaic manifested itself through a very high proportion of singletons within quadrats of the survey area, particularly in the Open-forest and Swamp and outcrop Communities Supergroups. Notably, over 25 % of species occurred in only a single quadrat (singletons - Table 8). A high pro- 256 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 4 Description of community types in the Dune vegetation Communities Supergroup (Communities Supergroup B) Community- Description Number of Species richness2 Land form/ soilsb type2 quadrats2 10 Shrubland. 2 34.0 Interdune plains in older D'Entrecasteaux dunes 11 Banksia woodland. 4 39.7 Seasonally inundated, interdune plains in D'Entrecasteaux dunes. 12 (a,b) Seasonally inundated coastal herbland. 3(1/ 2) 27.3 (27.0, 28.0) Seasonally inundated, freely drained D'Entrecasteaux dunes. 13 Tall shrubland in sumps 1 20.0 Sumps in BWp. 14 Shrubland on older dunes. 13 40.1 Older D'Entrecasteaux dunes. 15 Shrubland on recent dunes 10 36.0 Recent D'Entrecasteaux dunes. 16 (a,b) Closed shrubland on 4 (1, 3) 26.5 (28.0, 26.0) Limestone substrate in limestone substrate D'Entrecasteaux dunes. 17 Open shrubland to 2 29.5 Coastal granite headlands in low open heathland on coastal headlands D'Entrecasteaux dunes. a values for sub-communities in parentheses; b units of Churchward et al. (1988). Table 5 Description of community types in the Swamp and outcrop Communities Supergroup (Communities Supergroup C). Community type Description Number of quadrats Species richness Land form/ soilsa 18 Shrubland. 7 39.1 Broad peaty interdune plains. 19 Shrubland. 1 28.0 Permanently moist freely drained peaty swamps. 20 Closed shrubland. 3 23.0 Broad seasonally inundated peaty sands 21 Blackbutt woodland. 5 40.2 Seasonally inundated clay loams. 22 Melaleuca woodland 2 35.5 Seasonally inundated. 23 Melaleuca woodland 2 24.0 Seasonally inundated estuarine habitats. 24 Open sedgeland. 4 24.7 Seasonally inundated clay soils. 25 Melaleuca woodland 1 20.0 Seasonally inundated sumps. 26 Granite outcrop. 1 29.0 Coastal granite outcrop. 27 Granite outcrop. 2 29.5 Sand amongst granite outcrop. 28 Granite outcrop. 2 17.0 Loam amongst granite outcrop. 29 Tall Shrubland to Tall open-forest 2 17.0 Fertile minor valley floors. 30 Yate woodland. 1 55.0 Clay valley floors. 31 Granite outcrop. 1 66.0 Shallow soils on granite outcrop 32 Granite outcrop. 1 39.0 Shallow soils on granite outcrop 33 Granite outcrop. 5 60.2 Shallow soils on granite outcrop 34 Seasonally inundated blackbutt woodland. 2 55.0 Clay-loams on seasonally inundated valley floors 35 Wandoo woodland 1 26.0 Low valley slopes 36 Open woodland. 3 43.0 Clay in granitic outcrop 2 units of Churchward et al. (1988). 257 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 6 Description of community types in the Tall open-forest Communities Supergroup (Communities Supergroup D). Community-type Description Number of quadrats a Species richness3 Landform/soilsb 37(a, b) Tall-open karri /marri/yellow tingle forest 40 (12, 31) 34.0 (33.4, 38.7) Brown/yellow gravelly freely drained upland (Ky, My, COy, VI) 38(a, b, c, d, e) open jarrah/marri forest 31(3, 8, 6, 14) 42.8(55.7, 41.1,42.5, 40.0 Brown-gravelly freely drained upland (Kb, Mb, COb, VI) 39(a, b, c d, e) 40(a, b) Low-open to tall- open jarrah/marri forest Open jarrah/marri 47(9, 17, 6, 9, 5) 53.5(65.2, 54.8, 50.3, 41.2, 45.8) Gravelly upland sites - including block laterite (Ky, Kp, My, COy). forest 6(5, 1) 47.5(46.0, 54.0) Shallow gritty soils amongst rock outcrop (Ly, Ls, Lg, Ks). 41 Low-open jarrah/marri forest associated with granite outcrop. 12 55.4 Shallow gritty soils amongst rock outcrop with laterite (COp, COy Mtp, Mty, Ly, Lp). a values for sub-communities in parentheses; b units of Churchward et al. (1988). Tall open karri/red tingle forest Tall open karri/red tingle forest Tall open karri/red tingle forest Tall open karri /red tingle forest Tall open karri/red tingle forest Tall open karri forest Tall open karri forest Tall open karri forest Tall open karri forest Tall open karri forest Tall open karri forest Tall open karri forest 0.5610 0.6284 0.6958 0.7632 0.8306 0.8980 Figure 8. Dendrogram classification of the 89 sites (based on 132 vascular plant taxa) of the Tall open-forest Communities Supergroup (E); community type and sub-type are shown in brackets. Table 7 Description of community types in the Tall open-forest Communities Supergroup (Communities Supergroup E). Community-type a Description Number of quadrats3 Species richness3 Landform/soilsb 42(a, b, c) Tall-open karri /red tingle forest 78(12, 20, 10, 29, 7) 19.3(18.1, 21.8, 24.5, 17.0, 16.0) Brown-gravelly freely drained upland (Kb, Mb, COb, VI) 43(a, b, c, d, e, f) Tall-open karri forest 10(1, 1,5, 1,3,1) 19.1(16.0, 25.9, 17.2, 15.0, 33.0 16.0) Brown-gravelly freely drained upland (Kb) 44 Tall-open karri forest 1 12.0 Brown=gravelly freely drained upland (Kb) 3 values for sub-communities in parentheses; b units of Churchward et al. (1988). 258 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 8 Comparisons between three floristic surveys of the high rainfall zone (HRZ) of south-western Australia. Survey Authors Area km2 Species (analysed) Singletons (%) Quadrats Community types Tingle Mosaic Wardell-Johnson 3 700 (857) 214 (25) 441 44 (75 subtypes) Swan Coastal Plain Gibson et al. (1994) 4 000 1485 (1097) 272 (25) 509 30 (43 subtypes) South Coast Gibson (pers. comm.) 2 000 910 (877) 214 (24) 301 40 portion of singletons has also been noted in other floris¬ tic studies carried out in the HRZ (Table 8). Endemism in the Tingle Mosaic A total of 20 taxa, endemic to the Tingle Mosaic were encountered in quadrats located in this study area (Table 9). At least another 13 taxa are known to be endemic to the study area, but were not located in quadrats (Table 9). This included four dominant forest eucalypts, a hybrid and a eucalypt taxon (E. Virginia ms) previously collected in the area (1961) but not recognised as unique until the present study. Many of these locally endemic taxa are also rare (Table 9). A substantial number of collections made during this study require further clarification of taxonomic status. Thus additional taxa may subsequently be found to be restricted to the Tingle Mosaic. The considerable variation within what is currently accepted as a species suggests that several genera require major revision. These include Hcrnigenia, Astartea, Baeckea, Agon is, Calandrinia, Aotus, Chorizema, Daviesia, [acksonia , Latrobca , Leucopogon , Logania, Olearia and Hibbertia. The distribution of the endemic taxa is not even. For example 29 priority flora are known to occur within 10 km of Mt Lindesay (314 km2) compared with 130 for the whole of the Southern Forest Region (14 400 km2). This is a ratio of 10.2:1 on an area basis. Many taxa found most commonly in drier or more seasonal environments than the Tingle Mosaic have lim¬ ited ranges within the Tingle Mosaic (Table 10). These are usually confined to upland north-east facing slopes (e.g. Banksia gardneri var. brevidentata in the Soho Hills) or deep sands (e.g. B. coccinea in Redmond Forest Block) within the Tingle Mosaic. Many taxa that are confined to high rainfall, less seasonal areas and do not extend into the drier parts of the Tingle Mosaic, have range limits within the area (e.g. Anthocersis sylvicola, Lomandra ordii and Reedia spathacea, the latter being confined to peat swamps west of the Bow River catchment). Other spe¬ cies occurring in peat swamp habitats, such as Cosmelia rubra and Cephalotus follicidaris have distributions centred in the Tingle Mosaic. Table 9 Taxa endemic to the study area. Taxon r Actinotus sp Walpole (/ R Wheeler 3786) Alexgeorgea ganopoda Andersonia aff. setfolia (? A. macronema) Andersonia auriculata Andersonia sp Coll is (G Wardell-Johnson 5 A) Andersonia sp Middle Rd.( A R Annels 1059) Andersonia sp Mitchell River (B G Hammersley 925) w‘ Andersonia sp Mt. Lindesay (J A Cochrane 405) Anthocersis sylvicola ms. (P G Wilson 6312) Boronia virgata Borya longiscapa Bossiaea webbii * Caladenia evanescens r Calothanmus sp Mt. Lindesay (B G Hammersley 439) r Cryptandra congesta Eriochilus scaber subsp orbifolia ms Eucalyptus brevistylis Eucalyptus ficifolia Eucalyptus ficifolia x calophylla E u calyp tus g u i Ifoylei Eucalyptus jacksonii Eucalyptus Virginia ms (A R Annels 3107) Gastrolobium brown u Grevillea fuscolutea Lambertia aff. uniflora (A R Annels 1024) Microtis globula Rorippa dictyosperma (G J Keighery 11945) Sollya drommondii r Spyridium riparium #‘ Tetratheca elliplica *' Thelyrnitra jacksonii ms Trymaliutn venustum r Verticordia apecta Figure 9. Numbers of quadrats plotted against species richness for 304 quadrats in the Tingle Mosaic. 259 'priority taxa; "taxa not recorded during this survey. Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 10 Taxa with ranges ending in the Tingle Mosaic. Western limit Acacia biflora Grevillea umbellulata subsp acerosa Acacia luteola Hakea lasiantha Acacia sulcata Isopogon latifolius Agonis marginata Lambertia echinata subsp citrina Banksia coccinea # Latrobea sp South Coast (Ashby 1949) Banksia gardneri var gardneri # Lepidium pseudotasmanicum (SW limit) Banksia gardneri var brevidentata # Lepyrodia hermaphrodita Banksia goodii # Lepyrodia monoica Banksia verticillata # Lysinema lasianthum Billardiera sp South Coast (A R Annels 227) Melaleuca sp (A R Annels 863) Brachysema sericeum Melaleuca violacea * Chorizetna rcticulatum Monotoca tamariscina * Conostylis misera Neruda crenulata Dryandra serra Platy theca juniperina Eucalyptus angulosa Rinzia schollerifolia * Eucalyptus buprestium Schoenus trachy carpus 8 Eucalyptus decurva #* Sphen o to rna dru m tnon d i i Eucalyptus dorotoxylon # Sphenotorna parviflorum Eucalyptus missilis Synaphea polymorpha Eucalyptus occidentalis ' Verticordia endlicheriana var angustifolia Eucalyptus staeri # Verticordia fitnbrilepis subsp australis Eucalyptus wandoo (SW limit) # Xanthosia singuliflora Eastern limit ” Actinotus laxus ms Lepyrodia extensa ms Chaetanthus leptocarpoides # Lo man dr a ordii Chorizema rctrorsum * Melaleuca ringens Eucalyptus calcicola # Restio jacksonii Gafmia sp Yelverton (G J Keighery 10820) # Restio ustulatus Hemigenia microphylla (SE outliers) # Sporodanthus rivularis ms *’ Hypocalymma sp Scott River (A S George 1177) Patersonia umbrosa var xanthina Reedia spathaceae Stylidium laciniatum * Hypolaena caespitosa ms Stylidium pritzelianum ” Hypolaena viridis ms Taraxis glaucescens ms 8‘ Kennedia glabrata Tar axis grossa * Lambertia orbifolia Southern limits and outliers 8 Chamelaucium forrestii subsp forrestii ms #* Drakaea micrantha 8 Darwinia thymoides #* Epiblema grandiflorum var cyanea ms Diuris drummondii # Grevillea cirsiifolia 'priority taxa; 8 taxa not recorded during this survey. Discussion The Tingle Mosaic includes both great floristic rich¬ ness and many rare and locally endemic plant species. Detailed taxonomic work on the collections made dur¬ ing this survey is likely to provide further insight into the historical biogeography, and evolutionary history of the area. Other floristic studies of the high rainfall zone (HRZ) in the Swan Coastal Plain and along the south- coast, which have included over 1200 quadrats and 2000 species, demonstrate the individualistic nature of the floristics of the HRZ and that over 25% of the quadrat- based flora records are of taxa recorded only in a single quadrat (Gibson et al. 1994; N Gibson, CALM, pers. comm. 1995). Although new taxa continue to be discov¬ ered in forest sites, these are the least variable of the community types at a landscape scale. Sites in tall-open forest tend to be poorest in species (excepting extreme sites), with least variation across the landscape. The overall floristic diversity of the HRZ is very high. The richness of the flora of the region is related to this rich landscape pattern, which in turn is associated with a diverse climatic and edaphic history. Endemism in the Tingle Mosaic The Tingle Mosaic is notable for high species richness of locally endemic species. Several other south-western areas are also notable for high richness of local endemic and rare species (e.g. Whicher Range, Darling Scarp). However, none are notable for the high numbers of lo¬ cally endemic dominant species described by this study. For example, five species of dominant forest eucalypts are locally endemic to the Tingle Mosaic. These species were considered by Wardell-Johnson & Coates (1996) to be indicators of a non-mobile, small-scale relictual biota that is confined to the region. Swamp and outcrop sites are likely to have been im¬ portant refugia for both the mesic and dry country ele¬ ments of the biota during the major climate fluctuations 260 Journal of the Royal Society of Western Australia, 79(4), December 1996 since the mid Tertiary (Hopper 1979; Hopkins et al. 1983). Isolation of populations in these sites has led to differentiation and speciation in some woody plant gen¬ era (c.g. Agonis, Andersonia, Chamaelaucium and Leucopogon). Eucalyptus brevistylis, a tall forest tree, is the largest species endemic to granite outcrop sites. This species is also locally endemic to the Tingle Mosaic and associated with the moisture gaining sites at the base of granite outcrops in areas of high relief. The high levels of endemism associated with upland granite outcrop areas is no doubt associated with the geological and climatic history of the area. Thus sites relatively high in the landscape may have become is¬ lands during marine transgressions. Large islands such as Mt Lindesay may have retained a greater array of habitat-types than smaller islands such as Granite Peak, Mt Frankland and Mt Roe. These smaller peaks retain many rare and locally endemic taxa, but only a small proportion of the total within the region in comparison with Mt Lindesay. Association of floristics with environmental attributes The Tingle Mosaic features high landscape diversity encompassing hills and ridges, granite monadnocks, swamp, steep river valleys, dune systems and coastal cliffs. This area includes vegetation types ranging from tall open-forests to herblands and includes high levels of heterogeneity immediately adjacent to the forests and woodlands. Thus 36 of the 44 community types are out¬ side the forests. Of these, six are community types oc¬ curring on outcrops and 22 occur in swamp habitat. Both these landscape features include high gamma plant diversity. The degree to which community types are associated with the landform soils units of Churchward et al. (1988) is likely to vary between community types. Soil type appears to be stronger than landform in its associations with community types in hill and ridge areas of granitic base rock. However, at a fine-scale, considerable varia¬ tion has been noted within the community sub-types of the tall open-forests of the area. For example, Inions et al. (1990) examined variation in floristics within regener¬ ating karri forest over a major part of the range of karri. They defined 13 community types on the basis of floris- tic variation, each differing in productivity as measured by age-standardised top-height. This was despite the finding of Wardell-Johnson et al. (1989) that karri forest displayed the lowest alpha and gamma diversity of the 12 community types that they defined in the Walpole- Nornalup National Park. Twelve community sub-types are defined within the three community types of the Tall open-forest Communities Supergroup in this study. The swamps of the area are important features of the landscape and exhibit great variation from peat swamps, estuaries, lakes and playas. This diversity of swampland has been recognised in land form-soils mapping of the south-west. Of the 37 land form-soils units mapped by Churchward et al. (1988) along the south coast, 2 are valley units (17 sub-units), and 16 are units in swampy terrain. Thus half of the units identified by Churchward et al. (1988) are based on riparian or swampy terrain, although these occupy a minor proportion of the total landscape of the study area. Although the topography is muted, the origin and expression of this variation is not. and sharp ecotones between communities supergroups are a feature of the Tingle Mosaic (Wardell-Johnson et al. 1989). Floristic pattern in granite outcrop and swamp com¬ munities reflects high levels of complexity in landform soils mapping in these environments (Churchward et al. 1988). Swamp and outcrop communities have high gamma diversity. An expanded program of survey would be required to target the exceptional variety of environments in the Swamp and outcrop Communities Supergroup. The high water table in areas of swamp vegetation leads to a close link between water table and community structure in an area of great edaphic com¬ plexity. These community types are also likely to be most vulnerable to changes in land use. There is considerably less floristic diversity occurring in tall open-forest than in other vegetational structural types. Two community supergroups (D and E), repre¬ senting eight community types, occur in hills and pla¬ teau landform units and include forest. Open-forest, tall open- forest and woodland communities included most of the quadrats (356 of 441) and also occupied the larg¬ est area within the region. Thus hill and plateau units represent over 54% of the total survey area but include few of the community types. The open-forest areas in¬ clude high levels of a diversity, and occur on shallow and infertile soils of this high rainfall zone. Integration of floristic classifications and landform soil mapping There are many site-based floristic studies for, or near, the Tingle Mosaic. Wardell-Johnson et al. (1989) developed a floristic classification of the Walpole- Nornalup National Park based on 219 quadrats and 233 species. Inions et al. (1990) defined 13 community types on the basis of floristic variation within regenerating karri forest over a major part of the range of karri. Strelein (1988) defined seventeen site types based on the floristic composition of over 400 sites in the southern part of the range of jarrah using the methods of Havel (1968, 1975 a,b). Both Inions et al. (1990) and Wardell- Johnson et al. (1989) provided a means of allocating in¬ dependent sites to the classification using discriminant functions on species defined as indicators in the analysis (72 and 52 species respectively). Thus sites in one classi¬ fication can be defined according to another. Classifica¬ tions developed in both studies have used similar meth¬ ods and both schemes can be mapped (Ward & Wardell- Johnson 1993). However, although Hopper et al. (1992) concluded that an integration of site-based work (in the Warren Botanical Subdistrict) is desirable, considerable site revisiting would be required. The present study al¬ lows the integration of previous studies carried out over a small area, or within a subset of the variation in floris¬ tic composition (i.e. either in jarrah or karri forest) of the region. This work is required urgently and would allow an environmental context for the management of the re- gion. Floristic mapping Previous maps of the floristics of the Tingle Mosaic area include a vegetation map (Smith et al. 1991) which recognized the association of the twelve community types defined by Wardell-Johnson et al. (1989) with the 261 Journal of the Royal Society of Western Australia, 79(4), December 1996 landforms/soil units of Churchward et al. (1988). Never¬ theless, the complexity of the floristics in the Swamp and outcrop Communities Supergroup was underesti¬ mated by this work. Ward & Wardell-Johnson (1993) provided a trial mapping of the community types deter¬ mined by Inions et al. (1990) which found that these community types varied in a complex pattern in karri forest. They concluded that while mapping of commu¬ nity types was feasible, it was both expensive and time- consuming. They suggested that resources would be bet¬ ter spent integrating remote sensed imagery with site- based quadrat data to derive models of vegetation com¬ munities. The distribution of community types defined by Inions et al. (1990) is broadly geographically based (Wardell-Johnson & Christensen 1992), although overlap occurs within a single landform/soils unit (as defined by Churchward et al. 1988). The forests, however, have been mapped at fine scale. The whole of the HRZ (apart from gaps in the Leeuwin Naturaliste Ridge) was vegetation mapped at 1:25 000 scale, from aerial photographs of 1:40 000 scale, during the 1950s and 1960s. A series of 233 maps of the area were produced and the data digitised for the Forest Management Information System (FMIS) as 2 hectare square grid cells. These data were updated, usually us¬ ing 1:25 000 scale colour aerial photography during the late 1970s. These maps provide the structure, density and floristic composition of the overstorey of all of the south-western forests (excepting that area mentioned above). The Forests Department aerial photography in¬ terpretation (API) type-maps used a structural classifica¬ tion system of plant communities which combined height/ life form of dominant plants and the projective area of ground covered by the foliage of the dominant plants in the ecosystem. Within these structural subdivi¬ sions, species composition of the overstorey was used to define forest type. There is a need to amalgamate the plant communities defined by this study with those from other studies ( e.g . Strelein 1988; Wardell-Johnson et al. 1989; Inions et al. 1990; Gibson et al. 1994) and with broad scale vegetation mapping (e.g. Beard 1972-80; Smith 1972, Hopkins et al. 1995). There is also a need to access existing information in relation to the distribution of forest types particularly that available at 1:25 000 scale API maps prepared dur¬ ing the 1960s. This in conjunction with remote sensed imagery will allow more effective extrapolation of exist¬ ing quadrat-based data. The continuum of vegetation community types in the Tingle Mosaic requires acknowl¬ edgment in vegetation mapping, as with research in the northern jarrah forest by Heddle et al. (1980). Thus the landform/soils maps of Churchward et al. (1988) should form the basis of maps of complexes of community types in the Tingle Mosaic. The boundaries of these sub¬ communities are often diffuse. The presentation of sum¬ mary information of individual community types (and sub-types) and their geographic distribution would be helpful in the mapping of these environments. This would be of considerable benefit to managers of the en¬ vironment of the Tingle Mosaic. Hierarchies of mapping units for land resource and vegetation survey allow consistent approaches with mapping scales at varying levels of complexity. Higher levels of complexity of the biota are revealed at finer scales of mapping. At the local scale, a sound perspec¬ tive for reserve selection and design, and for ecosystem management is provided by studies covering a broad range of environments and incorporating a thorough as¬ sessment of the biota. A hierarchy of mapping units for land resource survey prepared for the Darling District (P Tille, Agriculture WA, pers. comm , 1995) includes the HRZ. Can this approach also be used in reserve system design and management? A similar hierarchy of vegetation types, based on the structural types of Beard (1980-1981) has been developed for the State of Western Australia (Hopkins et al, pers . comm.) using aerial photographic interpretation in com¬ bination with detailed field assessment. The digitised vegetation types were digitised from the original line- work and 1: 250 000 scale maps prepared by Beard (1980, 1981). His scheme is based on physiognomy simi¬ lar to that of Specht (1981), but he classified the vegeta¬ tion according to the ecologically dominant stratum rather than the tallest stratum. This recognises 823 types, 199 groups and 50 supergroups. The supergroups are used in producing a new 1: 3 000 000 scale map of the vegetation of Western Australia. Although the difficulty of mapping vegetation com¬ munities in topographically subdued landscapes with an apparently homogenous overstorey has been amply demonstrated (Havel 1975a,b), mapping complexes of such communities has become an important tool to de¬ termine reservation status and to guide its management (Heddle el al. 1980). Structural vegetation types and landform/soils mapping, both derived from the inter¬ pretation of remotely sensed imagery (usually aerial photography) are useful tools when used in conjunction with regional floristic survey. The amalgamation of flo¬ ristic studies from the HRZ will allow a more effective overall hierarchy to be developed of plant community assemblages in the area. Conclusions The Tingle Mosaic region had already been noted for its high diversity of herbaceous perennial taxa (Hopper et al. 1992), but is also notable for diversity among shrub and tree perennial taxa. This study identified many new taxa, and shows that the area is an important refuge for many high rainfall relictual woody taxa besides the tingle trees after which the area has been designated. The ancient and complex geological history marked by prolonged leaching and erosion, deposition and lateritization of the land-surface has resulted in a sub¬ dued landscape despite its complex ontogeny. This var¬ ied climatic and edaphic history has no doubt contrib¬ uted to the present richness of the flora. The present study quantifies this variation for the vascular plants, while Horwitz (1994) demonstrated the high levels of endemism in the freshwater invertebrates of the region. Wardell-Johnson & Horwitz (1996) demonstrated the need for a new approach in fine-scale management in the HRZ to reflect and account for the biotic variation in the region. This study has provided a means for deriving an un¬ derstanding of the local scale pattern of the biota in ar- 262 Journal of the Royal Society of Western Australia, 79(4), December 1996 eas of high biotic richness and high levels of land use pressure. The integration of site-based work, the defini¬ tion of complexes of community types, and the map¬ ping of these floristic assemblages are applications of this research which would be invaluable in the manage¬ ment for the conservation of biodiversity in the region. Acknowledgments : We thank T Annels, C Vellios, G I.iddelow and I Wheeler for assistance in the field, Y Winchcombe and N Gibson for ad¬ vice on database management data entry software. M Lyons converted the floristic database to the SEDIT scheme. R Hearn and T Annels contributed the data on endemism and rarity in the region. Many individuals who assisted with the locations of rare taxa or unusual sites in the Tingle Mo¬ saic. These included T Annels, B and G Jackson, A Syme, B Shur, C Chappelle, A Luscombe, B Hammersley and A Wardell-Johnson. Many colleagues are thanked for their identifications in their particular areas of expertise; T MacFarlane, G Keighery, N Gibson, P Wilson, B Rye, B Maslin, A Brown, J Wheeler, H White and R Cranficld from CALM; S Hopper and K Meeney from the Kings Park and Botanic Garden; P Short, National Botanic Garden, Melbourne; M Crisp, Australian National University; L Craven, CSIRO; K Wilson, Royal Botanical Gardens, Sydney, and A Lowrie, K Lemson, A George and F. George. Many volunteers who as¬ sisted with locations of rare taxa or unusual sites in the Tingle Mosaic included B Jackson, A Syme, B Shur, B Hammersley and A Wardell- Johnson. We also thank the Walpole-Nomalup National Parks Association and the Walpole District and Southern Region of the Department of Con¬ servation and Land Management for continued support. We thank CALM and the Australian Heritage Commission for support and funding. We also thank the Walpole-Nomalup National Park Association and the Walpole District staff for continued interest in the project, D Bell, D Coates, N Gibson, P Tille, M Churchward, A Hopkins, S Hopper, B Main, B York Main, A Wardell-Johnson and P Withers for constructive comments and for helpful discussion and advice. References Anderberg M R 1973 Cluster Analysis for Applications, Aca¬ demic Press, New York. Anderson D J 1981 Introductory notes. In: Vegetation Classifica¬ tion in Australia (ed A N Gillson & D J Anderson). CSIRO/ ANU Press, Canberra. Austin M P & Cunningham R B 1981 Observational analysis of environmental gradient. Proceedings of the Ecological Society of Australia 11:109-119. Beard J S 1972-80 Vegetation Survey of Western Australia. Vegmap Publications, Perth. Beard J S 1980 A new phytogeographic map of Western Austra¬ lia. Western Australian Herbarium Research Notes 3:37-58. Beard J S 1981 The Vegetation of Western Australia at the 1:3 000 000 scale. Explanatory notes. Dept Agriculture, Perth Belbin L 1993 PATN pattern analysis package. CSIRO, Canberra. Booth T H 1978 Numerical classification techniques applied to forest tree distribution data. 1. A comparison of methods. Australian Journal of Ecology 3:297-306. Burbidge A H & Boscacci L 1987 A biological survey of lands proposed for release for agriculture in the Manjimup Region. Report. Department of Conservation and Land Management, Perth. Campbell H, Armstrong R, Raynor M & Pearce C 1985 Site index curves for regrowth karri (Eucalyptus diversicolor) in south western Australia. In: Modelling Trees, Stands and Forests. CSIRO, Melbourne University, Melbourne. Christensen P 1980 A forest worth preserving. Report. Conserva¬ tor of Forests, Perth. Churchward H M, McArthur W M, Sewell P L & Bartle G A 1988 Landforms and soils of the south coast and hinterland, W.A. Northcliffe to Manypeaks. CSIRO Water Resources Divisional Report 88/1. CSIRO, Perth. Cope R N 1975 Tertiary epeirogeny in the southern part of West¬ ern Australia. Annual Report 1974. Western Australian Geo¬ logical Surveys Perth, 40-46. Cresswell I D & Bridgewater P B 1985 Dune vegetation of the Swan Coastal Plain, Western Australia. Journal of the Royal Society of Western Australia 67:137-148. Czekanowski J 1913 Zarys Method Statystycznyck. Warsaw. Diels L 1906 Die Pflanzenwelt von West-Austral ian Sudlich des Wendekreises. In: Die Vegetation der Erde Vol VII (eds Engler & O Drude). Engelman, Leipzig. Femie L & Fernie G 1989 In praise of a national park: The origins and history of the Walpole-Nornalup National Park. Fernie & Fernie, Perth. Foran B D Bastin G & Shaw K A 1986 Range assessment and monitoring in arid lands: the use of classification and ordina¬ tion in range survey. Journal of Environmental Management 22:67-84. Gentilli J 1971 Climates of Australia and New Zealand. In: World Survey of Climatology, Vol 13. Elsevier, Amsterdam. Gibson N, Keighery B J, Keighery G J, Burbidge A H & Lyons M 1994 A floristic survey of the southern Swan Coastal Plain. Report. Heritage Council of Western Australia, Perth. Green J W 1985 Census of the Vascular Plants of Western Austra¬ lia. Department of Agriculture, Perth. Green R H 1980 Multivariate approaches in ecology: the assess¬ ment of ecological similarity. Annual Review of Ecology and Systematic* 11:1-14. Havel J J 1968 The potential of the northern Swan Coastal Plain for Pinus pinaster Ait. plantations. Bulletin 76. Forests Depart¬ ment of Western Australia, Perth. Havel J J 1975a Site vegetation mapping in the northern jarrah forest (Darling Range): 1 Definition of site-vegetation types. Bulletin 86. Forest Department Western Australia, Perth. Havel J J 1975b Site vegetation mapping in the northern jarrah forest (Darling Range): 11 Location and mapping of site-veg¬ etation types. Bulletin 87. Forest Department of Western Australia, Perth. Havel j J 1981 Vegetation classification as a basis for land use planning. In: Vegetation Classification in Australia (eds A N Gillson & P J Anderson). CSIRO/ ANU Press, Canberra. Heddle E M, Loneragan O W & Havel J J 1980 Vegetation com¬ plexes of the Darling system, Western Australia. In: Atlas of Natural Resources: Darling system, Western Australia, De¬ partment of Conservation and Environment, Perth, 33-74. Hopkins A J M, Coker J, Beeston G R, Bowen P & Harvey J M 1995 Conservation Status of Vegetation Units Throughout Western Australia. Australian Nature Conservation Agency National Reserves Systems Co-operative Program Project No N703 Final Report. Hopkins A J M & Griffin E A 1984 Floristic patterns. In: Kwongan Plant Life of the Sandplain (eds J S Beard & J S Pate). University of Western Australia Press, Perth, 69-83. Hopkins A J M, Keighery G J & Marchant N G 1983 Species-rich uplands of south-western Australia. Proceedings of the Eco¬ logical Society of Australia 12:15-26. Hopper S D 1979 Biogeographical aspects of speciation in the south-western Australian flora. Annual Review of Ecology Systematics 10:399-422. Hopper S D 1992 Patterns of plant diversity at the population and species levels in south-west Australian mediterranean ecosystems. In: Biodiversity of Mediterranean Ecosystems in Australia (ed R J Hobbs). Surrey Beatty and Sons, Chipping Norton, 27-46. Hopper S D, Keighery G J & Wardell-Johnson G 1992 Flora of the karri forest and other communities in the Warren Subdistrict of Western Australia. In: Research on the Impact of Forest Management in South-west Western Australia. Occasional Paper 2/92, Department of Conservation and Land Manage¬ ment, Perth, 1-32. Horwitz P 1994 Patterns of endemism in the freshwater fauna of the far southern peatlands and shrublands of south-western Australia. Report. Australian Heritage Commission and the Heritage Council of Western Australia, Perth. 263 Journal of the Royal Society of Western Australia, 79(4), December 1996 Inions C. B, Wardell-Johnson G & Annels A 1990 Classification and evaluation of sites in karri (Eucalyptus diversicolor) regen¬ eration. 11. Floristic attributes. Forest Ecology and Manage¬ ment 32:135-154. Jutson J T 1914 An outline of the physical geology (physiogra¬ phy) of Western Australia. Bulletin 61. Western Australian Geological Survey, Perth. Logan B W 1968 Western Australia. In: Quarternary Shorelines Research in Australia and New Zealand (ed E D Gill). Austra¬ lian Journal of Science 31:110. Smith F G 1972 Vegetation survey of Western Australia. Vegeta¬ tion map of Pemberton and Irwin Inlet, 1:250 000. Depart¬ ment of Agriculture, Perth. Smith V, Annear R, Hanley P, Metcalf V, Sutton A & Wardell- Johnson G 1991 Walpole-Nornalup National Park Manage¬ ment Plan. Department of Conservation and Land Manage¬ ment, Perth. Sneath P H & Sokal R R 1973 Numerical Taxonomy. Freeman, San Francisco. Speck N H 1952 The ecology of the metropolitan sector of the Swan Coastal Plain. MSc Thesis, University of Western Aus¬ tralia, Perth. Specht R L 1981 Vegetation. In: The Australian Environment (ed G W Leeper). CSIRO and Melbourne University, Melbourne, 12-20. Strelein G 1988 Site classification in the southern jarrah forest of Western Australia. Research Bulletin 2. Department of Conservation and Land Management , Perth. Teakle L J H 1953 Soil survey of the Manypeaks district, Albany Road Board, Western Australia. Leaflet No 2070. Western Australian Department of Agriculture, Perth. Ward C & Wardell-Johnson G 1993 Trial mapping of community types in regenerating karri forest, south-western Australia. Report. Department of Conservation and Land Management, Perth. Wardell-Johnson G & Christensen P 1992 A review of the effects of disturbance on wildlife of the karri forest. Department of Conservation and Land Management, Perth, Occasional Paper 2/92, 35-57. Wardell-Johnson G & Coates D 1996 Links to the past: local endemism in four species of forest eucalypts in south-western Australia. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (eds S D Hopper, J Chappill, M Harvey & N Marchant), 137-154. Surrey Beatty & Sons, Chipping Norton. Wardell-Johnson G & Horwitz P 1996 Conserving biodiversity and the recognition of heterogeneity in ancient landscapes: a case study from south-western Australia. Forest Ecology and Management 85(l-3):219-238. Wardell-Johnson G & Smith V 1991 An ancient land. Landscope 5:15-21. Wardell-Johnson G, Inions G & Annels A 1989 A vegetation classification of the Walpole-Nornalup National Park, South¬ western Australia. Forest Ecology and Management 28:259- 279. Wardell-Johnson G, Williams J, Hill K & Cummings R in press Historical biogeography and current distribution patterns of eucalypts. In: Eucalypt Ecology: Individuals to Ecosystems (eds J Williams & J Woinowski). Cambridge University Press, Cambridge. Webb L J, Tracey J G & Williams W T 1976 The value of struc¬ tural features in tropical forest typology. Australian Journal of Ecology 1:1-28. Webb L J, Tracey J G & Williams W T 1984 A floristic framework of Australian rainforests. Australian Journal of Ecology 9:169- 198. Whittaker R H 1973 Handbook of Vegetation Science, Ordination and Classification of Communities. Dr W Junk, The Hague. 264 Journal of the Royal Society of Western Australia, 79(4), December 1996 Appendix List of taxa for the Tingle Mosaic with constancy (percentage of sites occupied by the species) for each taxon of each community group (A1-A4, B1-B2, Cl- C3, D1-D2, El). ' 7 & K Community Group: A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Number in Group: 64 14 53 10 29 11 14 8 13 71 65 89 Adiantaceae Cjieilanthes austrotenuifolia 2 25 8 3 2 Aizoaceae Carpobrotua modest us 14 Amaranthaceae Ptilotus stirlingii var laxus 10 Anthericaceae Agros tocrinuni sea brum 9 17 10 14 15 6 38 Borya longiscapa 8 Borya sphaerocephala 13 8 Caesia parviflora 9 7 13 23 6 Chamaescdla corymbosa 7 8 14 50 54 1 6 Corynotheca micrantha var panda Johnson ia lupidina 66 14 53 7 18 14 8 25 Laxmannia jamesii 8 Ijjxtnannia minor 23 Sowerbaea laxiflora 10 55 15 3 Thysanot us a renari us 2 8 1 Thysanotus gracilis 4 20 14 9 1 2 Thysanotus manglesianus 2 7 Thysanotus multiflorus 42 14 28 20 27 21 31 11 37 2 Thysanotus patersonii 1 Thysanotus pauciflorus 6 3 Thysanotus sp 3 Thysanotus sparteus 2 43 23 6 Thysanotus tenellus 2 2 Thysanotus thyrsoideus 3 15 6 12 ^'iconpie elatior 5 2 15 2 Tricoryne humilis 20 14 7 8 1 15 Apiaceae Actinotus glomeratus 14 4 9 8 2 Actinotus omnifertilis Apium prostratum var prostratum Centella asiatica Daucus glochidiatus 2 11 10 3 27 7 Hydrocotyle callicarpa 7 15 Hydrocotyle plebeya Hydrocotyle sp 1 30 8 Hyd rocolyle tetragonoca rpa Pentapeltis silvatica 2 20 24 3 32 Platysace anccps 7 2 Platysace compressa 33 28 10 17 13 35 46 4 Platysace filiformis 8 5 Platysace pendula 27 8 9 3 Platysace tenuissima Trachymcne anisocarpa Trachymenc pilosa 2 30 14 31 3 Xanthosia Candida 2 3 8 4 6 Xanthosia fruticulosa 8 2 Xanthosia hederifolia 3 4 3 1 Xanthosia huegelii 17 11 10 11 32 Xanthosia rotundifolia 30 30 8 21 48 1 Aspleni aceae Asplenium aethiopicum 8 1 Asplenium flabellifolium 1 2 1 Astera ceae Aster idea pulverulenta 20 38 Brachyscome iberid ifol ia Cotula coronopifolia Conyza bonariensis ' 10 17 7 2 C rasped ia pleiocephala Gnaphali u m sphaeri c u m 17 7 13 23 2 Helichrysum cordatum 10 14 1 Helichrysum macranthum Helichrysum ramosum 3 15 13 5 27 Hyalosperma cotula 8 Hypochaeris glabra ' 8 7 50 31 10 2 8 hi gen if era huegelii 10 21 15 6 6 Millntia myosot id ifolia 10 14 Olearia aff paucidentata (GWJ 2959) Olearia axillaris 2 40 72 7 Olearia cassmiae Olearia ciliata 10 10 13 Olearia paucidentata Podolepis gracilis Pseudognaphalium luteoalbum ' 10 7 8 13 3 Senecio gilbert ii 2 13 10 2 2 265 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 Senecio glomeratus Senecio lautus subsp maritimus Senecio quadridentatus 40 79 7 13 Senecio ramosissim us 20 Siloxerus filifolius Sdoxerus humifusus 2 10 7 14 13 8 Sonchus oleraceus ‘ 7 13 8 Trichocline spathulata Vellereophyton dealbatum ' Waitzia citrma 20 48 8 3 Brassicaceae Stenopetalum robustum 7 28 9 7 Campanulaceae Wahlenbergia communis Wahl cm bergia gracilen ta 13 1 Casuarinaceae Allocasuarina decussata 8 7 65 15 Allocasuarma fraseriana 86 15 7 9 8 1 15 Allocasuarina humilis 2 20 21 46 2 Centrolepidaceae Aphelia cyperoides 6 7 15 Central epis a ns tat a 2 Cephalotaceae Ccphalot us foil ic ularis 9 Chenopodiaceae Atriplex prostrata ' Rhagodia baccata Sarcocomia quinqueflora 20 17 7 13 Colchicaceae Burchardia monantha Burchardia multiflora 13 13 10 9 7 13 31 2 Burchardia umbellata 45 29 36 9 13 5 Wurmbea dioica subsp alba Wurmbea monantha 7 13 Convolvulaceae Dichondra repens 3 Cyperaceae Baumea sp (GWJ 5231) 9 Baumca sp (GWJ 3056) 9 Baumea juncea 9 14 13 Baumea vaginalis Carex sp (GWJ 5229) 7 13 Cyathochaeta avenacea 6 14 36 9 36 13 31 11 25 Cyathochaeta clandestina 8 6 5 Evandra aristata 3 36 15 10 73 5 Evandra pauciflora Gahniafilum6 4 7 Gahnia decomposita 4 20 18 57 3 2 Gymnoschoenus anceps 9 36 Isolepis marginata e 8 10 9 8 Isolepis nodosa 1 30 31 14 13 8 Isolepis prolifera ' 13 8 Lepidosperma aff gracile (GWJ 5257) 8 Lepidospenna aff angustatum Lepidosperma aff tenue (GWJ 5258) 8 2 Lep id os perma a ngus tatu m 27 15 30 76 18 21 38 77 10 37 Lepidospenna effusum 8 40 3 75 15 56 2 l £pidosperma glad ia t u m 60 28 Lepidospenna gracile 10 1 Lepidosperma leptostachyum 13 29 8 10 17 25 15 80 65 Lep idos penna Ion git u d inale lepidosperma squama t u m 2 10 13 Lepidosperma tenue 3 23 3 15 1 14 Lep idosperma tetraquetrum Lepidosperma viscid n m Mesomelaena gracil iceps 13 9 3 9 25 3 Mesomelaena stygia * 11 7 17 18 38 12 Mesomelaena tetragona 8 14 66 43 46 1 15 Mesomelaena uncinata 8 Iieedia spathacea Cyperaceae sp big 9 2 Cyperaceae sp fine 2 Cyperaceae sp 1 Schoenus acuminatus 2 2 15 1 Schoenus aff rodwayanus 2 2 7 18 Schoenus bifidus h 16 15 20 27 3 15 Schoenus brevisetis a 5 9 15 Schoenus brevisetis vel sp aff 2 7 4 10 7 9 Schoenus curvifolius 2 Schoenus grandiflorus 3 15 24 1 15 Schoenus odonocarpus 7 5 266 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Schoenus subbulbosiis 2 13 1 Schoenus subflavus c 13 9 8 1 Tetraria octandra 3 8 Tricostularia neesii var elatior 6 2 Tricostularia neesii 6 2 3 9 5 Dasypogonaceae Baxtcria australis 2 8 9 Cnlectasia grandiflora Chamacxeros sp nov Dasypogon bromeli ifol i us 100 57 66 3 45 13 23 5 1 Kirigia australis 5 36 47 7 7 31 Lxmiandra caespitosa 2 4 3 1 2 Lomandra drummondii 20 30 9 8 27 46 3 Ixmiandra hermaphrodita 2 8 2 hmiandra Integra 9 8 34 34 6 Lomandra micrantha 2 Ixmiandra nigricans 28 25 9 17 8 11 Lomandra pauaflora 5 25 10 14 25 83 55 7 Lomandra preissii 3 Ixmiandra purpurea 2 8 3 Lomandra sericea 5 8 38 31 Lomandra sonderi 2 2 Dennstaedtiaceae Pteridium esculenlum 6 29 15 18 38 92 20 75 Dilleniaceae Hibberlia sp (GWJ 4822) 8 Hibbertia aff pulchra (GWJ 4183) 20 6 7 54 5 Hibbertia a triplex ica ulis 13 28 20 17 14 13 31 65 74 Hibbetita commutata 11 4 7 13 15 28 55 3 Hibbertia cuneiformis 4 20 62 4 2 16 H ib bert ta cun n ingfia m i i Hibbertia desmophylla 6 9 20 10 21 13 4 20 2 Hibbertia furfuracea 20 7 8 10 7 27 6 35 f Iibbertia glaberrima 5 10 3 1 I ibbertia grossulari ifolia 20 55 Hibbertia hypencoides Hibbertia inconspicua Hibbertia microphylla 16 2 3 9 15 3 I Iibbertia racemosa Hibbertia serrata 2 21 21 3 42 Hibbertia stlvestris Hibbertia stellaris 2 21 8 1 5 Droseraceae Drosera sp (climber) 2 9 15 1 3 1 Drosera erythrorhiza Drosera glanduligera 2 10 7 9 7 15 3 Drosera hamiltonii Drosera huegelii 2 4 9 13 8 1 2 Drosera macrantha subsp macrantha 2 Drosera menziesii 2 14 13 45 14 13 38 1 2 Drosera neesii 5 13 7 Drosera pallida 30 79 42 21 18 25 23 23 38 1 Drosera pulchcll 15 2 Drosera sp (rosette) 8 9 18 3 Drosera stolonifera subsp s tolonifera 9 Epacridaceae Andersonia aff caerulea (GWJ 1563) 5 9 Anderson ia auriculata 23 6 9 Andersonia caerulea A ndersonia lehmanniana 44 50 32 10 18 8 15 subsp lehmanniana 2 Andersonia micrantha Andersonia sprengelioides 6 2 7 9 38 54 2 Astral orna baxteri Astroloma ciliaturn 3 8 1 2 Astroloma drummondii 5 4 8 Astroloma epacridis 3 54 1 5 Astroloma pallidum 3 13 20 Brachyloma preissii Cosmcha rubra 13 34 27 7 15 6 Ixmcopogon sp (GWJ 4828) Ixmcopogon altemifolius 17 9 36 15 35 31 Ixmcopogon australis 16 36 58 64 15 Dmcopogon capitellatus 22 14 21 30 69 9 14 38 38 62 1 63 Leucopogon concinnus Leucopogon distans 28 2 7 2 Leucopogon gilbertii 2 4 18 7 23 12 leucopogon glabellus 44 8 Ixmcopogon gracilis 2 29 11 1 Leucopogon obovatus 16 59 18 4 6 Leucopogon parviflorus 8 4 267 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El L eucopogon pendulus Leucopogon polystachyus 2 4 18 Leucopogon propi nqu us 22 36 14 7 13 20 25 1 Leucopogon pulchellus Leucopogon reflexus 2 2 7 15 Leucopogon unilateralis 13 19 10 18 7 25 23 8 32 Leu copogon vert i cilia t us 11 21 7 8 92 86 80 Lysinema ciliatum 5 2 10 31 9 Lysinema conspicuum 2 13 8 Monotoca tamariscina 11 29 25 10 5 3 Sphenotoma capitatum 2 Sphenotoma gracile 6 29 43 30 82 15 8 Sphenotoma squarrosum Stypheha tenuiflora 2 9 20 Euphorbiaceae Amperea ericoides 13 8 17 7 9 2 Amperea simulans 11 2 Amperea protensa Amperea volubilis 2 10 18 Phyllanthus calycinus Poranthera aff huegelii (GWJ 2837) 10 38 13 2 Poranthera huegelii 6 4 7 34 Ricinocarpos glaucus 13 1 2 Gentianaceae Centaurium erythraea ' 2 7 13 2 Geraniaceae Geranium solanderi 28 Pelargonium australe Pelargonium capitatum ’ Pelargonium littorale 2 21 13 8 Goodeniaceae Dampiera alata 2 31 8 Dampiera hederacea 2 14 15 9 13 70 9 58 Dampiera linearis 92 93 83 60 64 71 31 17 54 3 Dampiera trigona Diaspasis filifolia 8 10 27 7 2 Goodenia caerulea Gooderua eatoniana 2 21 14 8 6 42 Goodenia filiformis var filiformis Goodenia leptoclada 9 7 Goodenia tenella 7 9 Lcchenaultia expansa 5 11 10 3 Scaevola crassifolia 24 Scaevola globulifera 3 9 1 Scaevola microphylla 2 2 9 25 6 38 Scaevola striata 84 66 10 3 18 7 25 23 25 55 1 Scaevola thesioides Velleia macrophylla 10 8 Velleia trinervis 2 4 40 28 9 21 23 Gyrostemonaceae Gyrostemon sheathii 3 Haemodoraceae Anigozanthos flavidus 14 8 80 13 34 9 21 Conostylis aculeata subsp aculeata 6 20 83 14 3 6 Conostylis setigera 27 9 12 Haemodorum laxum 2 2 20 9 21 8 3 3 Haemodoru m pani culat u m 2 Haemodorum spicatum 31 79 19 3 18 8 1 2 Phlcbocarya ciliata 23 2 3 Tribonanthes australis Haloragaceae Glischroearyon aureum var aureum Gonocarpus benthamii Gonocarpus dijfusus Gonocarpus paniculatus Gonocarpus simplex Haloragis broumii Haloragodend ron racemosu m Iridaceae Gladiolus undulatus ' Orthrosanthus laxus var l ax us Patersoma babianoides Pater sonia occidentalis Patersonia pygmaea Patersonia umbrosa var urnbrosa Patersonia umbrosa var xanthina Romulea rosea ' Juncaceae Juncus kraussii subsp australiensis Juncus kraussii Juncus pallidus 18 2 33 14 21 26 20 60 3 34 3 3 43 7 13 38 38 35 10 25 6 2 38 15 10 268 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Jiincus pauciflorus 10 Juncus planifolius 2 2 21 Lamiaceae Hemigenia sp (GWJ 4119) 2 2 8 3 8 Hemigenia sp (GWJ 4517) Hem igett tarn icrophylla 8 2 Lauraceae Cassytha capillars Cassytha mkrantha 16 29 23 10 3 36 7 8 2 1 Cassytha racernosa h 47 29 26 30 21 55 50 13 15 55 54 16 Lentibulariaceae Utricularia meniicsii Utncularia multifida Utricularia simplex 7 2 9 21 13 Utricularia tmella 9 29 31 Utricularia violacea 7 Lindsaeaceae Lindsaea linearis 58 36 75 15 35 49 Lobeliaceae Isotoma hypocrateriformis 2 2 10 7 7 8 2 Lobelia alata 80 7 14 13 Lobelia gibbosa 2 2 10 3 Lobelia heterophylla Lobelia rarifolia 7 2 10 41 2 lobelia rhombifolia 3 Lobelia tenuinr 3 50 79 Loganiaceae Logania aff serpyllifolia (GWJ 2743) 2 13 8 4 9 Logania campanulata 8 5 Logania serpyllifolia 20 11 30 28 7 15 23 63 1 Logania vaginalis 10 21 3 1 Mitrasacme paradoxa 29 8 7 23 1 Loranthaceae Nuytsia flori bun da 8 11 7 3 Menyanthaceae Villarsia lasiosperma 10 Villarsia pamassifolia 70 43 8 Mimosaceae Acacia aff pentadenia (GWJ 3700) 5 8 7 31 Acacia alata 2 3 2 1 Acacia biflora 2 2 Acacia broumiana 21 7 38 26 Acacia cochlea ns 21 8 Acacia crispula 2 14 23 Acacia divergens 5 7 15 10 21 32 17 2 Acacia extensa 2 2 8 18 Acacia hastulata Acacia littorea Acacia luteola 5 6 50 55 6 Acacia myrtifulia 33 19 9 14 50 8 13 28 1 Acacia nervosa 15 Acacia pentadenia 11 7 13 15 77 31 97 Acacia pulchella Acacia scapeltiformis 14 6 40 17 29 25 54 1 14 Acacia stenoptera 2 36 54 2 Acacia sidcata 15 Acacia triptycha Acacia uligmosa Acacia urophylla 27 15 8 9 Acacia varia var varia ms 8 5 Acacia mUdenowana Paraserian thes lophan tha 8 2 Myoporaceae My open i m opposit ifol iu m 20 Myopontm tel randru m 10 7 Myrtaceae Agonis sp (GWJ 2113) Agon is sp (GWJ 4768) 2 9 Agon is sp (GWJ 1953) Agon is aff jumperina (GWJ 3628) 3 18 Agon is flexuosa 22 4 70 83 9 7 25 13 15 Agon i s hypericifi ilia Agon is juniper in a 69 14 47 10 18 38 4 75 1 Agonis linear if olia 4 64 43 38 31 6 2 Agonis marginata 13 Agonis parviceps 88 86 98 73 14 13 23 42 68 As tar tea fascic 1 1 laris 3 14 15 10 55 50 13 31 Baeckea aff crispiflora (GWJ 4827) Baeckea aff preissii (GWJ 4804) 7 8 Baeckea astarteoides Baeckea camphorosmae 2 9 14 23 269 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Beaufortia decussata 31 21 49 9 Beaufort ia sparsa Calytrix acutifolia 5 21 15 3 82 2 Calytrix asperula Callistemon speciosus Calathamn us gracilis 2 4 18 7 8 Calothamnus lateralis 8 9 21 8 Chamdaucmm sp (GWJ 3784) Conotha mn us neglect us Danvinia citriodora 2 13 8 Danoinia oederoides 3 4 23 Darwinia vestita Eucalyptus angulosa 11 4 10 7 31 Eucalyptus an ceps E ucal ypt us brevis tyl is 8 25 7 Eucalyptus calcicola Eucalyptus calophylla 17 53 7 7 38 46 85 100 18 Eucalyptus comuta Eucalyptus calophylla x ficifolia 13 8 20 7 25 1 2 Eucalyptus missilis Eucalyptus decipiens 2 7 23 Eucalyptus diversicolor 4 13 55 3 75 Eucalyptus doratoxylon 2 Eucalyptus ficifolia 30 9 18 1 Eucalyptus guilfoylei 8 13 51 11 48 Eucalyptus jacksonii 4 4 47 Eucalyptus rnarginata 77 79 89 18 14 13 38 46 100 1 Eucalyptus megacarpa Eucalyptus occidentalis 6 7 4 20 7 13 8 3 Eucalyptus patens Eucalyptus rudis 6 14 2 18 36 13 23 1 1 Eucalyptus staeri Eucalyptus Virginia ms Eucalyptus wandoo Hamalospermum firmum 16 2 100 23 6 Hypoca lym ma an gust ifoli it m 2 8 7 13 46 9 1 1 ypocalynmia cordifolium Uypocalymnia strictum 50 21 2 9 14 Kunzea aff micrantha 7 8 Kunzea ericifolia Kunzea recurva 17 71 4 20 17 9 7 Kunzea sulphurea Melaleuca acerosa 19 7 6 10 17 18 7 8 Melaleuca aff polygaloides (GWJ 269) 7 Melaleuca baxteri Melaleuca cuticularis 2 10 43 13 Melaleuca densa Melaleu ca d wsm ifolia 2 7 79 13 Melaleuca incana 7 9 Melaleuca niicrophylla 9 13 4 Melaleuca pauciflora Melaleuca polygaloides 3 9 21 Melaleuca preissiana 2 8 36 14 8 Melaleuca rhaphiophylla 14 25 8 Melaleuca scabra 8 Melaleuca spathulata Melaleuca thymoides 83 8 21 18 14 8 2 Melaleuca violacea Pericalymmp crassipes Pericalymma ellipticum Rinzia scholler folia Scholtzia sp tom Verticordia sp (GWJ 4830) Verticordia habrantha Verticordia plumosa Olacaceae Olax benthamiana Olax phyllanthi Orchidac eae Burnet tia /arrest ii Bumettia nigricans Caladema aphylla Caladema corynephora Caladema flava Caladenia huegelii Caladema inter jacens Caladenia latifolia Caladenia longiclavata var longiclavata Caladema rnarginata Caladenia pectinata Caladema natia 15 7 29 23 13 31 2 11 7 14 60 21 10 7 28 13 13 13 23 15 13 1 17 3 3 2 2 2 11 2 270 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Caladcma sericea 2 Caladenia sp 1 Conflicts nbditus 2 9 8 5 Corybas recurvus 14 6 7 Ctyptostylis ovata Cyrtostyhs huegelri 2 2 10 10 13 2 3 Cyrtostylis robust a Diuris p ana flora 20 3 13 3 Diuris lands 13 Diuris Ion gif alia Drakaea elastica 3 14 10 7 1 3 Drakaea glyptodon 2 7 Eh/tfmmthera brio uni is 2 4 10 14 9 7 13 38 1 Elythranthera emarginata 7 3 3 Eriochilus dilatatus EriiKitilus pukhellus 3 11 34 25 6 3 1 Eriochilus sea her 7 15 1 Gastrodia lacista 3 10 4 Lcporella fimbriata Lap toe mis ntenziesii Lyperanthus serratus Microtis alba 5 4 8 5 1 Microtis media subsp media 2 7 Microtis sp Microtis unifolia 4 10 21 25 8 2 2 Monadenia bracteata * 4 13 15 Paracaleana mgnta Prasophyllutn braum i i Prasophyll um drummond ii 3 7 9 7 7 5 Prasophyllutn datum 7 Pmsophyllum fimbria 21 4 Prase iphyllum g ibbo.su m 7 8 Prasophyllutn hians 7 Prasophyllutn odoratum Prasophyll u m pa rvifoli u m Prasophyll u tn regi u m 21 2 10 7 8 Pterostylis aff mm Pterostylis barbata 2 2 13 2 2 Pterostylis narta 7 25 8 11 6 1 Pterostylis plutnosa 8 2 Pterostylis recurva Pterostylis scabra 2 3 13 1 9 Pterostylis vittata var vittata 3 2 7 9 13 8 10 11 2 Thdymitra aiitennifera 2 25 31 1 Thelymitra ben thamiana 1 2 Thdymitra crinita 5 7 38 3 14 Thdymitra nuda 14 6 15 4 11 Thelym it ra pa uciflora 6 14 4 20 7 9 8 1 9 Thdymitra sp 2 2 Thdymitra sp 3 2 1 3 Orobanchaceae Orobanche minor ' 3 2 Oxalidaceae Oxalis purpurea ' Oxalis corniculata 3 13 8 Papilionaceae Aotus getiist aides 2 14 4 27 7 Actus intermedia Bossiaea disticha 27 1 3 Bossiaea eriocarpa 10 11 Bossiaea linnphylla 13 4 30 38 13 11 35 Bossiaea ornata 2 23 4 51 Bossiaea rufa 63 14 15 3 9 8 1 9 Bossiaea tvebbii 3 32 24 9 8 Brachysema minor 7 2 Brachysema sericeum 2 36 Burtonia conferta 55 71 23 10 18 7 8 14 Burtonia scabra 9 Burtonia villosa Chorizema aciculare 5 8 Chorizema aff aciculare (GWJ 2579) 2 2 3 9 Chorizema diver sifol i u m 2 2 10 28 6 Chorizema ilicifol in m Chorizema nanum 4 21 9 Chorizema ret r or sum 2 58 22 61 Chorizema rhombeum 2 8 6 5 Daviesia aff incrassata (GWJ 4526) 2 Daviesia cordata 23 1 12 Daviesia decurrens 44 4 9 8 2 Daviesia horrida 38 3 Daviesia incrassata 2 6 6 271 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 Daviesia oppositifolia Daviesia polyphylla 8 2 Daviesia preissii 8 6 15 14 Dillwynia sp A E u ch i lops is linea ris Eutaxia densifolia 2 8 38 6 6 Eutaxia obovata 30 7 38 3 3 Eutaxia parvifolia Eutaxia virgata Gastrolobium bilobum Gastrolobium brownii 2 2 14 25 15 4 5 Gastrolobium forrestii Gompholobi um aris tatu m 2 2 7 3 2 Gompholobium burton ioides 3 7 Gompholobium capitation 34 9 2 Gompholobiu m kn ightian um 6 4 8 12 Gompholobium ovatum 3 17 15 1 58 Gompholobium polymorphism 2 7 13 38 3 26 Gompholobium preissii 2 Gompholobium tomentosum 2 2 10 21 1 3 Gompholobium venustum Hardenbergia comptoniana 6 20 24 13 15 1 2 Hovea chorizemifolia 5 28 23 7 71 Hovea elliptica 5 15 69 69 Hovea trisperma Isotropis cuneifolia 2 11 55 7 8 5 Jacksonia aff furcellata (GWJ 1411) 16 2 10 48 9 Jackson ia furcellata 11 10 Jacksonia horrida 2 Jacksonia spinosa 2 Kennedia coccinea 8 19 7 14 29 Ken tied ia m icrophylla Latrobea genistoides 14 8 Latrobea tenella var tenella 2 2 Lotus suaveolens ' 2 2 Lotus uliginosus ' 2 15 Mirbelia dilatata 13 8 9 Nemcia crcnulata 23 Oxylobium lanceolatum Phyllota bar bat a Pultenaea aff obcordata 6 20 3 9 7 38 8 Pultenaea barbata Pultenaea ericifolia 2 15 2 Pultenaea reticulata 77 79 25 30 55 6 2 Pultenaea strobilifera 2 Sphaerolobium alatum 2 8 1 14 Sphaerolobium grandiflorum 3 14 8 7 8 1 Sphaerolobium macranthum 3 21 20 55 21 13 23 2 S pha erolob ium m ed ium 5 19 7 8 14 62 Sphaerolobium nudiflorum 2 S phaerolobium racem ulos u m Sphaerolobium vimineum 2 20 9 14 8 2 Templetonia retusa Trifolium dubium * 2 - 3 8 Viminaria juncea 2 43 13 15 El Philydraceae Phitydrella pygmaea Phormiaceae Dianella revaluta Stypandra glauca Pittosporaceae B ilia rd i era cuerul eo-p unctata Billardiera floribunda Billardiera variifolia Sollya heterophylla Poaceae Agrostis avenacea Air a caryophyllea ' Amphipogon amphipogonoides Amphipogon avenaceus Amphipogon debilis Amphipogon laguroides Amphipogon turbinatus Briza maxima ' Briza minor * Rromus hordeaceus ' Danthonia caespitosa Dichelachne crinita Echinopogon ovatus Festuca littoralis Grass sp 1 21 2 23 2 2 2 49 2 4 11 2 2 4 4 2 4 10 10 17 7 14 14 7 14 25 25 13 18 10 10 31 13 13 38 46 15 23 8 38 46 8 8 8 15 46 1 14 87 1 11 5 26 71 8 28 6 19 83 272 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Grass sp 2 Holcus lanatus ' 2 4 9 7 8 1 2 Lolium perenne' 2 2 7 Microlaena stipoides Neurachne alopecuroidea 2 17 13 8 6 2 1 Pou drummondiana 31 7 1 Poa poiformis Poa porphyroclados 2 30 28 7 13 3 2 Poa serpen turn Poa sp 10 7 13 Potypogon monspeliensis ' 13 Stipn cornpressa 10 13 2 Stipa fldPescens 28 Stipa temdfolia Stipa scmibarbata 2 3 38 2 Stipa sp 1 13 8 Tetrarrhena laevis 8 3 7 38 23 89 38 85 Vulpia bromoides * 2 Vulpia myuros ' 2 Podocarpaceae Podocarpus drouyn ianus 27 11 9 58 77 6 Polygalaceae Comesperma ca lymega 36 43 6 10 7 8 9 Comespertna confertum 33 14 38 20 31 55 43 8 35 60 8 Comesperma flavurn 3 21 6 3 9 Comesperma volubile 4 7 23 3 9 Polygonaceae M ueh Jen beckia a dpressa Rumex acetosella * 40 24 2 Rumex pulcher subsp pulcher ' 13 Portulacac eae Calandrinia brevipedata Calandrinia calyptrata 48 9 Primulaceae Anagallis arvensis var arvensis * Anagallis arvensis var caerulea ' 10 3 25 8 Samotus juncetts 20 10 7 Samolus repens 10 3 Proteaccae Adenanthos cuneatus 33 2 20 9 Adenanthvs obomlus 83 57 58 40 3 73 7 5 Banksia attenuate 36 10 Banksia coecinea 2 Banksia gardneri var brevidentata 2 2 2 Banksia gardneri var gardneri Banksia goodii 2 2 Banksia grandis 19 36 31 9 7 38 68 1 Banksia ilicifolia Banksia littoral is 33 8 40 7 18 14 3 2 Banksia occidentals subsp occidentals 9 7 Banksia qu&cifolia 36 9 20 45 7 Banksia seminuda Banksia sphaerocarpa Banksia verlidllata 2 7 13 1 2 Cortospemtum caeruleum 27 9 12 Conosperm um capita tu m 5 2 3 Conospermum flexuosum 2 1 3 Dnjandra armata 2 23 3 Dryandra fomtosa 62 4 9 Dn/andra nivea 8 14 Diyandra serra Dn/andra sessilis Franklandia jucifalia Grevilhi acerosa 2 28 8 5 Grevillea aff martgtesioides Grevillea bronwynae 7 2 Grevillea brownii 2 23 6 Grevillea cirsufolia Grevillea diversifolia subsp subterseri 2 21 15 Grevillea fuscolulea Grevillea occidentalis 6 8 3 26 Grevillea pulchella 8 11 28 Grevillea quern folia 4 14 3 11 Grevillea trifid a 2 8 15 28 Hakea a mplexica u l is 2 21 15 35 66 Hakea ceratophylla Hakea corymbosa 6 7 36 10 29 8 Hakea falcata 9 21 15 11 Hakea falcata short leaved form 15 25 2 Hakea florida 2 8 1 17 273 Journal of the Royal Society of Western Australia, 79(4), December 1996 Al A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Hakea lasianthoides Hakea lasiantha Hakea linearis Hakea lissocarpha Hakea oleifolia Hakea prostrata Hakea mscifolia Hakea sp 1 Hakea trifurcata Hakea undulata Hakea varia Isopogon attenuatus Isopogon axillaris Isopogon formosus Isopogon latifolius Isopogon sphaerocephalus Isopogon teretifolius Lambertia echinata Lambertia uniflora Persoonia aff longifolia Persoonia elliptica Persoonia longifolia Persoonia microcardia Petrophile divaricata Petrophile dwersifolia Petrophile linearis Petrophile longifolia Petrophile squamata subsp A Petrophile squamata subsp B Stirlingia tenuifolia Strangea stenocarpoides Synaphea petiolaris Synaphea polymorpha Synaphea preissii Synaphea reticulata Pteridaceae Pteris vittata Ranunculaceae Clematis pubesems Ranunculus colonorum Restionaceae Alexgeorgea ganopoda Anarthria gracilis Anarthria laevis Anarthria prolifera Anarthria scabra Empodisma gracillimum Hypolaena exsulca Hypolaena ramosissima Leptocarpus sp (GWJ 5249) Leptocarpus aff tephrinus (GWJ 5255) Leptocarpus difftisus Leptocarpus aristatus m Leptocarpus can us 1 Leptoca rp us coa n gust at us Lepyrodia drummondiana * Lepyrodia monoica Restio amblycoleus Restio confer tospi cat us Restio tremulus Restio ustulatus Tar axis glaucescens ms Rhamnaceae Spyridium globulosum Trymalium floribundum Trymalium vemistum Trymalium ledifolium var ledifolium 2 9 3 2 3 2 2 59 71 3 39 2 7 8 97 100 95 100 2 30 2 9 2 83 79 15 2 50 30 45 72 10 10 10 10 40 19 4 2 26 2 4 9 55 10 10 30 10 20 27 18 36 36 9 27 18 21 14 7 21 21 7 7 7 7 43 21 7 7 13 57 14 13 46 13 38 23 15 13 10 15 38 15 15 31 1 70 49 15 8 30 45 3 3 13 17 3 22 2 5 15 3 8 9 11 35 3 2 11 82 2 46 2 2 14 12 2 5 8 58 66 17 Sporodanthus strictus 2 14 Leptocarpus sp 1 3 7 8 Leptocarpus sp 2 6 2 9 7 1 Leptocarpus sp 3 2 2 21 Leptocarpus tenax 39 29 19 55 36 2 Leptocarpus tenellus 14 2 Desmocladus fasciculatus 20 34 14 85 8 69 Loxocarya flexuosa 1 19 7 15 50 97 27 14 25 8 8 31 Lyginia barbata 59 11 10 7 18 7 8 2 Meeboldina denmarkica 2 6 7 Pseudoloxocarya grossa ms 9 Restio sp (GWJ 5235) 9 Restio sp (GWJ 5239) 2 Restio aff tremulus (GWJ 4534) ' 4 9 7 21 38 64 6 274 Journal of the Royal Society of Western Australia, 79(4), December 1996 Rosaceae Acaena echinata var retrorsumpilosa Rubiaceae Opercularia hispidula Opercularia vaginata Ofiercularia vol ubili s Rutaceae Boronia crenulata var crenulata Boronia dcnticulata Boronia gracilipes Boronia jnncea Boronia megastigma Boronia mulloyae Boronia nnnatophylla Boronia virgata Boronia spathulata Boronia stricta Cho rila ena querdfd la Crowca angustifolia var angnstifolia Crozoea angnstifolia var dentata Phebalium anceps Santalaceae Exocarpos odoratus Exocarpos sparteus Lep tamer ia cunn ingha ni i i Leptomeria pauciflora Leptomeria scrob iculata Leptomeria squarrulosa Sapindaceae Dodonaea aptera Dodonaea ceratocarpa Saxifragaceae Ercmosyne pectinata Scrophulariaceae Bellardia trixago ' Gratiola peruviana Parentucellia latifolia * Parentucellia viscnsa ' Veronica distant Veronica plebeta Solanac eae Ant hocerc is s ilvicola Anthocercis viscosa Solatium nigrum ' Stackhousiaceae Stackhousia huegelu Stackhousia monogyna Tripterocnccus brunonis Sterculiaceae Ijis iopetalu m cord if ilium Lasiopetalum floribund um Lasiopelalum floribundum subsp nov. Thomasia foliosa Thomasia pauciflora Thomasia pauciflora var paniculata Thomasia quercifolia Stylidiaceac I jpvenhookia dubia U'venhookia pusilla Stylidium sp (GWJ 3615) Stylidium adnatum Stylidium aff luteum (GWJ 2245) Stylidium aff scan dens Stylidium amnenum Stylidium assimile Stylidium breviscapum Stylidium brunonianum Stylidium caespitosum Stylidium calcaratum Stylidium ciliatum S tyhdiu m divers ifol i u m Stylidiu m etna rgma turn Stylidium fasciculatum Stylidium valioides h Stylidium glaucum subsp angust folium Stylidium glaucum subsp glaucum Stylidium guttatum S tyl idii i m imbri c a turn Stylidium inundatum Stylidium junceum Stylidium laaniatum A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El 13 36 25 70 52 18 7 63 23 76 75 16 2 10 7 10 7 13 25 46 23 36 17 20 3 14 15 3 12 2 2 17 9 7 13 80 43 44 2 21 9 21 2 10 13 2 7 2 10 25 55 18 21 31 1 29 21 2 36 13 1 8 49 2 2 56 22 25 40 21 10 7 14 50 11 18 14 25 8 11 26 1 4 7 5 2 14 1 2 2 2 3 13 3 13 23 7 14 13 8 20 15 10 14 7 3 13 1 3 10 1 5 2 28 3 3 5 2 7 59 8 70 2 8 14 50 3 2 4 52 6 31 2 15 2 2 3 9 7 31 7 11 10 17 5 10 7 2 2 2 5 6 9 8 7 45 18 2 31 2 7 2 2 7 9 14 2 14 9 14 23 1 2 8 8 18 7 2 21 2 9 21 46 7 20 28 8 21 2 3 2 9 13 15 2 9 13 8 50 9 3 2 9 275 Journal of the Royal Society of Western Australia, 79(4), December 1996 A1 A2 A3 A4 B1 B2 Cl C2 C3 D1 D2 El Stylidium lepidum 2 Stylidium luleum 3 9 9 29 8 7 15 Stylidium pilifenim 8 14 25 9 8 3 9 Stylidium pritzelianum 10 12 Stylidium repens 20 43 13 20 9 8 2 Stylidium rhy nchocarpum 41 22 7 Stylidium scandens 8 43 34 3 18 25 8 4 11 Stylidium schoenoides 20 8 9 15 6 Stylidium sp 1 2 Stylidium sp 2 6 7 2 Stylidium sp 3 3 5 Stylidium spathulatum subsp spathulata 2 4 14 31 3 20 Stylidium spathulatum subsp acuminata Stylidium violaceum 17 2 5 Thymelaeaceae Pimelea sp (GWJ 4204) Pimelea angustifolia Pimelea clavata 4 10 14 1 2 4 Pimelea ferruginea Pimelea hispida 13 30 3 14 23 5 Pimelea imbricata 38 31 Pimelea longiflora 69 32 15 3 2 Pimelea rosea 6 40 66 14 8 1 8 Pimelea spectabilis 2 24 28 8 Pimelea suaveolens subsp suaveolens 1 8 Pi melea sy Ives t ris 20 11 11 Tremandraceae Platytheca galioides 3 Tetratheca affinis 2 4 23 38 Tetratheca fitiformis 2 27 Tetratheca hirsuta 11 9 32 43 2 T et ra theca hispid is si ma 2 Tetratheca setigera p 9 6 8 3 Tremandra diffusa 8 38 34 1 Tremandra slelligera 4 13 73 8 72 Violaceae Uybanthus debilissimus 2 20 28 2 Xanthorrhoeaceae Kanthorrhuea gracilis 3 8 14 63 Xanthorrhoca preissii 27 57 75 30 41 27 36 25 77 8 40 Xyridaceae Xyris flexifolia 18 Xyris lanata 2 6 36 Zamiaceae Macrozamia riedlei 13 25 38 8 61 42 16 weed species; ' some specimens subsequently confirmed as Schoenus caespititus; '' specimens at quadrat 1009 subsequently confirmed as Schoenus rnultigluniis ; specimens at quadrat 423b subsequently confirmed as Schoertus tr achy carpus; d some specimens subsequently confirmed as Gahma sp Yelverton (G J. Keighery 10820) and others as Gahnia sp aff insignis ; ■ some specimens subsequently confirmed as Schoenus subflavus ; 1 some specimens subsequently confirmed as Schoenus subtnilTOSlachyus; * specimens at quadrat 1041 subsequently confirmed as Mesmclaena graciliceps; h members of this species group amalgamated in analysis; 1 some specimens subsequently confirmed as Restio crascens; 1 some amalgamation of this taxon in analysis including Desmocladus flexuosus ms, Empodisma gracillimum, Taraxis glaucescens ms; k some specimens subsequently confirmed as Lepyrodia extensa; 1 some specimens subsequently confirmed as Lepyrodia muirii; m some specimens subsequently confirmed as Chaetanthus leptocarpoides ; n subsequently confirmed as Stylidium spathulatum ; p some specimens subsequently confirmed as Tetratheca hispidissima. 276 Journal of the Royal Society of Western Australia, 79:277-280, 1996 Terrestrial invertebrates in south-west Australian forests: the role of relict species and habitats in reserve design Barbara York Main Department of Zoology, University of Western Australia, Nedlands WA 6907 Abstract Invertebrates are an integral and functional component of any ecosystem, including the south¬ west Australian forests. However they are often overlooked in biological studies. This is partly due to the incompleteness of taxonomic knowledge, as much of the invertebrate fauna is un¬ named. Also, while a great deal of distributional data exists for invertebrates, it is generally not readily accessible, often associated only with individually stored museum specimens and/or scattered taxonomic descriptions in the literature. It is proposed here that a pragmatic approach to invertebrate conservation via reservation is to consider the known distribution of selected relict species (because of their sensitivity to disturbance) and their microhabitat requirements, and (a) deduce whether such species/habitats are adequately included in the current reserve system and (b) if not, then predict where in the overall domain they are likely to occur, with a view to ensuring their persistence. A short-cut method is to note the location of Gondwanan-type sites and argue for their preservation on the grounds that they will contain relict species assemblages. This approach assumes that widespread and/or ecologically tolerant species are more likely to have other options in the face of artificial disturbance. Introduction In the catch-phrases of an attention-seeking society, while "big may be beautiful" in reference to the role of invertebrates in biological processes and ecosystems, perhaps the apt phrase is "small and significant". Wil¬ son (1987) convincingly argued for the dependance on invertebrates of higher life forms, and the support base provided by invertebrates in any ecosystem. The decline of some local insectivorous birds is commonly remarked upon. Hence it is ironic that although nature conserva¬ tion is now a well established and even "popular" disci¬ pline with a practical and ethical base, little attention has been given to invertebrates (in their own right) until relatively recently. In Australia, compared to vertebrates and flora, invertebrates have fared poorly with respect to conservation. Amongst the general publications de¬ voted to invertebrates, those of New (1984, 1995) are notable, and there are also studies dealing with particular groups, for example a recent review on arachnid conservation by Yen (1995). Under the blanket concept and ethos of conservation must be considered the practical manifestation: reserva¬ tion of places for invertebrates. How then, if at all, does the practicality of reservation for terrestrial invertebrates differ from that for flora or vertebrates? Or can they be assumed to be safe-guarded along with whatever meth¬ ods are applied in reserve establishment and manage¬ ment for the more obvious biota such as vascular plants and vertebrates? It must be noted that in spite of their significance, invertebrates are frequently overlooked in Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 broad sweep ecological studies and biological surveys because of their small size and often cryptozoic behaviour. Nevertheless there are statutory requirements in Western Australia concerning the conservation of all fauna and dissemination of knowledge related to that fauna. The Strategic Plan 1995 of the Department of Con¬ servation and Land Management (Anon 1995a) states under the major functions of CALM firstly, " Conserva¬ tion of Nature. To conserve the indigenous biota and ecological processes in natural habitats throughout the State " and fourthly " Knowledge . To ensure that .... functions [i.e. concerning conservation] are underpinned by up-to-date and reliable science-based knowledge." These objectives surely partner in intent the Western Australian Museum Act which charges the Western Australian Museum "to make and preserve ... collections representative . of the natural history of the State" (Anon 1969). Furthermore the Science and Information Division of CALM has been developed to meet several inclusive needs; "to document the biota , ecological processes & biological resources of the state", "to conserve threatened species & ecological communities by ameliorating inimical processes ", and "to ensure that land & biological resources are used sustainably" (Anon 1995a). This would suggest that all is well for our natural heritage (including the significant smaller " taxa") and that a beneficent authority is taking good care of our fauna and is about to inform us (if not already having done so) of what we have and will continue to have, within our south-west forests. Invertebrates - a special case However I argue here that if conservation reserves within the region under discussion within the 800mm rainfall isohyet (for instance forested landscapes) are to adequately cater for conservation of invertebrates then 277 Journal of the Royal Society of Western Australia, 79(4), December 1996 the special characteristics and needs of this fauna must be considered. These needs stem primarily from one or more of the following characteristics and phenomena; • the sheer diversity and taxonomic range of inver¬ tebrates. Some idea of this diversity can be gauged from spiders: this arachnid order is represented by at least 50 families in Western Australia of which all but a few occur in the south-west forest and some of which occur only in the region (Main 1985 and unpublished); • the range of foraging methods from vegetative to predatory and parasitic; • the range of life histories and individual longevity ranging from a few days to over 30 years; • the differing dispersal methods, particularly of ju¬ veniles, and mobility of various life history stages; • the range of site fidelity from sedentary to par¬ tially territorial or transitory; • the interaction of life cycle patterns with other biota; • the association with particular physiographic con¬ figurations; and • the antiquity of many forms/taxa. All the above phenomena /characteristics are a conse¬ quence of the historical ecology of the particular taxon especially of the "ancient" groups which has been deter¬ mined by the environmental history of the landscape including geological and climatic changes and changes in the vegetation which relate to earlier continental con¬ figurations and their positions. Because of their small size and often specialised behaviour, particularly of relict forms, many inverte¬ brates (in contrast to most vertebrates) are confined to topographically or geographically restricted areas and specialised microhabitats which may not be readily perceived by broader scale surveys. Such microhabitats are vulnerable to artificial disturbances imposed by agriculture, forestry and other rural and urban disrup¬ tions to the landscape, for instance roads and other human constructions. Current knowledge-base of invertebrates We might digress now and look at our current knowl¬ edge-base relating to invertebrates of the south-west forest of Western Australia by asking the following question. Is there a taxonomic database (either as lists of taxa contained in the Western Australian Museum and/ or other collections or in the published literature), arranged or accessible according to the region in the present context, habitat data, and whether present in the current reserve system? Firstly there is a great deal of information on local distributions of invertebrate taxa for the south-west forest region in the taxonomic literature. Nearly 3000 papers dealing with all aspects of invertebrates of Western Australia are listed in the excellent bibliography compiled over 15 years ago by Majer & Chia (1980). In spite of this, the taxonomic and distributional data needed now are not readily or rapidly accessible and are largely uncoordinated, except for particular groups which specialists have in their personal research files. For example a recent paper by Abbott (1994) discusses the distribution of earthworms in the south-west and shows a spatter of dots on a map within the 400mm rainfall zone. The data were derived from museum col¬ lections and published records but are not readily acces¬ sible. Similarly, many systematists including myself (for spiders) have unpublished species lists from particular parks or localities. Secondly, some associated habitat data accompany stored specimens and taxonomic literature records, but except for a very few groups are either uncoordinated or unpublished. Thirdly, the greater proportion of the invertebrate fauna (and not just insects) is unnamed. Hence biologi¬ cal and distribution data associated with stored speci¬ mens, as well as being not readily accessible, cannot be manipulated in the same way as vertebrate data. Statements such as, " All states and the Commonwealth have a database of threatened and rare species observations .... ", from the Deferred Forests Assessment document (DFA, Anon 1995b, p 5) are largely irrelevant for invertebrates as such databases refer almost exclusively to vertebrates. However, Abbott (1995) has produced a valuable resource in his selective list of insects recorded in the literature from the jarrah and karri forests of southwestern Western Australia; he noted " nearly 1800 insect and closely allied species of ... Hexapoda" from south¬ west forests but estimated that the total number could be 15 000 to 20 000. Suggested guide-lines for reservations In spite of the absence of comprehensive, usable databases of invertebrates, a case study could be made of selected taxa to determine guidelines for reserve selection. This could be based on what knowledge we do have regarding relictual taxa including Gondwanan elements (e.g. spiders and other arachnids, selected insects, amphipods, earthworms, nemerteans, Onycophora etc), which are confined to certain habitats and in some cases geographic areas. Bearing in mind the foregoing points regarding the restricted nature and vulnerability of habitats of relic taxa, the converse is assumed for later evolved /adapted groups and/or more widespread taxa, when such taxa are likely to be more resilient to artificial disturbance and fragmentation of habitats. Definition of relict taxa and habitats Some definition of relict taxa and their associated habitats needs to be made. Generally such taxa are rep¬ resentative of a fauna from a more humid, less mark- edly-seasonal climate associated with a mesophytic for¬ est with closed canopy and as such they are relicts of a pre-fire-prone environment. With progressively dryer and more seasonal climatic conditions, the most favourable habitats have become increasingly fragmented until now such fauna are restricted to specialised microhabitats which simulate on a small scale an earlier more widespread habitat. Relict taxa include extremely ancient representatives of Gondwanan 278 Journal of the Royal Society of Western Australia, 79(4), December 1996 elements and with lineages reaching back to pre- Cretaceous times (120 - 140 million years ago) e.g. the trapdoor spider Moggridgea (see Main 1991) and velvet worms (Onycophora) or to the early Tertiary Eocene/ Oligocene period 40 - 50 mya (Main 1987). Striking austral zoogeographic affinities are found within many spider families, including the Archaeidae and Pararchaeidae (Forster & Platnick 1984; Platnick 1991; Main 1995), Orsolobidae (Forster & Platnick 1985; Griswold & Platnick 1987), which can best be interpreted as relict distributions from a contracted early Tertiary environment. Presence of the Micropholcommatidae and Archaeidae in south-western Western Australia has been explained in this manner (Main 1974, 1995). The distri¬ bution of certain terrestrial insect groups such as thrips of the family Aeolothripidae (Mound 1972) can also be interpreted to be of this era. Many groups of invertebrates from both periods have taxonomic affinites with south-easten Australia, Tasma¬ nia, New Zealand and/or other southern continents. It is outside the scope of this paper to document the many examples which, from perusal of the taxonomic litera¬ ture indicate Gondwanan affinities. Some relationships are listed by Main & Main (1991) and Hopper et al. (1996). The associated habitats of relict species are perma¬ nently moist and shaded, with such conditions provided by high rainfall. However other formative physiographic conditions include topography, proximity to coast and directional orientation. Main & Main (1991), Hopper et al. (1996), and A R Main (1996) have summarised the major characteristics of persistent Gondwanan habitats in south-west Western Australia. In addition B Y Main (1996) has demonstrated the occurrence of relictual Gondwanan microhabitats on the plateau region of southern Western Australia. Location of relict habitat sites The major high rainfall areas on geologically old ter¬ rain include the Walpole/Nornalup topographically high region, the high etched region west of Manjimup and Pemberton and farther north, and isolated regions near Collie, Dwellingup and Jarrahdale. Areas which are wet by virtue of old erosional phenomena combined with orientation include the northern valleys of the Dar¬ ling Scarp such as near Mundaring and Kalamunda, and farther south in the sinuous configurations of rivers, such as the Deep River, which have further significance by having a continuity with the fauna of the Walpole/ Nornalup area. Noteworthy also are some Pleistocene sites (including cave complexes) and areas of the Perth Coastal Plain, as at Jandakot, where relict fauna has encroached from "older" areas. Farther south the limestones of the Mammoth Cave area and Boranup with high rainfall and wet karri forest contain certain relicts. Typical minor or "special" areas are sites associated with granite outcrops which benefit from run-off (some¬ times created simply by dew and light mist), summits of emergent monadnocks and south facing slopes, swampy headwaters of river systems, perched "swamps" and so on. Even within the general forest, regardless of minor topography, there are further secretive "micro-sites" such as; the litter "stacks" or "rings" around the butts of karri and accumulated debris in the buttresses of tingle (Wallis 1992); the persistent spongy (if not recently burnt) bark of red tingle; long unburnt rotten logs; stag- heads of old trees and debris in tree forks (particularly sheoaks) and trunk crevices. As well as high rainfall, the prevalence of fog whether induced by topography or coastal proximity, promotes persistence of wet "microhabitats". Examples are rem¬ nants of the old plateau (including Mt Cooke and Mt Saddleback in the jarrah forest), Mt Lind es ay near Den¬ mark, granite outcrops and south coast peninsulas such as Torndirrup and West Cape Howe. Again, within such sites subtle microhabitats can be further specified. From the foregoing it is apparent that when considering reservations of sites to include invertebrates (whether at the species or community level), different, smaller scale criteria need to be used than is the case for vertebrates. The DFA document (Anon 1995b) emphasises the im¬ portance of using physical environmental data in the absence of adequate community data "to ensure that representativeness is met". This approach is paramount in the selection of reserves encompassing invertebrate preservation. Bench-mark information Contained within the stored collections and literature records is extractable bench-mark information of distributions of taxa prior to continuing disturbance. Notable are the reports of the Michaelsen & Hartmeyer Expedition to south-western Western Australia in 1905 (Michaelsen & Hartmeyer 1907-30). Nearly 90 sites were sampled, of which about 25 were in the forest-zone extending from Jarrahdale to Nornalup and Albany. Within the forest and neighbouring areas, accessibility was both facilitated and constrained by forestry activities ranging from logging and milling to timber transport and loading of ships at Bunbury, Hamelin Bay and Torbay. Ironically it is the contemporary timber industry that is provoking most research and fortuitous gathering of collections and knowledge of much of the invertebrate fauna of the forest. Collection methods in 1905 were both haphazard and less refined than some of the current techniques and the collectors did not have our hindsight of environmental and taxonomic knowledge to be able to predict where relict (Gondwanan) species would be likely to occur. Nevertheless it is notable that this expedition was staged on the premise that the south-west peninsula of Australia would be the biogeographic complement of the southern areas of South America and South Africa which regions the German team had already sampled. Later taxonomic studies confirm this premise. The relevance of the bench-mark collections, in com¬ bination with later records, to the selection of reserves must not be ignored. Summary I suggest that by selecting named relict taxa, noting their microhabitat requirements, predicting where such 279 Journal of the Royal Society of Western Australia, 79(4), December 1996 microhabitats are likely to occur in the forest, and com¬ bining the above with recorded locality data for such relict taxa, then it should be possible to deduce whether all known relict taxa and their full geographic ranges are represented within the present reserve system, and conversely whether all "relict" sites and concomitant, potential unnamed species are catered for in the present reserve system. Obviously we don't have the luxury of time to ferret out even the recorded information of species, habitats and distributions to impose these data as a template on the geographic area. However, this needs to be done, taxon by taxon by various specialists in a co-ordinated manner. The practical short-cut suggested here places priority on physiographic features likely to contain any of the microhabitat categories described above and as delin¬ eated by Main &: Main (1991), A R Main (1996), B Y Main (1996) and Hopper et al. (1996), and indeed where our major biotrove lies. I suggest that this approach is an ideal guide for selection of reserves containing micro¬ habitats appropriate to relict, geographically restricted and specialised invertebrates. The more widespread and/or ecologically tolerant species have a wider range of natural options and are also better able to cope with artificially disrupted habitats. References Abbott 1 1994 Distribution of the native earthworm fauna of Aus¬ tralia: a continent-wide perspective. Australian Journal of Soil Research 32:117-126. Abbott I 1995 Prodomus of the occurrence and distribution of insect species in the forested part of south-west Western Aus¬ tralia. Calm Science 1 (4):365-464. Anon. 1969 Western Australian Museum Act. Western Austra¬ lian Government, Perth. Anon. 1995a Strategic Plan 1995, Science and Information Divi¬ sion. Department of Conservation and Land Management, Perth. Anon. 1995b Deferred Forest Assessments. Commonwealth Government, Canberra. Forster R R & Platnick N 1 1984 A review of the archaeid spiders and their relatives, with notes on the limits of the superfamily Palpimanoidea (Arachnida, Araneae). Bulletin of the American Museum of Natural History 178:1-106. Forster R R & Platnick N I 1985 A review of the austral spider family Orsolobidae (Arachnida, Araneae), with notes on the superfamily Dysderoidea. Bulletin of the American Museum of Natural History 181:1-230. Griswold C E & Platnick N I 1987 On the first African spiders of the family Orsolobidae (Araneae, Dysderoidea). American Museum Novitates 2892:1-14. Hopper S D, Harvey M S, Chappill J A , Main A R & Main B Y 1996 The Western Australian biota as Gondwanan heritage - a review. In: Gondwanan Heritage : Past, Present and Future of the Western Australian Biota (eds S D Hopper, J A Chappill, M S Harvey & A S George). Surrey Beatty and Sons , Chipping Norton, 1-46. Majer J D & Chia J 1980 An inventory of information on terres¬ trial invertebrates occurring in Western Australia. Depart¬ ment of Biology Bulletin 1, Western Australian Institute of Technology, Perth. Main A R 1996 Forest reservations: an overview'. Journal of the Royal Society of Western Australia 79:301-304. Main B Y 1974 Occurrence of the lungless spider Micropholcomma Crosby and Bishop in south-west Western Australia (Araneae: Symphytognathidae). Journal of the Australian entomological Society 13: 79. Main B Y 1985 Richness of spiders (Araneae) in south-western Australia. In: Soil and Litter Invertebrates of Australian Medi¬ terranean - Type Ecosystems (eds P Greenslade & J D Majer). School of Biology Bulletin 12. Western Australian Institute of Technology, Perth, 10-11. Main B Y 1987 Ecological disturbance and conservation of spi¬ ders: implications for biogeographic relics in southwestern Australia. In: The Role of Invertebrates in Conservation and Biological Survey (ed J D Majer). Department of Conservation and Land Management Report, Perth, 89-97. Main B Y 1991 Occurrence of the trapdoor spider genus Moggridgea in Australia with descriptions of two new species (Araneae: Mygalomorphae: Migidae). Journal of Natural His¬ tory 25:383-397. Main B Y 1995 Additional records of the Gondwanan spider Austrardmea from southwestern Australia. The Western Aus¬ tralian Naturalist 20:151-154. Main B Y 1996 Microcosmic biogeography: trapdoor spiders in a time warp at Durokoppin. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (eds S D Hopper, J A Chappill, M S Harvey & AS George). Surrey Beatty, Chipping Norton, 163-171. Main A R & Main B Y 1991 Report on the southern forest region of Western Australia. Report to the Australian Heritage Commission, Canberra. Michaelsen W and Hartmeyer R 1907 - 1930 Die Fauna Sud-west Australiens. Vols. 1 - 5. Fischer, Jena. Mound L A 1972 Further studies on Australian Aeolothripidae (Thvsanoptera). Journal of the Australian entomological Soci¬ ety 11:37-54. New T R 1984 Insect Conservation - An Australian Perspective. W Junk, Dordrecht. New T R 1995 An Introduction to Invertebrate Conservation. Ox¬ ford University Press, Oxford. Platnick N I 1991 On Western Australian Austrardmea (Araneae: Archaeidae). Bulletin of the British arachnological Society 8:259-261. Wallis N W 1992 Variation in the leaf litter and associated inver¬ tebrates around the bases of red tingle (Eucalyptus jacksonii ) and karri (Eucalyptus diversicolor) trees. Honours Thesis. Curtin University, Perth. Wilson E O 1987 The little things that run the world. Conserva¬ tion Biology 1:344-346. Yen A 1995 Australian spiders: An opportunity for conservation. Records of the Western Australian Museum, Supplement 52:39-47. 280 Journal of the Royal Society of Western Australia, 79:281-291, 1996 Aquatic fauna of the Warren bioregion, south-west Western Australia: Does reservation guarantee preservation? K M Trayler1, J A Davis1, P Horwitz2 & D Morgan1 1 School of Biological and Environmental Sciences, Murdoch University, Murdoch WA 6150; 2 Centre for Ecosystem Management, Edith Cowan University, Joondalup WA 6027 Abstract The Warren Bioregion, in the extreme south-west of Western Australia, has a unique assem¬ blage of aquatic invertebrates, fish and amphibians. Current literature indicates that 192 fully described species have been collected, of which 10 invertebrate, 1 fish and 6 frog species could be considered locally endemic. We estimate that secure nature reserves (A-Class and National Parks) in the Warren Bioregion provide a refuge for 86% of the aquatic faunal elements. Reservation alone, however, may not be sufficient to protect certain of the aquatic fauna. Adverse impacts occurring within catchments, including erosion and deposition of sediment, salinization, fire, land clearing, the presence of dams and the introduction of exotic fish, may adversely affect the aquatic fauna within a reserve. Management of protected habitats must ensure that only anthro¬ pogenic activities which are sympathetic to the long term persistence of all elements of the biota occur within, and adjacent to, the reserve systems. Introduction The natural landscape of the south-western corner of Australia is characterised by rivers which arise on an ancient and flat semi-arid inland plateau. From here, the rivers flow sluggishly towards the coast, before passing through a zone where the topography steepens and rainfall increases (Mulcahv et al. 1972; Churchward et al. 1988). Beyond this, the rivers again slow as they traverse the coastal lowlands, which feature extensive wetland systems and terminate in lagoon-like estuaries. Early settlers found the coastal lowlands to be largely infertile soils and used much of the area as pastoral land, preferring to use the rich alluvial soils along the rivers for intensive agriculture (Jarvis 1979). Pressures of urbanization and intensification of agriculture, associated with a burgeoning population in the century since colonization, have led to the loss of a large proportion of coastal wetland systems and many of those that remain have been altered from their natural state (see Halse 1989). A number of the larger rivers have been damaged through siltation and salinization, and many of the remainder are now impounded or subject to varied degrees of alteration through activities such as mining and logging (Olsen & Skitmore 1991). The effect of these changes on the aquatic fauna cannot be fully assessed due to the lack of historical or baseline data. Locally or even regionally endemic aquatic species may have been lost from south-western Australia, or suffered substantial range reductions, particularly where these fauna were sensitive to changed hydrological regimes, increased eutrophication and/or salinization. This is evident in our aquatic environments today. Highly eutrophic wetlands contain few rare taxa. Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 have high densities of "pollution tolerant" species and, overall, support an invertebrate composition which is distinct from that occurring in low to moderately en¬ riched or coloured wetlands (Davis et al. 1993; Edward et al. 1994). The highly saline Hotham River and Thirty- Four Mile Brook, south-east of Perth, are dominated by salt tolerant crustaceans (60% total abundance), while insects contribute only 14% to the overall invertebrate composition (Bunn & Davies 1992). This situation is atypical of undisturbed streams of the jarrah forest, where insects usually comprise 70-80% of the fauna (Bunn et al. 1986), and was attributed to increased salinity (Bunn & Davies 1992). Williams et al. (1991) found little or no longitudinality in the faunal composition of the Blackwood River despite the presence of a distinct salinity gradient along its length. Although this could be interpreted as evidence of a halotolerant fauna, it may also indicate elimination of less tolerant fauna, which once characterised a more diverse system. Public awareness of the need to conserve aquatic habitats has increased markedly over the last decade and this is reflected in the number of scientific studies that have been undertaken on wetlands, streams and rivers in the south-west since the mid 1980s (see Balia 1994). Considerable attention has been devoted to the aquatic fauna and their habitats in highly populated areas of the south-west, such as wetlands on the Swan Coastal Plain and streams in the adjacent jarrah forest, where monitoring to assess the impact of pollution or habitat change has often been the principal research or management objective (e.g. Storey et al. 1990; Bunn & Davies 1992; Crowns et al. 1992). Further south, the relatively undisturbed aquatic habitats within the Warren Bioregion (sensu Thackway & Cresswell 1995) have also provided a focus for scientific studies (e.g. Christensen 1982; Pusey & Edward 1990a, b; Horwitz 1996; Edward et al. 1994). This region is regarded as an important centre for endemism from a 281 Journal of the Royal Society of Western Australia, 79(4), December 1996 botanical viewpoint (Hopper et al. 1992), but its signifi¬ cance to the aquatic fauna has not been assessed in its entirety. This paper summarizes existing knowledge of the aquatic fauna in the Warren Bioregion and estimates the degree of endemicity found in the region. In addi¬ tion, it presents a preliminary assessment of the reserva¬ tion status of known aquatic species and makes recom¬ mendations for the ongoing protection of fauna within the reserve system. Approaches The aquatic invertebrates (>110 pm), fishes and amphibians which occur within the Warren Bioregion were compiled from published literature and museum records, and are therefore limited by the habitats examined and sampling protocol of these sources. The distribution of each species, determined from published literature, is described as either widespread, regionally endemic or locally endemic in accordance with Horwitz (1996). The reservation status of each species within the Warren Bioregion has been determined according to their presence or absence within reserves which fall under categories I and II of the World Conservation Union's classification of protected areas (Anon. 1994). This equates only to nature reserves (A-Class) and national parks which have the conservation of flora and fauna as a primary objective and require legislative provisions to alter their boundaries or status. Other forms of reserves in the region do not have this level of security and therefore have not been included in this study. We do, however, acknowledge that these other areas also play an important role in conservation. For the purpose of this study, a species which is described as reserved must have been recorded at least once in a secure reserve, as defined above. To date, there have been no comprehensive surveys of aquatic invertebrates within all nature reserves and national parks of the Warren Bioregion and therefore the reference in Table 1 to species as absent from these reserve categories should be considered preliminary. In addition, species which are widespread or regionally endemic may not be found in nature reserves or national parks within this region, but may be found within similar reserves located elsewhere. Although more invertebrate taxa are found in the Warren Bioregion than have been recorded in Table 1, many of these have not yet been described and specimens only exist in voucher or reference collections of individual researchers. To prevent confusion, whereby authors may have assigned different voucher names to the same taxon, only species for which published descriptions exist were considered here. Emergent adult invertebrates found in the vicinity of aquatic habitats were excluded because they could not be definitively linked with specific waterbodies. In addition, waterbirds have been excluded from this review, but information relating to their occurrence and reservation within the larger southern forest region may be found in Christensen (1992). The tabulated information on aquatic invertebrates, fishes and amphibians provides a preliminary account of the level of endemism of aquatic fauna in the Warren Bioregion and emphasises the importance of the reserve system to both the restricted and more common fauna of the region. Local endemicity in the aquatic fauna of the Warren Bioregion Invertebrates Aquatic habitats within the Warren Bioregion support a large proportion of invertebrate taxa which are endemic to south-west Western Australia, with approximately 17% of these considered locally restricted on the basis of surveys (Table 1). This number, however, may be an underestimate as only fully-described species have been considered, thereby excluding some taxa. There are, for example, only eight oligochaete species with widespread distributions shown in Table 1, yet there are at least a further four species which are restricted to the Warren Bioregion (Horwitz 1996) for which descriptions are pending (A Pinder pers. comm.). This situation may also extend to other groups, notably the dipterans and arachnids. Locally endemic invertebrates include the freshwater crayfish Engaewa subcoerulea and E. similis and the trichopteran Kosrheithrus boorarus (see Horwitz 1994; Crowns & Davis 1994, respectively). A rich suite of micro-crustaceans have been found in the area and, in the case of the copepods, this is highly distinctive (Table 1). Bayly (1992) examined temporary ponds near Northcliffe and identified several copepods which are locally endemic, including Calamoecia elongata, Boeckella geniculata and Paracyclops sp nov, as well as a unique form of C. tasmanica si. In addition, the copepod Hemiboeckella powellensis is known only from Lake Powell, between Albany and Denmark (Bayly, 1979). Locally endemic cladocerans, Biaptera imitaria and Daphnia occidentalis were also found near Northcliffe by Bayly (1992). Daphnia occidentalis is a relictual species thought to have originated 70 mybp (Benzie 1986, 1988). This species was found in large numbers in a single pool, surrounded by swamp land within the D'Entrecasteaux National Park. Fish Eight species of freshwater fish are endemic to the south-west of Western Australia (Table 2) and four of these are now principally confined to the Warren Bioregion (Morgan et al. 1996). These include the Western Australian salamanderfish Lepidogalaxias salamandroides , a relictual Gondwanaland species (Allen 1982), which is now predominantly restricted to ephemeral pools within the coastal peat flats between Windy Harbour and Walpole (Morgan et al. 1996). The black stripe minnow Galaxiella nigrostriata and Balston's pygmy perch Nannatherina balstoni are similarly con¬ fined, although the former species is found in low num¬ bers in a few lakes of the region, while the latter occurs in low numbers in both lakes and rivers of the region (Morgan et al. 1996). The mud minnow Galaxiella munda has the widest distribution of these four species, occur¬ ring in the headwaters and tributaries of rivers, as well as on the coastal peat flats (Morgan et al. 1996). Disjunct and isolated populations of some of these 282 Journal of the Royal Society of Western Australia, 79(4), December 1996 Table 1 Invertebrate fauna found in aquatic habitats within the Warren bioregion. Reservation status refers to their presence or absence in a nature reserve or national park within the bioregion. Distribution: W, widespread; RE, regionally endemic; LE, locally endemic. Sources; 1, Pusey & Edward (1990a); 2, Growns & Davis (1994); 3, Edward ct al (1994); 4, Bayly (1982); 5, Horwitz (1994); 6, Williams et al.( 1991); 7, Bayly (1992); 8, Bayly (1979); 9 Harvey (1996); 10, Morton (1990). Species1 Source Lotic Lentic Reservation Distribution MOLLUSCA Hydriidae Westralunio carteri Iredale 2,3 Glacidorbidae Glacidorbis occidentals Bunn & Stoddart 1 Ancylidae Ferrissia petterdi Johnston 1 Hydrobiidae Potamopyrgus niger Quoy & Gaimard 6 ANNELIDA Phreodrilidae Insulodrilus lacustris Benham 5 Insulodrilus nudus Brinkhurst & Fulton 5 Tubificidae Limnodrilus hoffmeisteri Cleperede 5 Naididae Pristina longiseta Ehrenberg 5 Pristina aequiseta Bourne 5 Derofurcatus Muller 5 Dero digitata Muller 5 Chaetogaster diastrophus Gruithuisen 5 ARACHN1DA Hydryphantidae Pseudohydryphantes doegi Harvey 5 Aturidae Wheenyoides cooki Harvey 5 Limnocharidae Linmochares australica Lundblad 5 Pionidae Australotiphys barmutai Harvey 9 Larri laffa Harvey 9 Fiona cumberlandensis Rainbow 9 Arrenuridae Arrenurus sp nr lasmanicus Lundblad 5 OSTRACODA Limnocytheridae Lininocy there mowbrayensis Chapman 1,3 Gomphodella aff rnaia De Deckker 3 Cyprididae Candonocypris novaezelandiae Baird 2,3 Neivnhamia fenestra King 3 Cypretta viridis Thomson 7 Cypretta baylyi McKenzie 3,4,7 Eucypris virens Jurine 7 Kapcypridopsus asymmetra De Deckker 4 Sarcypridopsis aculeata Costa 3 Alboa wooroa De Deckker 3 Ilyodromus candonites De Decker 4 llyodromus varrovillus King Bennelongia australis Brady 8 Candonidae Cando7iopsis tettuis Brady 3 COPEPODA Centropagidae Calamoecia attenuate Fairbridge 1,3,7 Calamoecia tastnanica Smith si 1,3,7 Calamoecia elongata Bayly 7 Hemiboeckella searh Sars 3,7 Hemiboeckella andersonae Bayly 3,7 Hemiboeckella powellensis Bayly 8 Boeckella geniculata Bayly 7 V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V RE RE w w + w + w + w + w - w - w + w - w + RE + RE + W + RE + RE + W + RE + W + W + w + w + w + w + w + LE + W + W + LE + W + W + RE + RE + W + LE + W + RE + LE + LE 283 Journal of the Royal Society of Western Australia, 79(4), December 1996 Species Source Gladioferens imparipes Thomson 3 Cyclopidae Macrocyclops albidus Jurine 1,7 Australocy clops australis Sars 7 Eucyclops spatulatus Morton 10 CLADOCERA Chydoridae Biapertura cf macrocopa Sars 4,7 Biapertura rigidicaudis Smirnov 4 Biapertura affinis Leydig 1 Biapertura longinqua Smirnov 7 Biapertura nr setigera Brehm 3 Biapertura imitatoria Smirnov 7 Graptolebersis testudinaria Fischer 3 Camptocercus cf australis Sars 3 Chydorus barroisi Richard 4 Chydorus cf sphaericus O F Muller 7 Pleuroxus inemiis Sars 7 Pleuroxus jugosus Henry 4 Alonella cf excisa Fischer 7 Rak obustus Smirnov & Timms 8 Monope reticula Henry 7 Macro thricidae Neothrix arrnata Gurney 1,3, 4,7 Daphniidae Simocephalus acutirostratus King 1,3 Daphnia carinata King si 3 Daphnia occidentalis Benzie 7 Scapholeberis kingi Sars 7 Bosminidae Bostnina meridionalis Sars 3 ISOPODA Amphisopidae Amphisopus annectans Nicholls 1,5 Amphisopus llintoni Nicholls 1 Hyperodesipus ?plurnosus Nicholls & Milner 2 AMPHIPODA Perthidae Perthia acutitelson Straskraba 1,2 Perthia branchialis Nicholls 3 Ceinidae Austrochiltonia subtenuis Sayce 3 DECAPODA Parastacidae Cherax quinquecarinatus Gray 3,5 Cher ax tenuimanus Smith 3,5 Cherax preissii Erich son 5 Cherax crassimanus Riek 5 Cherax destructor Clark 3 Engaeiva subcoerulea Riek 5 Engaeiva similis Riek 5 Grapsidae Leptograpsodes octodentatus Milne-Ed wards 5 Palaemonidae "Palaemonetes" australis Dakin 3,5 ANISOPTERA Aeshnidae Aeshna brevis ty la Rambur 3 Austroaeschna anacantha Tillyard 2,5,6 Corduliidae Orthetrum caledonicum Brauer 3,5,6 Austrothemis nigrescens Martin 3,5 Diplacodes bipunctata Brauer 5 Nannophya dalei occidentalis Tillyard 5 Dithrocordulia metallica Tillyard 3 Hesperocordulia berthoudi Tillyard 3 Lotic Lentic Reservation Distribution V + RE v v + W V - w V - w V + w V + w V V + W V + W V + W V + LE V + W V + w V + w V - w V - w V + w V + w V - w V + RE V V + w V V + w V + w V + LE V + W V + W V + RE V V + RE V - RE V V + RE V - RE V + W v + RE V + RE V + RE V + RE V + W V + LE V + LE V + W V + RE V + w V V + RE V V + W V + W V - w V + RE V + RE V + RE 284 Journal of the Royal Society of Western Australia, 79(4), December 1996 Species Source Lotic Lentic Reservation Hetnicordulia australiae Rambur 3,6 V V + Hemicordulia tau Selys 5 V + Procordulia affinis Selys 3,5 V + Synthemis cyanitincta Tillyard 2,3,5 V V + Gomphidae Hemigomphus arttliger Tillyard 2 V Austrogomphus lateralis Selys 2,3 V + Austrogomphus collaris Hagen 3,5 V + Austro gotnphus ochraceus Selys 6 V - ZYGOPTERA Lestidae Austrolestes annulosus Selys 3,5 V + Austrolestes analis Rambur 5 V + Megapodagriidae Argiolestes minimus Tillyard 5 + Austroagrion cyane Selys 3 V + Xantluigrion erythroneurum Selys 6 V - EPHEMEROPTERA Leptophlebidae Bibulmena kadjina Dean 1,2,3 V V + Neboissophlebia occidentals Dean 3 V + Nyungara bunni Dean 1,2 V + Nyungara ellitasha Dean 2 V _ Caenidae Tasmanocoenis tillyardi Lestage 2,3 V V + PLECOPTERA Gripopterygidae Newmanoperla exigua Kimmins 1/2 V V + Leptoperla australica Enderlain 1,2 V + MEGALOPTERA Corydalidae Archichaulibdes cervulus Theischinger 2 V - TRICHOPTERA Leptoceridae Lectrides parilis Neboiss 1,2 V V + Triplectides australis Navas 3 V + Condocerus nr aptus Neboiss 2 V - Notoperata tenax Neboiss 3 V + Atriplectididae Atriplectides dubius Mosely 2,3 V V + Hydroptilidae Acroptila globosa Wells 1,3 V + Ecnomidae Ecnomina ?trulla Neboiss 1,3 V + Ecnomus pansus Neboiss 1,3 V + Ecnomus turgidus Neboiss 3 V - Philorheithridae Kosrheithrus boorarus Neboiss 2 V _ Hydrobiosidae Apsilochorcma urdalum Neboiss 2 V - Taschorema pallescens Banks 2 V - Hydropsychidae Smicrophylax australis Ulmer 2 V - Polycentropodidae Plectrocnemia exirnia Neboiss 3 V + Adectophylax volutus Neboiss 2 V - COLEOPTERA Dytiscidae Sternopriscus Ibrowni Sharp 1,3 V V + Sternopriscus marginatus Watt 1 V V + Homoeodytes scutel laris Germar 1,3 V + Rhantus suturalis MacLeay 1 V + Liodessus inornatus Sharp 3 V + Eiodessus dispar Sharp 3 V + Distribution 285 Journal of the Royal Society of Western Australia, 79(4), December 1996 Species Source Lotic Lentic Reservation Distribution Megaporus solidus Sharp 3 / V + RE Necterosoma darwini Babington 3 V + RE Antiporus fernoralis Boheman 3 / V + W Lancetes lanceolatus Clark 3 V + W DIPTERA Simulidae A us tros imul i u m fu riosum Skuse 1,5 V V + W Cnephia tonnoiri tonnoiri Drummond 5 V - W Chironomidae Aphroteniellafilicornis Brundin 1,2, 3,5 V V + W Aphroteniella ten ui corn is Brundin 5 V - W Paramerina levidensis Skuse 1,2, 3, 5 V V + W Procladius palludicola Skuse 3,5 V + W Procladius ?villosimanus Kieffer 3 V + W Alotanypus dalyupensis Freeman 1,3,5 V + W Coelopynia pruinosa Freeman 3 V + W Corynoneura ?scutellata Winnertz 5 V + W Stictocladius uniserialis Freeman 1,5 V V + W Cricotopus annuliventrus Skuse 1, 2,3,5 V V + W Paralimnophyes pullulus Skuse 3,5 V + W Cladopelma curtivalva Kieffer 1,3,5 V V + W Cryptochironomis griseidorsum Kieffer 1,3,5 V V + W Chironomus occidental is Skuse 3 V + W Chironomus aff alternans Walker 1,3,5 V + W Chironomus tepperi Skuse 4 V + W Dicrotendipes Iconjunctus Walker 1,3,5 V V + W Kiefferulus martini Freeman 1,3,5 V V + W Kiefferulus intertinctus Skuse 3,5 V + W Paratany tarsus grimmii Schneider 5 V + W Stempellina laustraliensis Freeman 3,5 V + W 1 fully described species only Table 2 Native fish species found in freshwater habitats in the Warren Bioregion. Reservation refers to their presence or absence within a nature reserve (A-class) or national park within the region. Distribution: W, widespread; RE, regionally endemic; LE, locally endemic. Sources of information include Christensen (1982), Jaensch (1992), Morgan et al. (1996) and Western Australian Museum records. Species Lotic Lentic Reservation Distribution AGNATHA Geotriidae Geotria australis Gray V + W TELEOSTEI Lepidogalaxiidae Lepidogalaxias salamandroides Mees V + LE Galaxiidae Galaxiella nigrostriata Shipway V V + RE Galaxiella tnunda McDowall V + RE Galaxias occidentals Ogilby v V + RE Galaxias truttaceus Cuvier V V + W Galaxias maculatus Jeyns V V + W Percicthyidae Bostockia porosa Castelnau V V + RE Nannopercidae Edelia vittata Castelnau V V + RE Nannatherina balstoni Regan V V + RE Plotosidae Tandanus bostocki Whitley V V 4- RE Gobiidae Pseudogobius olorum Sauvage V d + W1 Afurcagobius suppositus Sauvage V d + RE1 Atheriniidae Leptatherina wallacei Prince, Ivantsoff & Potter V V + RE' 1 not strictly a freshwater species. 286 Journal of the Royal Society of Western Australia, 79(4), December 1996 species have been recorded from Margaret River, Bunbury, Gingin and Two Peoples Bay (Museum Records; Morgan et al 1996) suggesting that they were once more widely distributed throughout the coastal en¬ vironment of south-west Western Australia. All of these fishes are either included or have been recommended for inclusion in the list of Australian threatened fishes (Anon. 1994). Amphibians Twenty two species of frogs are found in the Warren Bioregion, with three of these occurring just within its boundaries (Table 3). Six of the remaining eighteen spe¬ cies could be considered to be locally endemic, includ¬ ing the south coast froglet Ranidella subinsignifera, the roseate frog Geocrinia rosea, the Walpole frog G. hi lea, the white-bellied frog G. alba, the orange-bellied frog G. Table 3 Frog species known to occur within the boundaries of the War¬ ren Bioregion. Reservation refers to their presence or absence in a nature reserve (A-class) or national park within the region. Dis¬ tribution: RE, regionally endemic; LE, locally endemic. Sources of information include Main (1965), Christensen (1992), Tyler et al. (1994), Roberts et al. (in press) and Wardell-Johnson ( pers . comm.). Species Reservation Distribution Hylidae Litoria adelaidensis Gray + RE Litoria moorei Copeland + RE Myobatrachidae Limnodynastes dorsalis Gray + RE Heleioporus momatus Lee & Main + RE Heleioporus cyrei Gray + RE Heleioporus psammophilus Lee & Main + RE Heleioporus albopunctatus Gray n/a RE1 Neobatrachus pelobatoides Wemer + RE Crinia georgiana Tschudi + RE Ranidella glauerti Loveridge + RE Ranidella pseud insigni) era Main + RE Ranidella insignifera Moore n/a RE1 Ranidella subinsignifera Littlejohn + LE2 gen. et sp. nov Roberts et al. + LE Geocrinia rosea Harrison + LE Geocrinia lutea Fletcher + LE Geocrinia alba Wardell-Johnson & Roberts + LE Geocrinia vitellina Wardell-Johnson & Roberts - LE Geocrinia leai Fletcher + RE Metacrinia nichollsi Harrison + RE Myobatrachus gouldii Gray n/a RE1 Pseudophryne guentheri Boulenger + RE 1 may be found within the boundary of the Warren Bioregion, but only peripherally; reservation status in this bioregion is therefore not applicable (n/a). 2 found in the lower south-west, a distribu¬ tion broadly approximating the Warren bioregion. vitellina and the sunset frog (Myobatrachidae). All four Geocrinia species and the sunset frog have highly limited geographic distributions with G.vitellina and G. alba formally gazetted in the schedules of the Commonweath's Endangered Species Protection Act, 1992. The Reserve System The Warren Bioregion extends over 1 042 000 ha with approximately 25% of that area held as national parks or nature reserves (A-class). We estimate that these reserves incorporate 86% of the aquatic faunal elements found in the region, but 7% of locally restricted species are apparently not included (Table 4). The further inclusion of locally endemic species or assemblages into the reserve system is made difficult because they are often rare as well as restricted in their distribution. Large-scale surveys across all aquatic habitats within the lower south-west would be necessary to locate all of these aquatic fauna. This is unlikely to occur in the near future and so the ability to predict the occurrence of these fauna at unsurveyed sites would be advantageous. Predictable patterns of faunal communities often occur in wetlands with similar physical and chemical characteristics (e.g. Edward et al. 1994). Thus, it may be possible to give priority for reservation to aquatic habi¬ tats with particular biophysical traits which, elsewhere, support high levels of endemic species. Highly acidic environments and aquatic habitats associated with gran¬ ite outcrops provide good examples. The four species of copepod, and two species of cladoceran found by Bayly (1992) were all acidophilic and were collected from ponds with low pH. Two species of ostracod, llyodromus candonites and Kapcypridopsis asymmetra, and one species of chironomid, Allotrissocladius sp, which are endemic to the Warren Bioregion, have been found in temporary pools associated with granite outcrops (Bayly 1982). In addition, the sunset frog, which is thought to have origi¬ nated approximately 30-36 million years ago was re¬ cently found in an organic-rich swamp at the base of a granite outcrop near Walpole (Roberts et al 1997). Granite outcrops and other areas elevated above the level of the Eocene marine incursion, such as headwater regions, permanent freshwater flows, and elevated coastal locations receiving orographic rainfall, often sup¬ port relictual fauna (Main & Main 1991; Hopper et al. 1996). These relictual habitats generally provide organi¬ cally rich and permanently moist microhabitats (Main & Main 1991) for a fauna which has persisted since eustatic changes and the onset of seasonal aridity in the early Tertiary (Keast 1981). Relictual fauna are of exceptional importance from a nature conservation perspective and thus the systematic evaluation of existing reserves and Table 4 Summary of the distribution and reservation status of aquatic fauna found in the Warren Bioregion. Number of species found, thus far, in nature reserves (A-class) or national parks within the region are shown in parentheses. Taxa Species1 Locally Endemic Regionally Endemic Widespread Invertebrates 156 10 (9) 49 (41) 97 (80) Amphibians 22 6 (5) 13 (13) 0 Fish2 14 1 (1) 9 (9) 4 (4) 1 fully described species only; 2 native species only. 287 Journal of the Royal Society of Western Australia, 79(4), December 1996 the inclusion of new reserves to ensure adequate repre¬ sentation of relictual habitats should be an urgent task. Are the fauna protected in the reserve system? National parks and nature reserves provide a refuge for a substantial proportion of the aquatic fauna. Their protection, however, cannot be assured unless on- reserve management procedures and activities outside the reserve system are sympathetic to the preservation and maintenance of their habitat. The practice of conducting fuel reduction burns dur¬ ing late spring, summer and autumn, when the soil is dry, may inadvertently threaten some aquatic fauna either directly or through the loss of soil as habitat. Species such as L. salamandroides oversummer in the moist substrate of peatlands and shrublands (Pusey 1990), while other species deposit drought resistant eggs. These organically-rich soils bum readily and hence the fauna within the substratum may be lost. In addition, there may be changes in local hydrology, such as the creation of more surface pools, or improved drainage of sandy soils, which alter the proportional occurrence of habitats and some aquatic species. Activities occurring within catchments of conservation reserves may impact adversely on the fauna. Harvey (1996) noted that Poorginup Swamp, within the Lake Muir Nature Reserve, was threatened with increased salinization as a result of nearby agricultural clearing. The swamp is the type locality for two species of water mite, Acercella poorginup and Pseudoln/dn/phantes doegi , both of which were thought to be extinct as a result of recent hydrological changes occurring within the swamp (Harvey 1996). The latter species, however, has been found in one other location, within the Shannon River National Park (Horwitz 1994). Adverse land-use activities within catchments are even more apparent in flowing waters where anthropogenic activity upstream may directly influence downstream habitats (e.g. Walker 1985; Davey et at. 1987; Campbell & Doeg 1989). Land clearing for agriculture in Western Australia has resulted in increased salinity (Schofield 1990) and sedimentation (Williams 1992) of many rivers and clear- fell logging which occurred in the last decade is still affecting streams today (Crowns & Davis 1991; Trayler & Davis, unpubl. obs .). A river and stream zone system was introduced into the State Forest in the mid-70s (Anon 1977). This was later modified and today this system is estimated to include some 63 100 ha of land in the southern forest region (Anon 1992a). In addition to increasing the size of the conservation estate, these zones of undisturbed vegetation are designed to act as a buffer and minimize the effect of logging opera¬ tions on the water quality of nearby streams and rivers (Anon 1992a). Large buffer zones (100 m) have been proven effective in reducing the input of sediment into second order streams (Borg et al. 1987), but the effective¬ ness of smaller zones (20-30 m), which are routinely used on these streams, has not been assessed in Western Australia. While there is ample evidence to suggest that these buffers may prevent increased sedimentation (Davies & Nelson 1994; Clinnick 1985), it is unlikely that such small buffers would prevent increases in stream salinity. Borg et al. (1987) found that even 100 m buffers would not prevent a rise in salinity in streams adjacent to logged coups. There is, however, some evidence to suggest that a large buffer will reduce the period that salinities remain elevated from fifteen years, as esti¬ mated by (Borg et al. 1988), to eight years, as docu¬ mented by Growns & Davis (1991). In the past, rising salinity has not been considered important in high rain¬ fall areas where salinities in logged catchments gener¬ ally remain within the range considered acceptable for drinking water (Anon 1992b). There is, however, in¬ creasing evidence that the invertebrate fauna of this re¬ gion may be intolerant of relatively small increases in salinity (Growns & Davis 1991; Trayler & Davis, unpubl obs.). The river and stream zone system plays an important role in the conservation of the endemic invertebrate fauna of lotic origin. Many of these fauna have, thus far, not been found elsewhere in the conservation estate (Table 1) and it is therefore essential that riparian buffer zones adjacent to logging coupes are of an adequate size and that these areas are managed properly. Growns (1992) argued that these buffer strips would be compromised if poorly constructed roads or accessways crossed streams to access coupes. In addition, the headwater streams within the karri forest often comprise low gradient, marshy areas with ill-defined and ephemeral water courses which may not be recognised or mapped as first-order streams. There is, therefore, potential for these areas to be overlooked and not included as part of the stream reserve system. Translocated and exotic fish species are widespread in the waterways of the south-west with increasing anecdotal evidence that these introduced species have a serious impact on the distribution of the native fish fauna (see Morgan et al. 1996). Of particular concern is the translocation of piscivorous species such as the golden perch Macquaria ambigua and the silver perch Bidyanus bidyanus to private dams within the catchment of the D'Entrecasteaux National Park (Morgan et al. 1996). These species might further restrict populations of the native fishes if they were to escape into natural waterways. The non-aestivating native fish species, N. balstoni, Galaxias occidental is, G. munda , Edelia vittata and Bostockia porosa are particularly vulnerable to predation by introduced fishes during summer and autumn dry periods when they are forced to retreat to permanent pools and streams (Morgan et al. 1996). With the exception of N. balstoni , all of these species as well as the lamprey Geotria australis, were once found in abundance in the headwaters of Big Brook near Pemberton (Pen et al. 1988, 1991), but today their numbers have declined dramatically (Morgan & Gill 1996). In particular, G. munda, which was once extremely common to the system (Pen et al. 1988, 1991), has now disappeared up¬ stream of the dam and this has been attributed to the recent introduction of the voracious and piscivorous redfin perch Perea fluviatilis to Big Brook Dam (Morgan et al. 1996). Big Brook Dam also provides an ideal habitat and potential refuge for a number of other predatory species including Salmo trutta, Gambusia holbrooki and 288 Journal of the Royal Society of Western Australia, 79(4), December 1996 Oncorhynchus mykiss (Morgan et al. 1996), whose presence may have also adversely affected the native fish fauna in the Big Brook area. While this is yet to be documented for the Warren Bioregion, elsewhere these fishes have been widely implicated in the fragmentation of fish distributions (e.g. Tilzey 1976; Lloyd 1990; Hutchinson 1991; Crowl et al. 1992). A variety of other anthropogenic activities and threatening processes operate within nature reserves in the region. Road building activities and the construction of fire breaks, or scrub rolling for fire suppression, in¬ crease the likelihood of sediment deposition into wet¬ land systems and enhance the potential for the spread of soil borne fungal pathogens. The effects of these activi¬ ties have not been fully documented to date. Recommendations Conservation reserves within the Warren Bioregion in south-west of Western Australia play an important role in the protection of both the endemic and cosmopolitan aquatic fauna of the entire south-west. The wetlands within these reserves are important refuges for aquatic invertebrates, fishes and amphibians and will become increasingly so as the pressure of urbanisation and agriculture intensifies. Priority for further reservation should be given to aquatic habitats which are known to support a high level of endemic or relictual fauna. It should, however, be acknowledged that the presence of rare fauna is often difficult to detect, and unique faunal assemblages may occur in many of the undisturbed and unsurveyed wetlands of this region. Thus, all relatively undisturbed aquatic habitats in this region should be afforded some protection, at least until they have been properly surveyed. Since aquatic habitats do not exist in isolation, they cannot be protected within a reserve system without reference to activities occurring within the catchment. Adverse impacts can arise through land clearing, the construction of dams, the introduction of exotic fishes, habitat removal, salinization, sedimentation and eutrophication associated with agriculture, mining, logging and road construction. Careful planning, which is sensitive to the fragility of aquatic habitats, as well as adequate buffers between anthropogenic activities and reserves are essential for the preservation of the fauna in this region. The effectiveness of a reserve system is also dependent on adequate management within the reserves themselves, where it is essential that due consideration be given to the habitat requirements of the aquatic fauna. This is particularly important for aquatic environments which may be dry for six months or more each year. Seasonal and ephemeral waterbodies must be recognised as comprising an important part of the diversity of aquatic habitats and the absence of water in these envi¬ ronments does not necessarily preclude the presence of aquatic fauna. Differences between management policy and practice can lead to detrimental effects on aquatic environments (see Davies & Nelson 1994). The environmental awareness of workers carrying out management procedures may be improved through training specifically in ecological and environmental issues. This important aspect of the management of natural areas currently receives little attention. Finally, we cannot afford to be complacent with respect to the conservation of the aquatic fauna. Although reserved lands within the Warren Bioregion are extensive, these areas may not be protected in perpetuity. This has been demonstrated by the recent excision of land from the D'Entrecasteaux National Park, near Lake Jasper for the potential purpose of sand mining. References Allen G R 1982 A Field Guide to Inland Fishes of Western Australia. Western Australian Museum, Perth. Anon. 1977 General Working Plan 86. Forests Department, Perth. Anon. 1992a Management strategies for the south-west forests of Western Australia: A review. Department of Conservation and Land Management, Perth. Anon. 1992b Australian water quality guidelines for fresh and marine waters. National Water Quality Management Strategy. Australian and New Zealand Environment and Conservation Council, Canberra. Anon. 1994 Australian Threatened Fishes 1994 supplement. Australian Society of Fish Biology Newsletter 24:19-24. Anon. 1994 Guidelines for Protected Area Management Categories. IUCN, Gland, Switzerland. Balia S A 1994 Wetlands of the Swan Coastal Plain Vol 1: Their Nature and Management. Water Authority of Western Australia & the Department of Environmental Protection, Perth. Bayly I A E 1979 Further contributions to a knowledge of the centropagid genera Boeckella, Hemiboeckella and Calamoecia (athalassic calanoid copepods). Australian Journal of Marine and Freshwater Research 30:103-127. Bayly I A E 1982 Invertebrate fauna and ecology of temporary pools on granite outcrops in southern Western Australia. Australian Journal of Marine and Freshwater Research 33:599- 606. Bayly I A E 1992 The micro-crustacea and physico-chemical features of temporary ponds near Northcliffe, Western Australia. Journal of the Royal Society of Western Australia 75:99-106. Benzie J A FI 1986 Daphnia occidentalis, new species (Cladocera: Daphniidae) from Western Australia: new evidence on the evolution of North American D. atnbigua-D. middendorffiana group. Journal of Crustacean Biology 6:232-245. Benzie J A Ff 1988 The systematics of Australian Daphnia (Cladocera: Daphniidae). Species descriptions and keys. Hydrobiologia 166:95-161. Borg H, King P D & Loh I C 1987 Stream and ground water response to logging and subsequent regeneration in the southern forest of Western Australia: Interim results from paired catchment studies. Water Authority of Western Aus¬ tralia WH 34, Perth. Borg H, Stoneman G L & Ward C G 1988 The effect of logging and regeneration on groundwater, streamflow and stream sa¬ linity in the southern forest of Western Australia. Journal of Hydrology 99:253-270. Bunn S E & Davies P M 1992 Community structure of the macroinvertebrate fauna and water quality of a saline river system in south-western Australia. Hydrobiologia 248:143- 160. Bunn S E, Edward D H & Loneragan N R 1986 Spatial and temporal variation in the macroinvertebrate fauna of streams of the northern jarrah forest. Western Australia: community structure. Freshwater Biology 16:67-91. 289 Journal of the Royal Society of Western Australia, 79(4), December 1996 Campbell I C & Doeg T J 1989 Impact of timber harvesting and production on streams: A review. Australian Journal of Marine and Freshwater Research 40:519-539. Christensen P 1982 The distribution of Lepidogalaxias salamandroides and other small freshwater fishes in the lower south-west of Western Australia. Journal of the Royal Society of Western Australia 65:131-141. Christensen P 1992 The karri forest: Its conservation significance and management. Department of Conservation & Land Man¬ agement, Perth. Churchward H M, McArthur W M, Sewell P L & Bartle G A 1988 Land forms and soils of the south-coasts and hinterland. Western Australia: Northcliffe to Manypeaks. CSIRO Divi¬ sional Report 88/1, Perth. Clinnick P F 1985 Buffer strip management in forest operations: a review. Australian Forestry 48:34-45, Crowl T A, Townsend A R & McIntosh A R 1992 The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Reviews in Fish Biology and Fisheries 2:217- 241. Davey G W, Doeg T J & Blyth J D 1987 Changes in benthic sediment in the Thomson River, South-Eastern Australia, during construction of the Thomson Dam. Regulated Rivers 1:71-84. Davis J A, Rosich R S, Bradley J S, Growns J E, Schmidt L G & Cheal F 1993 Wetlands of the Swan Coastal Plain Vol 6: Wetland classification on the basis of water quality and invertebrate community data. Water Authority of Western Australia and the Environmental Protection Authority, Perth. Davies P E & Nelson M 1994 Relationships between riparian buffer widths and the effects of logging on stream habitat, invertebrate community composition and fish abundance. Australian Journal of Marine and Freshwater Research 45:1289-1305. Edward D H D, Gazey P & Davies P M 1994 Invertebrate community structure related to physico-chemical parameters of permanent lakes of the south coast of Western Australia. Journal of the Royal Society of Western Australia 77:51-63. Growns I O 1992 Macroinvertebrate community structure in the streams of the southern forests of Western Australia: The influence of seasonality, longitudinal gradients and forestry activities. PhD Thesis, Murdoch University, Perth. Growns I O & Davis J A 1991 Differences in the macroinvertebrate communities of streams in clearfelled and undisturbed areas in experimental catchments 8 years after harvesting. Australian Journal of Marine and Freshwater Research 42:689-706. Growns I O & Davis J A 1994 Longitudinal changes in near-bed flows and macroinvertebrate communities in a Western Australian stream. Journal of the North American Benthological Society 13:417-438. Growns J E, Davis J A, Cheal F, Schmidt R G, Rosich R S & Bradley S J 1992 Multivariate pattern analysis of wetland in¬ vertebrate communities and environmental variables in West¬ ern Australia. Australian Journal of Ecology 17:275-288. Halse S A 1989 Wetlands of the Swan Coastal Plain - past to present. In: Swan Coastal Plain Groundwater Management Conference Proceedings (ed G Lowe). Western Australian Water Resources Council, Perth, 105-112. Harvey M 1996 A review of the water mite family Pionidae in Australia (Acarina: Hygrobatoidea). Records of the Western Australian Museum 17:361-393. Hopper S D, Keighery G J & Wardell-Johnson G 1992 Flora of the karri forest and other communities in the Warren Botanical Subdistrict of Western Australia. In: Research on the Impact of Forest Management in South-West Western Australia (ed M Lewis). Department of Conservation and Land Management Occasional Paper 2/92, 1-32. Hopper S D, Harvey M, Chappel J, Main A R & Main B Y (1996) Gondwanan heritage - A review. In: The Gondwanan Heritage: Past, Present and Future of the Western Australian Biota, (eds S D Hopper, J Chappel, M Harvey & A George). Surrey Beatty & Sons and Kings Park and Botanic Garden, Chipping Norton and Perth, 1-46. Horwitz P 1994 Patterns of endemism in the freshwater fauna of the far southern peatlands and shrublands of south-western Australia. Report. Australian Heritage Commission, Perth. Horwitz P 1996 Comparative endemism and richness of the aquatic invertebrate fauna in peatlands and shrublands of far south-western Australia. Memoirs of the Museum of Victoria 56:313-322. Hutchinson M J 1991 Distribution patterns of redfin perch Perea fluviatilis Linnaeus and western pygmy perch Edelia vittata Castelnau in the Murray River system Western Australia. Records of the Western Australian Museum 15:295-301. Jaensch R P 1992 Fishes in wetlands on the south coast of Western Australia. Technical Paper. Department of Conservation and Land Management, Perth. Jarvis N T 1979 Western Australia: An atlas of human endeavour. Education and Lands and Surveys Departments of Western Australia, Perth. Keast A 1981 Distributional patterns, regional biotas, and adaptations in the Australian biota: A synthesis. In Ecological Biogeography of Australia (ed A Keast). Dr W Junk, The Hague. Lloyd L 1990 Ecological interactions of Gambusia holbrooki with Australian native fishes. In: Introduced and Translocated Fishes and their Ecological Effects (ed D.A. Pollard). Australian Society for Fish Biology, Bureau of Rural Resources Proceedings. Australian Government Publishing Service, Canberra, 94-97. Main A R 1965 Frogs of Southern Western Australia. Handbook 8, Western Australian Naturalists' Club, Perth. Main A R & Main B Y 1991 Report on the Southern Forest Region of Western Australia. Report. Australian Heritage Commission, Perth. Morgan D, Gill H & Potter I C 1996 The distribution of freshwa¬ ter fish in the south-western comer of Western Australia. Re¬ port. Water Authority of Western Australia, Perth. Morgan D & Gill H 1996 The effect of Big Brook Dam during drought on the fish communities of the Lefroy and Big Brooks. Report. Water and Rivers Commission, Perth. Morton D W 1990 Revision of the Australian Cyclopidae (Copepoda: Cyclopoida). II. Eucyctops Claus and Ectocyclops Brady. Australian Journal of Marine and Freshwater Research 41:657-675. Mulcahy W M, Churchward H M & Dimmock G M 1972 Landforms and soils on an uplifted peneplain in the Darling Range, Western Australia. Australian Journal of Soil Research 10:1-14. Olsen G & Sfcitmore E 1991 State of the Rivers of the South West Drainage Division. Western Australian Water Resources Council 2/91, Perth. Pen L J, Hilliard R W & Potter I C 1988 The Effect of Big Brook Dam and the Pemberton Weir on the migration of native fish. Report. Water Authority of Western Australia , Perth. Pen L J, Hilliard R W & Power G W 1991 Fish and lamprey monitoring in the Lefroy/ Big Brook system in south-west Western Australia. Report. Water Authority of Western Aus¬ tralia, Perth. Pusey B J 1990 Seasonality, aestivation and the life history of the salamanderfish Lepidogalaxias salamandroides (Pisces: Lepidogalaxiidae). Environmental Biology of Fishes 29:15-26. Pusev B J & Edward D H 1990a Limnology of the southern acid peat flats, South-Western Australia, journal of the Royal Society of Western Austalia 73:29-46. Pusey B J & Edward D H 1990b Structure of fish assemblages in waters of the southern acid peat flats, south-western Australia. Australian Journal of Marine and Freshwater Research 41: 721-734. Roberts J D, Horwitz P, Wardell-Johnson G, Maxson L R & Mahony M 1997 Taxonomy, relationships and conservation of 290 Journal of the Royal Society of Western Australia, 79(4), December 1996 a new genus and species of myobatrachid frog from the high rainfall region of south-western Australia. Copeia 1997: 373- 381. Schofield N J 1990 Water Interactions with land use and climate in south western Australia. Western Australian Water Authority Report WS60, Perth. Storey A W, Bunn S E, Davies P M & Edward D H 1990 Classifi¬ cation of the macroinvertebrate fauna of the river systems in southwestern Australia in relation to physical and chemical parameters. Regulated Rivers: Research & Management 5:217-232. Thackway R & Cresswell I D 1995 An interim biogeographic regionalisation for Australia: A framework for setting priori¬ ties in the national reserves system cooperative program. Aus¬ tralian Nature Conservation Agency, Canberra. Tilzey R D 1976 Observations of interactions between indigenous Galaxiidae and introduced Salmoniidae in the Lake Eucumbene catchment. New South Wales. Australian Journal of Marine and Freshwater Research 27:551-564. Tyler M J, Smith L A & Johnstone R E 1994. Frogs of Western Australia. Western Australian Museum, Perth. Walker K F 1985 A review of the ecological effects of river regulation in Australia. Hydrobiologia 125:111-129. Williams P 1992 The state of Western Australian rivers. Land and Water Research News 13:8-15. Williams W D, Taaffe R G & Boulton A J 1991 Longitudinal distribution of macroinvertebrates in two rivers subject to salinization. Hydrobiologia 210: 151-161. 291 Journal of the Royal Society of Western Australia, 79:293-300, 1996 Ecosystem dynamics and management in relation to conservation in forest systems R J Hobbs CSIRO, Division of Wildlife & Ecology, LMB 4, PO Midland WA 6056 Abstract Any system of conservation reserves sits within the context of the surrounding ecosystem. Modifications to this surrounding matrix will inevitably have some impact on the reserve system. While forest ecosystems in south-western Australia are significantly less modified than other ecosystems, particularly in the agricultural area, they are nevertheless subjected to marked human-induced modifications. These take the form of forest management practices including timber harvesting and fuel reduction burning, and the impacts of introduced species including the pathogen Phy toph thorn cimmmomi. The level of information available to assess the impacts of these modifications is for the most part inadequate. Changing ideas on the nature and dynamics of ecosystems require a reappraisal of how we manage ecosystems for conservation and production purposes. The recognition of interconnections between different impacts and between different ecosystem components has resulted in the development of the concept of "ecosystem management". This aims to integrate the various management goals and allow production to take place in such a way that both the long term productive potential and the biodiversity of the forest are maintained. Introduction While there is currently considerable debate over the selection of reserves within forests in Australia, less at¬ tention has been paid to how these reserves are likely to fare within the context of the forest ecosystem as a whole. I argue here that conservation reserves do not form discrete entities which can be considered and man¬ aged separately from the rest of the forest. Rather, the whole forest has to be considered as a collection of inter¬ acting parcels of land, and events in one parcel are liable to impact those in surrounding areas. This has three ma¬ jor implications. Firstly, reserves must be managed not only to meet the needs of the biota being conserved, but also in the context of the main ecosystem processes pre¬ vailing in the forest. Secondly, the impacts of manage¬ ment activities in the forest as a whole have to be as¬ sessed. Finally, the interconnected nature of natural sys¬ tems means that the forest as a whole, including those areas outside conservation reserves, plays a vital role in the overall conservation of biodiversity. I explore these points in this paper, and discuss their implications for forest management, with particular reference to Western Australia. Ecosystem dynamics The forest ecosystem consists of the forest biota and its environment (Fig 1). The biotic components are divided up according to their primary functions (i.e. primary producers, consumers, decomposers etc.). These, and other non-living components (dead organic matter, inorganic material) form pools within which Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 carbon, other nutrients, water and other important ele¬ ments accumulate. Ecosystem dynamics describe the flows of energy and elements into and out of the system, and between the various pools (e.g. Waring & Schlesinger 1985; Aber & Melillo 1991). In addition, the responses to, and recovery from, disturbance form a fur¬ ther important set of ecosystem dynamics, incorporating the ideas of succession and resilience. Ecosystem processes have frequently been ignored in conservation management, presumably because conservation has been primarily directed at species and biotic communities. Ecosystem ecology has largely developed separately from population and community ecology, and has tended to subsume biotic components into larger "black boxes" (Aber & Melillo 1991), as illustrated in Figure 1. While population and community Exchange with other ecosystems Figure 1. A simplified representation of an ecosystem, indicating the major pools and flows of carbon and nutrients. 293 Journal of the Royal Society of Western Australia, 79(4), December 1996 A B Ecosystem response Figure 2. Characteristics of complex ecosystems which render the search for simple cause and effect relationships difficult. A Non-linear system response to human activities. The system may show little or no response until a certain level of impact is experienced, at which stage a sudden system change occurs. B Complex interactions between system components and human activities. The diagram shows the main factors discussed for Western Australian forest, and potential linkages between them. ecology have been concerned with entities such as species and populations, ecosystem ecology is concerned with the flows of materials between components - what Pickett et al. (1994) have termed the "things versus stuff" dichotomy. However, it is increasingly recognized that the links between species and ecosystems are important, and attempts are now being made to integrate the two streams of ecology (Pickett et al. 1994; Jones & Lawton 1995). In forest systems of Western Australia, the major ecosystem dynamics to be considered include the natural dynamics of the forest, and a set of dynamics imposed by humans. Here, natural forest dynamics are determined largely by three factors which predominate in the area i.e. a mediterranean-type climate, with its well-defined summer drought; soils with low nutrient status; and the incidence of fire and other disturbances. These factors, and ecosystem responses to them, have been considered in detail elsewhere (Kruger et al 1983; Dell et al. 1986; Dell et al. 1989; Davis &c Richardson 1995), and I will not dwell on them here. Rather, I will concentrate on the set of imposed dynamics. The most 200 r 160- * Stream 120 - • flow (mm) 80 # 40 # • • 0 - - - f - 0 20 40 60 Area affected by dieback (%) Figure 3. Streamflow in catchments with different proportions affected by Phytophthora cinnamomi (redrawn from Schofield et al. 1989). important of these are forest management practices, in particular timber harvesting and fuel reduction burning, and the impacts of invasive species, including the intro¬ duced pathogen, Phytophthora cinnamomi. Introduced species Phytophthora cinnamomi is undoubtedly a major factor influencing the forest ecosystem in Western Australia and in other parts of the country (Dell & Malajczuk 1989). A recent symposium has highlighted the impacts of the disease on a variety of forest components (Withers et al. 1994). From an ecosystem perspective, the disease is important because of its impacts on forest structure and composition and resulting environmental changes. Depending on the severity of attack, Phytophthora causes the loss of overstorey and understorey plant species, which then presumably alters the microclimate and reduces evapotranspiration. This in turn leads to increased input of water to the system, which has been recorded as increasing stream flows with increasing degrees of dieback incidence (Fig 3; Schofield et al. 1989). This change in local hydrology could potentially lead to localized vegetation change (Davidson 1994), but this has not been investigated in detail. Introduced predators, particularly foxes and cats, are thought to be one of the major causal factors leading to the complete or near extinction of many Australian marsupials (Burbidge & McKenzie 1989; Friend 1990). Indeed, Western Australian forests have provided the last refuge for a number of species, presumably because fox numbers remained lower than in other areas, and/ or they arrived later. Predator control is now practiced in many areas, with obvious success indicated by increases in abundances of native mammals (Kinnear et al. 1988; Friend 1990). Changes in abundances of mammals in ecosystems are liable to have impacts on other system components due to their herbivory, digging activities and dispersal of fungal spores (Lamont et al. 1985; Noble 1993; Lamont 1995). Such interactions 294 Journal of the Royal Society of Western Australia, 79(4), December 1996 and their disruption may have important, but difficult to detect, effects on ecosystem function (Hobbs et al. 1995). Invasive plants are another major threat to many ecosystems across Australia (Humphries et al. 1991; Humphries 1993). Individual invaders, such as bridal creeper (MyrsiphyUum asparagoides), have the potential to crowd out native species and alter vegetation composition and structure. Herbaceous species, especially grasses, also have the potential to alter fire regimes by changing the structure and availability of fuel (D' Antonio & Vitousek 1992). In the region of 1500 species are currently naturalized across Australia, with 1032 species recorded from Western Australia (Keighery 1995). Two hundred and twenty species are recognized as noxious weeds across Australia (Parsons & Cuthbertson 1992), and many are problems in native ecosystems. In addition to existing problems, there are likely to be many more species which could become a problem in the future (Hobbs 1993a). These include, for instance, pine species planted in native forest areas which have been found to be invasive elsewhere in the world (Richardson et al. 1994). Australia continues to import plant species without due consideration for the potential threats of invasiveness. There is a clear need for an integrated approach to weed management in the forests and elsewhere (Hobbs & Humphries 1995). Forest management practices Opinions vary as to the extent to which current forest management practices affect ecosystem processes (Abbott & Christensen 1994; Calver et al. 1996). A repeated assertion is that there is no evidence to show that the ecological processes that maintain the forests have been impaired or that forest biodiversity is impacted by forest management (Abbott & Christensen 1994; Anon 1994). However, this may be partially due to the lack of monitoring and research into such potential impacts. In most parts of Australia, little research work has been conducted into the long-term impacts of timber or burning operations. The impact of disturbance- causing activities concentrates on individual species, and no long term monitoring has been implemented, apparently because it is "very difficult" (Anon 1994, p52). This then leads to a dearth of information with which to assess statements on the impacts of management operations. For instance, Schofield et al. (1989) stated that "No long term detailed studies on the effects of different silvicultural systems on the hydrology of the jarrah forest have been carried out". This situation is being redressed (Stoneman 1993), and some information is available from other studies. These indicate that timber harvesting can lead to increases in streamflow over a period of years. Streamflow increases of 91-182% were reported by Borg et al (1987) following heavy logging, and Bari et al. (1994) have shown marked increases in streamflow and groundwater discharge following clear felling. These impacts were highest immediately after logging, but persisted for at least 8 years. These findings concur with experimental work carried out elsewhere (Bormann & Likens 1981). Impacts of logging on other system components are equally poorly documented. What, for instance, are the impacts of timber harvesting on the nutrient pools in the forest? There are indications from forests elsewhere that impacts on soil properties can be substantial (Rab 1994). While findings from one forest type are not necessarily directly applicable to another, such results indicate that efforts should be made to assess possible impacts in Western Australian forests. Considerable effort has been expended in Tasmania, for instance, in assessing impacts of logging systems on system components (e.g. Hickey 1994; Taylor & Haseler 1995) and clear recognition of the need to consider impacts of management on biodiversity are evident in management manuals (Taylor 1991; Anon 1993; Duncan & Packham 1994; Jackson & Taylor 1994). Little of this type of assessment has been carried out in Western Australia. Work by Vlawson & Long (1994) has suggested that current logging practices significantly alter habitat suitability for some bird species, although the validity of the methods used has been questioned by Burrows et al. (1995). Similar discussions on the impacts of logging have occurred in Victoria, where Attiwill (1994a,b) has implied that logging is in many ways equivalent to natural forest disturbance, a conclusion which has been contested (Lindenmayer 1995; see also subsequent response by Attiwill 1995). A similar story is apparent when the impacts of fire management are examined (Williams & Gill 1995). The jarrah forest is currently subjected to widespread short- rotation fuel reduction burning, which has been developed to reduce the risk of destructive wildfires. Controversy surrounds the questions of whether such a burning regime is effective and whether it has adverse impacts on the forest ecosystem (McGrath 1985; Tingay 1985; Underwood et al. 1985). It has been claimed that the current regime mimics the regime prevailing prior to European settlement (Burrows et al. 1995), although the evidence for this is not compelling. Burrows et al. (1995) state that fire scars on Eucalyptus marginata trees indicate the occurrence of moderate to severe fires in the forest occurred with a mean interval of 81 years. They then use historical accounts of aboriginal burning and lightning records to conclude that these severe fires must have been accompanied by low intensity fires every 2-5 years. The question remains as to whether such a regime of low intensity fires prevailed over the whole forest or was restricted to areas most frequented by aborigines. If it prevailed over the whole forest, this then suggests that high intensity fires are possible even under a regime of fuel reduction burning, as currently practiced. Indeed, observations in other forest systems suggests that intensive fuel reduction measures do not necessarily "fire proof" forests (DellaSala et al. 1995). Abbott & Christensen (1994) suggest that the current fire regime (and logging activities) are "....a minor , irregular and relatively insignificant perturbation.. .". On the other hand, McCaw & Burrows (1989) conclude that " While many studies have examined the effects of one , or occasionally several fires , on plant and animal communities in the forest , the basis for predicting longer term effects of different fire regimes is limited". Indeed, the impact of one fire may be minimal, although even this conclusion is open to question (e.g. Majer & Abbott 1989). Again, evidence from Victoria points to a potentially detrimental impact of fuel reduction burning (Hamilton 295 Journal of the Royal Society of Western Australia, 79(4), December 1996 el al. 1991), although the conclusions of this study again have been questioned (McCaw 1993). Nevertheless, it is the overall fire regime ( i.e . frequency, intensity, season, size etc.) which shapes the vegetation in the long term, and long-term data on vegetation changes under current fire management practices are not available. Analyses such as that by Abbott & Christensen (1994) look at relatively short time scales and suggest no significant ecosystem changes. For instance, the present fuel reduction fire regime has been in place for less than 40 years, a short time span in terms of forest dynamics. Longer term impacts could be significant, but will not be noticed if relevant monitoring systems are not in place. Even if impacts are detected, it could be some considerable time before a policy or management response is implimented. A clear example of the types of lag involved in responding to problems is the salination of agricultural land caused by past land clearance. This problem was first documented in the 1920s (Wood 1924) but is only now being acted on at the policy level. Complexity and non-linearity The degree of debate over the importance of changes to forest ecosystems arising from management practices indicates the difficulty in reaching conclusions on the issue. The problem is further compounded by two characteristics of natural systems. The first is the likelihood that system components exhibit non-linear responses to particular activities (Fig 2A). In other words, the system may change unpredictably or may show relatively little change in response as the intensity of an activity increases, until a threshold level is reached, at which point system behavior changes dramatically. Alternatively, the system may show no response to a particular activity for a period of time, and then change rapidly. Such non-linear behaviour is a recognized characteristic of complex systems (e.g. Roberts 1994). The second characteristic is the complexity of interactions between system components and processes (Fig 2B). This may be referred to as the "ECWEE" principle i.e. "everything is connected with everything else" (see Oppenheimer 1995). While this may be an over-generalization, it is nevertheless the case that complex interactions and feedback loops are common in natural systems. Classical scientific approaches to environmental questions attempt to deny the importance of these interconnections since they render the search for simple cause-and-effect relationships almost impossible. Certainly, much can still be gained by single factor studies, but failure to recognize the potential for complex interactions between factors can also lead to simplistic and misleading conclusions. For instance, while there has been a considerable body of research on the various individual components in Figure 2B, the possible interactions have received little attention. Are there, for example, interactions between the effects of forest management and the spread of Phytophthora ? An important part of the problem, however, lies in the continued assertion that current management practices are having little or no impact, even in the absence of data (Abbott &Christensen 1994). It is clear that all management has some impact on the ecosystem (even if the management is to do nothing). The Figure 4. North Bungulla nature reserve in the Western Australian wheatbelt, illustrating the location of the reserve within a greatly- modified agricultural matrix. Although less obvious, forest reserves also sit within an altered matrix. (Photograph by Dion Steven). 296 Journal of the Royal Society of Western Australia, 79(4), December 1996 important question is whether the level of impact is acceptable or not. The acceptability or otherwise of any particular impact will change as society's expectations and priorities change. It is clear that forest managers need to be responsive to such changes (Gordon 1994). However, the acceptability of the level of impact can be assessed only if the relevant information is available. In 1977, the Senate Standing Committee on Science and the Environment (Anon 1977) stated that "the extreme lack of knowledge on the biological sphere . is hampering responsible decision making In 1993, the Resource Assessment Commission (Anon 1993) still had to conclude that " the level of information on impacts appears insufficient for most current uses". This is echoed internationally by the US National Research Council (Anon 1990), which concluded that //.. the existing level of knowledge is inadequate to develop sound management practices". Obtaining this knowledge for Western Australian forests is problematic in the face of current underfunding for research within state agencies and in an environment which is not conducive to open scientific debate on forest issues. Ecosystem management How is all this relevant to the selection and management of conservation reserves? Surely it could be argued that impacts in the parts of the forest managed for production are irrelevant if adequate areas are set asiue for conservation? Unfortunately, it is becoming increasingly recognized that this is not the case, and th.it conservation management and production management have to be integrated to achieve the goal of sustainability. Biota and ecosystem processes do not respect legal boundaries (Newmark 1985), and different parts of the landscape interact. Reserves are located within a surrounding altered or managed matrix. This dichotomy between reserve and matrix is obvious in cases where the matrix is noticeably altered, for instance in an agricultural situation (Fig 4). In these situations, there are clear impacts of the surrounding matrix on the remnant vegetation within reserves (Saunders el al. 1991; Hobbs 1993b; Hobbs 1994). In the case of forests, the impacts are less obvious, because forestry operations do not necessarily create an entirely transformed matrix. Nevertheless, modifications outside reserves can have impacts within the reserves, and reserves are not immune from factors arising in the surrounding matrix, such as dieback, fire, feral animals and so on. At the same time, it is also becoming recognized that reserve systems will not be sufficient on their own to conserve the biodiversity of a region, and that conservation and production management have to be integrated to achieve the goals of sustainability and conservation. For instance. Sample et al. (1993) suggest that "Many ecologists agree that neither our current system of forest reserves... nor any conceivable such system will be sufficient to provide adequate protection of biodiversity" . They continue, "We are urged .. to consider all lands within the ecosystem as important to its overall functioning and sustainability" . Further, they suggest, "We are also urged. ..to discover ways in which the protection of biodiversity ... can be thoroughly incorporated into the management of lands for a variety of uses and values". In other words, we need to ensure that the areas set aside for conservation sit in a matrix which is managed in a way which ensures the continued integrity of the reserves and provides some conservation benefit as well as productive outputs. This includes forests both on public and private land, especially where private holdings constitute a relatively large component of the conservation and production resource (e.g. Braithwaite et al. 1993). Attempts to develop this approach are being made in many different types of forest (e.g. Hansen et al. 1991; Caraher & Knapp 1995; Frumhaff 1995). Such a suggestion could be viewed as an attempt by those interested in conservation to grab as much of the forest as possible and prevent further productive use. On the other hand, it seems likely that continued productive output also strongly depends, on the maintenance of healthy, functioning forest ecosystems. The challenge is to find management regimes that optimize both conservation and production and retain the functionality of the ecosystem. Is this a pipe dream? A suggested approach to these challenges is what has been termed "ecosystem management". This approach to management tries to develop a holistic framework and move away from the fragmented and frequently contradictory practices conducted in parts of the forest managed for different goals. Ecosystem management recognizes that a variety of scales are important in management, from the individual site to the landscape and regional scale, and that management needs to be coordinated across these scales (Franklin 1993; Salwasser et al. 1993). Ecosystem management also recognizes a key set of ecosystem characteristics, outlined by Costanza (1992), Norton (1992) and Grumbine (1994) : 1. Dynamism. The classical idea of the "balance of nature" is being replaced by the concept of the "flux of nature"; i.e. ecosystems are constantly changing and should not be regarded as static en¬ tities (Botkin 1990; Pickett et al. 1992); 2. Relatedness. The "ECWEE" principle discussed above, and the need for cross-boundary manage¬ ment; 3. Hierarchy. The idea that natural systems and pro¬ cesses are nested, and the importance of manag¬ ing at the right scales and recognizing connections between scales; 4. Creativity and ecological integrity. Natural systems are self-organizing, and the processes which maintain this organization need to be maintained; and 5. Differential fragility. Systems vary in their resil¬ ience and thus have to be managed accordingly, as no one prescription will be suitable across a range of ecosystems. Gordon (1994) and Grumbine (1994) suggested the following principles of ecosystem management: 1. Manage where you are. Emphasis on site-specific properties, and the objectives of the management; 2. Manage with people in mind. Management needs to consider human desires, influences and responsi¬ bilities. Human values play a dominant role in determining management goals. In addition. Sample et al. (1993) suggest, "An ecosystem ap- 297 Journal of the Royal Society of Western Australia, 79(4), December 1996 proach must be not only ecologically sound but also economically viable and socially responsible" and "A focus on biophysical factors , with little or no consideration of social and economic needs, is doomed from the start". Magerum & Born (1995) also point to interaction with stakeholders and the public as a key operational component for integrated management; 3. Manage across boundaries. The recognition and management of neighbouring influences, and in¬ tegration of management goals through inter¬ agency cooperation. Decision support systems and allocation modelling procedures can be used to facilitate this ( e.g . Ive & Cocks 1989; Kilgour & Lau 1994) 4. Manage based on mechanisms rather than " rules of thumb". Use existing knowledge on processes and interactions, and seek to improve this knowledge. Assume that current knowledge is provisional and be prepared to adapt management practices in the light of new information; and 5. Manage without externalities. Include all known components and interactions when management decisions are made. The goal of ecosystem management, according to Gordon (1994), is a "sustained forest", which exhibits a "full range of characteristics and organisms, not just a sustained supply of wood". Add .ing this will not neces¬ sarily be easy. There will undoubtedly be problems with definitions and operationalising these definitions (e.g. Burroughs & Clark 1995). The whole concept of ecosystem management requires detailed knowledge of ecosystem processes and often, as discussed above, this knowledge is lacking. This lack of knowledge cannot, however, be used as an excuse or rationale for inaction or for ceasing current practices. Rather, management practices need to be adaptive and experimental, and adequately monitored. Changes in practice should be implimented as more and better information becomes available. However, current management philosophies and structures are not necessarily malleable enough to incorporate the necessary changes in approach. For instance. Sample et al. (1993) have suggested that "Another challenge is the reorganization of resource¬ managing organizations, both public and private, away from function-based, target oriented hierarchies towards open organizations conducive to multidisciplinary approaches to achieving desired future resource conditions". This problem has been discussed more generally by Holling (1995), who suggests that "The very success of managing a target variable for sustained production of food or fiber apparently leads inevitably to an ultimate pathology of less resilient and more vulnerable ecosystems, more rigid and unresponsive management agencies, and more dependent societies" He sees the underlying causes of this to be the prevalence of a single target and a piece-meal policy, a single scale of focus (typically on the short term and the local), no realization that all policies are experimental, and rigid management with no priority to design interventions as ways to test hypotheses underlying policies. An obvious corollary is that the way to deal with the problem is to reverse these causes, and to develop an integrated cross-scale management policy and an adaptive management strategy which monitors the impacts of management activities and modifies the regime where necessary (Gunderson et al. 1995). Conclusions The selection and design of nature reserves and reserve systems is but one part of the strategy required for adequate protection and maintenance of biodiversity in forest ecosystems (or any other type of ecosystem). Reserves sit within a matrix of lands managed for purposes other than nature conservation. While they may be treated as separate* legal or administrative entities, they are not separate in ecosystem terms. Ecosystem flows ensure that reserve systems are con¬ nected to the surrounding matrix, and the integrity of the reserves is in large measure dependent on what hap¬ pens in that matrix. The best reserve system in the world will not do what it is supposed to do in the long term if the surrounding matrix is degrading. The surrounding matrix of production lands thus plays a part in main¬ taining the biodiversity of a region. We thus need to move away from a piece-meal approach to management in which conservation and production management are considered separately. Management goals are now much more complex than previously and aim to maintain not only a sustainable harvest of timber, but also the struc¬ ture and complexity of the forest ecosystem. The con¬ cept of ecosystem management aims to integrate the various management goals and allow production to take place in such a way that both the long term productive potential and the biodiversity of the forest are main¬ tained. The challenge is to make the concept operational and to convince everyone involved of the urgent need to do so. Acknowledgments: I thank B Main, D Spratt, R Wills and two referees for constructive comments on the draft manuscript. Part of this paper appeared previously in the Proceedings of the Symposium "Professional Forestry: Evolving Issues in Forest Management", 1996, School of Forestry, University of Christchurch, New Zealand. References Abbott I & Christensen P 1994 Application of ecological and evolutionary principles to forest management in Western Australia. Australian Forestry 57:109-122. Aber J D & Melillo J M 1991 Terrestrial Ecosystems. Saunders College Publishing, Philadelphia. Anon 1977 Woodchips and the Environment. Senate Standing Committee on Science and the Environment. Australian Gov¬ ernment Printing Service, Canberra. Anon 1990 Forestry Research: A Mandate for Change. National Research Council. National Academy Press, Washington, DC. Anon 1993 Ecological Impacts of Forest Use: A Survey of Completed Research. Resource Assessment Commission Research Paper 9. Australian Government Printing Service, Canberra. Anon 1994 Forest Management Plan 1994-2003. Department of Conservation and Land Management, Perth. Attiwill P M 1994a The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management 63:247-300. Attiwill P M 1994b Ecological disturbance and the conservative management of eucalypt forests in Australia. Forest Ecology and Management 63:301-346. 298 Journal of the Royal Society of Western Australia, 79(4), December 1996 Attiwill P M 1995 Managing Leadbeater's possum in the mountain ash forests of Victoria - reply. Forest Ecology and Management 74:233-237. Bari M A, Smith N J, Boyd D W & Ruprecht J K 1994 Generation of streamflow following clearfell logging and regeneration at March Road catchment. Report WS119. Water Authority of Western Australia, Perth. Borg H, King P D & Loh I C 1987 Stream and ground water response to logging and subsequent regeneration in the southern forest of Western Australia. Interim results from paired catchment studies. Surface Water Branch, Water Authority of Western Australia. Report WH 34. Water Authority of Western Australia, Perth. Bormann F H & Likens G E 1981 Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York. Botkin D B 1990 Discordant Harmonies. A New Ecology for the Twenty-first Century. Oxford University Press, Oxford. Braithwaite W, Belbin L, Ive J & Austin M 1993 Landuse allocation and biological conservation in the Batemans Bay forests of New South Wales. Australian Forestry 56:4-21. Burbidge A A & McKenzie N L 1989 Patterns in the modern decline of Western Australia's vertebrate fauna: causes and conservation implications. Biological Conservation 50:143- 198. Burroughs R H & Clark T W 1995 Ecosystem management: a comparison of Greater Yellowstone and Georges Bank. Envi¬ ronmental Management 19:649-663. Burrows N D, Ward B & Robinson A D 1995 Jarrah forest fire history from stem analysis and anthropological evidence. Australian Forestry 58:7-16. Calver M C, Hobbs R J, Horwitz P & Main A R 1996 Science, principles and forest management: a response to Abbott and Christensen. Australian Forestry 59:1-6. Caraher D & Knapp W H 1995 Assessing ecosystem health in the Blue Mountains. In: Silviculture: From the Cradle of Forestry to Ecosystem Management. Southeast Forest Experiment Station General Technical Report SE-88. United States Department of Agriculture, Forest Service, Hendersonville, North Carolina, 34-41. Costanza R 1992 Toward an operational definition of ecosystem health. In: Ecosystem Health: New Goals for Environmental Management (ed R Costanza, B G Norton & B D Haskell). Island Press, Washington, DC, 239-256. D'Antonio C M & Vitousek P M 1992 Biological invasions by exotic grasses: the grass/ fire cycle, and global change. Annual Review of Ecology and Systematics 23:68-87. Davidson E M 1994 Role of environment in dieback of jarrah: effects of waterlogging on jarrah and Phytophthora cinnamomi and infection of jarrah by P. cinnamomi. Journal of the Royal Society of Western Australia 77:123-126. Davis G W & Richardson D M 1995 Mediterranean-type Ecosystems: The Function of Biodiversity. Springer, New York. Dell B, Havel J J & Malajczuk N, ed 1989 The Jarrah Forest: A Complex Mediterranean Ecosystem. Kluwer, Dordrecht. Dell B, Hopkins A J M & Lamont B B 1986 Resilience in Mediterranean-type Ecosystems. Junk, The Hague. Dell B & Malajczuk N 1989 Jarrah dieback - a disease caused by Phytopthora cinammomi . In: The Jarrah Forest: A Complex Mediterranean Ecosystem (ed B Dell, J J Havel & N Malajczuk). Kluwer, Dordrecht, 67-87. DellaSala D A, Olson D M, Barth S E, Crane S L & Primm S A 1995 Forest health: moving beyond rhetoric to restore healthy landscapes in the inland Northwest. Wildlife Society Bulletin 23:346-356. Duncan F & Packham J 1994 Forest Botany Manual. Nature Conservation Region 9. Forestry Commission, Hobart. Franklin J F 1993 The fundamentals of ecosystem management with applications in the Pacific Northwest. In: Defining Sustainable Forestry (ed G H Aplet, N Johnson, J T Olson & V A Sample). Island Press, Washington DC, 127-144. Friend J A 1990 The numbat Myrmecobius fasciatus (Myrmecobiidae): History of decline and potential for recovery. Proceedings of the Ecological Society of Australia 16:369-377. Frumhaff P C 1995 Conserving wildlife in tropical forests managed for timber: to provide a more viable complement to protected areas. BioScience 45:456-464. Gordon J C 1994 From vision to policy: a role for foresters. Journal of Forestry 92:16-19. Grumbine R E 1994 What is ecosystem management? Conservation Biology 8:27-38. Gunderson L H, Holling C S & Light S S 1995 Barriers broken and bridges built: a synthesis. In: Barriers and Bridges to the Renewal of Ecosystems and Institutions (eds L H Gunderson, C S Holling & S S Light). Columbia University Press, New York, 489-532. Hamilton S D, Lawrie A C, Hopmans P & Leonard B V 1991 Effects of fuel-reduction burning on a Eucalyptus obliqua forest ecosystem in Victoria. Australian Journal of Botany 39:203- 217. Hansen A J, Spies T A, Swanson F J & Ohmann J L 1991 Conserving biodiversity in managed forests: lessons from natural forests. BioScience 41:382-392. Hickey J E 1994 A floristic comparison of vascular species in Tasmanian oldgrowth mixed forest with regeneration resulting from logging and wildfire. Australian Journal of Botany 42:383-404. Hobbs R J 1993a Dynamics of weed invasion: implications for control. In: Proceedings of the 10th Australian Weeds Conference and 14th Asian Pacific Weed Society Conference. Vol. 1. Weed Society of Queensland, Brisbane, 461-465. Hobbs R J 1993b Effects of landscape fragmentation on ecosystem processes in the Western Australian wheatbelt. Biological Conservation 64:193-201. Hobbs R J 1994 Fragmentation in the wheatbelt of Western Australia: landscape scale problems and solutions. In: Fragmentation in Agricultural Landscapes (ed J Dover). I ALE, Garstang, UK, 3-20. Hobbs R J, Groves R H, Hopper S D, Lambeck R J, Lamont B B, Lavorel S, Main A R, Majer J D & Saunders D A 1995 Function of biodiversity in the mediterranean-type ecosystems of southwestern Australia. In: Mediterranean-type Ecosystems: The Function of Biodiversity (eds G W Davis & D M Richardson). Springer, New York, 233-284. Hobbs R J & Humphries S E 1995 An integrated approcah to the ecology and management of plant invasions. Conservation Biology 9:761-770. Holling C S 1995 What barriers? What bridges? In: Barriers and Bridges to the Renewal of Ecosystems and Institutions (ed L H Gunderson, C S Holling & S S Light). Columbia University Press, New York, 3-34. Humphries S E 1993 Environmental impact of weeds. In: Proceedings of the 10th Australian Weeds Conference and 14th Asian Pacific Weed Society Conference. Vol 2. Weed Society of Queensland, Brisbane, 1-11. Humphries S E, Groves R H & Mitchell D S 1991 Plant invasions of Australian ecosystems. A status review and management directions. In: Kowari 2. Plant invasions. The incidence of environmental weeds in Australia. Australian National Parks and Wildlife Service, Canberra, 1-127. Ive J R & Cocks K D 1989 Incorporating multiparty preferences into land-use planning. Environment and Planning B: Planning and Design 16:99-109. Jackson J & Taylor R 1994 Threatened Fauna Manual for Production Forests in Tasmania. Forestry Commission, Hobart. Jones C G & Lawton J H 1995 Linking Species and Ecosystems. Chapman and Hall, New York. 299 Journal of the Royal Society of Western Australia, 79(4), December 1996 Keighery G J 1995 How many weeds? In: Invasive weeds and regenerating ecosystems in Western Australia (ed G Burke). Institute for Science and Technology Policy, Murdoch University, Perth, Kilgour B & Lau A 1994 Decision support systems for Victorian forests - future durections using GIS. In: Resource Technology '94 Conference Proceedings: New Opportunities - Best Practice (ed I Bishop). University of Melbourne, Parkville, 199-206. Kinnear J E, Onus M L & Bromilow R N 1988 Fox control and rock wallaby population dynamics. Australian Wildlife Research 15:435-450. Kruger F J, Mitchell D T & Jarvis J U M 1983 Mediterranean-type Ecosystems: The Role of Nutrients. Springer, New York. Lamont B B 1995 Interdependence of woody plants, higher fungi and small marsupials in the context of fire. CALMScience Supplement 4:151-158. Lamont B B, Ralph C S & Christensen P E 1985 Mycophagous marsupials as dispersal agents for ectomycorrhizal fungi on Eucalyptus calophylla and Gastrolobium bilobum. New Phytologist 101:651-656. Lindenmayer D B 1995 Forest disturbance, forest wildlife conservation and the conservative basis for forest management in the mountain ash forests of Victoria - comment. Forest Ecology and Management 74:223-231. Magerum R D & Born S M 1995 Integrated environmental man¬ agement: moving from theory to practice. Journal of Environ¬ mental Planning and Management 38:371-391 . Majer J D & Abbott I 1989 Invert* b rates of the jarrah forest. In: The Jarrah Forest: A Complex Mediterranean Ecosystem (ed B Dell, J J Havel & N Malajczuk). Kluwer, Dordrecht, 111-122. Mawson P R Long I L 1994 Siz- and age parameters of nest trees used by four species of parrot and one species of cockatoo in south-west Australia. Emu 94:149-155. McCaw W L 1993 Effects of fuel-reduction burning on a Eucalyptus obliqua forest ecosystem in Victoria. Comment and Response. Australian Journal of Botany 41:413-414. McCaw W L & Burrows N D 1989 Fire management. In: The Jarrah Forest: A Complex Mediterranean Ecosystem (ed B Dell, J J Havel & N Malajczuk). Kluwer, Dordrecht, 317-334. McGrath M 1985 Fire planning and management: an overview. In: Fire Ecology and Management in Western Australian Ecosystems (ed J W Ford). WAIT Environmental Studies Group, Perth, Western Australia, 219-221. Newmark W D 1985 Legal and biotic boundaries of western North American national parks, a problem of congruence. Biological Conservation 33:197-208. Noble J C 1993 Relict surface-soil features on semi-arid mulga (Acacia aneura) woodlands. Rangelands Journal 15:48-70. Norton B G 1992 A new paradigm for environmental management. In: Ecosystem Health: New Goals for Environmental Management (ed R Costanza, B G Norton & B D Haskell). Island Press, Washington, DC, 23-41. Oppenheimer M 1995 Context, connection and opportunity in environmental problem solving. Environment 37:10-15, 34. Parsons W T & Cuthbertson E G 1992 Noxious weeds of Australia. Inkata Press, Melbourne. Pickett S T A, Kolasa J & Jones C G 1994 Ecological Understanding: The Nature of Theory and the Theory of Nature. Academic Press, New York. Pickett S T A, Parker V T & Fielder P 1992 The new paradigm in ecology: implications for conservation biology above the species level. In: Conservation Biology: The Theory and Practice of Nature Conservation (ed P Fielder & S Jain). Chapman and Hall, New York, 65-88. Rab M A 1994 Changes in physical properties of a soil associated with logging of Eucalyptus regtians forest in southeastern Australia. Forest Ecology and Management 70:215-229. Richardson D M, Williams P A & Hobbs R J 1994 Pine invasions in the Southern Hemisphere: determinants of spread and invasibility. Journal of Biogeography 21:511-52 7. Roberts N 1994 The global environmental future. In: The Changing Global Environment (ed N Roberts). Blackwell, Oxford, 3-21. Salwasser H, MacCleary D W & Snellgrove T A 1993 An ecosystem perspective on sustainable forestry and new directions for the U.S. National Forest system. In: Defining Sustainable Forestry (ed G H Aplet, N Johnson, J T Olson & V A Sample). Island Press, Washington DC, 44-89. Sample V A, Johnson N, Aplet G H & Olson J T 1993 Introduction. Defining Sustainable Forestry. In: Defining Sustainable Forestry (ed G H Aplet, N Johnson, J T Olson & V A Sample). Island Press, Washington DC, 3-15. Saunders D A, Hobbs R J & Margules C R 1991 Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32. Schofield N J, Stoneman G L & Loh I C 1989 Hydrology of the jarrah forest. In: The Jarrah Forest: A Complex Mediterranean Ecosystem (ed B Dell, J J Havel & N Malajczuk). Kluwer, Dordrecht, 179-201. Stoneman G L 1993 Hydrological response to thinning a small jarrah (Eucalyptus marginata) forest catchment. Journal of Hydrology 150:393-407. Taylor R J 1991 Fauna Conservation in Production Forests in Tasmania. Forestry Commission, Hobart. Taylor R J & Haseler M E 1995 Effects of partial logging systems on bird assemblages in Tasmania. Forest Ecology and Management 72:131-149. Tingay A 1985 Contemporary views of the voluntary conservation movement on the use of fuel reduction burns as a land management technique. In: Fire Ecology and Management in Western Australian Ecosystems (ed J W Ford). WAIT Environmental Studies Group, Perth, 215-218. Underwood R J, Sneeuwjagt R J & Styles H G 1985 The contribution of prescribed burning to forest fire control in Western Australia. In: Fire Ecology and Management in Western Australian Ecosystems (ed J W Ford). WAIT Environmental Studies Group, Perth, 153-170. Waring R H & Schlesinger W H 1985 Forest Ecosystems: Concepts and Management. Academic Press, New York. Williams J E & Gill A M 1995 The impact of fire regimes on native forests in Eastern New South Wales. New South Wales National Parks and Wildlife Service, Hurtsville, NSW. Withers P C, Cowling W A & Wills R T 1994 Plant Diseases in Ecosystems; Threats and Impacts in South-western Australia. Journal of the Royal Society of Western Australia Volume 77(4). Perth, Royal Society of Western Australia. Wood W E 1924 Increase of salt in soil and streams following destruction of native vegetation. Journal of the Royal Society of Western Australia 10:35-47. 300 Journal of the Royal Society of Western Australia, 79:301-304, 1996 Forest reservations: an overview A R Main Zoology Department, University of Western Australia, Nedlands WA 6907 Abstract This paper is placed in the context of the assumptions made when suggesting criteria for selecting areas for reservation. It is suggested that, in the absence of detailed taxonomic knowledge, site selection might be improved by recognising the possibility that relict occurrences of past environments will include equally ancient biotic assemblages. Moreover, the reasons for the persistence of such past environments offer guide-lines for management which will ensure their retention. Introduction When I was first asked to provide an overview for this Symposium the proposed title began "Requirements for Reservation....", the latest version is " The design of Reservations". It is easy to see why the title has been changed; to require is to insist upon having. In a biologi¬ cal sense it is broadly known what is required of reser¬ vations. But requirement also carries a connotation of asking or claiming by right or authority. So, the latter title is less likely to be misconstrued because design, as a scheme or plan, purpose or intention, is less coloured. The other papers in this issue have embraced these subtleties. In this overview I wish to emphasise the importance of including historically significant biologi¬ cal elements as a requirement of adequate design for reservations, allude to scientific procedures so that non¬ scientists may understand why interpretations change, and show that historical knowledge of the biota offers a guide for management. The Commonwealth has proposed a set of criteria for National Forests Conservation Reserves defining forests as woody vegetation with a potential height >5m (Anon 1995:41). In the proposed criteria it is frequently stated that there is a need to emphasise national estate values. Moreover, "our knowledge of forest species is limited , large numbers of invertebrate species are reasonably presumed to be undiscovered and/or undescribed" (Anon 1995:24). It is proposed that this deficiency can be overcome by identifying habitats as surrogates for these values. This is in contrast to various studies which suggest that plants and vertebrates are such surrogates, while the Commonwealth position is that consultation and 15% of the original or pre-European extent of forest present will adequately reserve forest communities. A common approach to the problem is to accept the present as given, then after survey and classification, select areas for reservation i.e. assume tomorrow will be like today. Yet other perspectives and procedures on these topics are in terms of whether reserves are ad¬ equate, comprehensive and connected. Another possibility is as follows; some of the National Estate Symposium on the Design of Reserves for Nature Conservation in South-western Australia © Royal Society of Western Australia 1996 values reside in historic elements of the biota which can be readily recognised. Their occurrence is tied to historical /geological events and land forms resulting from them. Their origins go back to the time when there was only one great southern continent, the land mass known as Gondwana. The plants and animals which survived from this time contribute much to the distinctive character of the biota and sets aside the bioregion as being unique i.e. having National Estate values. The aquatic and wetland habitats favoured by animals are readily recognised and thus present themselves as potential surrogates when identifying areas for reservation. Goals As a minimum we wish to retain replicated, repre¬ sentative areas of natural biodiversity so that those who come after us may know what the present day world was like. While we are interested in retaining biodiversity, usually conceived as species or genetic richness, within reservations, we need to ask the ques¬ tions: Is this goal reasonable? Can it be achieved? We cannot foresee the future and can only know the present, so does this give a sound basis for choosing areas or expecting success? A central question which arises with the goal of conserving through reservation is: will tomorrow (the future) be like today (the present)? i.e. will what we know and cherish about nature be able to persist tomorrow? Common sense tells us that today is like yesterday or at least not too different and so common sense suggests that tomorrow will be like today. But in science, common sense is not taken as a good guide; rather, it is a conjecture which must then be subject to rigorous tests to show whether these conjectures are true. Only when repeated attempts at refutation have failed can any faith be placed in the conjectures and the interpretations that follow from them. The world is neither static nor stable and nature persists in a non-equililbrium state i.e. selection and ge¬ netic composition of populations or the populations composing communities are the result of past selection and temporary assemblages arising from transient cir¬ cumstances. But this happened in the past also, so we might conclude that tomorrow will be like today. Espe- 301 Journal of the Royal Society of Western Australia, 79(4), December 1996 dally so because the same classes of biological interac¬ tions will occur, methods of interaction will be similar but rates will be different and the resource base reduced unless we are careful, so tomorrow may not be like to¬ day. The foregoing paints a picture of great uncertainty and is the central issue facing us when selecting areas for reservation and managing them. Management can address uncertainties by adopting a system of adaptive management as advocated. Landscapes rather than re¬ serves might be the management unit. But our concepts of reserve design and adequacy are pitifully incomplete because of the poor and very slowly expanding knowl¬ edge base. Additionally, there is the question of whether 15% (or any other arbitrary percentage) of an area can cover the foregoing uncertainties. Ideally, selection of areas for reservation would fol¬ low comprehensive surveys. These are time-consuming, expensive and hindered by inadequate taxonomic exper¬ tise. Moreover, the results might not be timely in the sense that for various reasons decisions need to be made before the survey can be completed. So what is a prudent course of action? The more particular and specific are the goals the greater the likelihood of misjudgement i.e. specific goals require de¬ tailed knowledge, but this takes an inordinately long time to acquire. However, generality based on understandable principles relating to biological needs of animals of gondwanan origin might be a way of selecting areas. Moreover, the general principles might be confirmed from a knowledge of what happened following changes in the past and might incidentally give guidance for management. History Australia is a fragment of a former super continent, Gondwana. Since it broke away some 60 mybp, Austra¬ lia has traversed 35-40 degrees of latitude and has come from a region with a cool damp climate to the unreliable climate with seasonal drought characteristic of its present low latitude situation. So today is not like yesterday. The last major sculpting of the western landscape occurred during the Permian glaciation 250 mybp. The subdued landscape that resulted from the glaciation has dominated stream flows and erosion patterns ever since. The current system of salt lakes for example are the relicts of early Tertiary rivers and these in turn may represent Mesozoic drainage channels. While the drainage channels show their origin they also show the influence of more recent events, such as the sagging of the southern margin of the Western Australian part of the Australian plate as it fragmented from Antarctica and the uplift of the western margin along the Darling Fault and Meckering line. These events changed the direction of streams affected by the tectonic events to either the west or south (Hocking & Cockbain 1990). The nett result was a new cycle of erosion with the head waters of these streams, in some cases, cutting back into the streams of the plateau. More recently, rain¬ fall has become seasonal, and evaporaton higher result¬ ing in the accumulation of cyclic salt in the soil profiles in areas of lower rainfall (Ghassemi et al. 1995: 155, 180). From the taxonomic literature it is possible to identify many invertebrate groups which have Gondwanan affinities occur in South East Australia, New Caledonia, New Zealand, South America or South Africa (Main & Main 1991; Hopper et al. 1996; Table 4). A field inspec¬ tion of the localities from which animals have been re¬ ported reveals the easily recognised environmental char¬ acteristics where habitat favourable for invertebrate ani¬ mals with Gondwanan affinities may be found. Such sites are damp or wet, often with impeded drainage and showing little or no signs of salinisation. Such sites can be considered to resemble or be representative of habi¬ tats which were more widespread in earlier times. Thus with respect to the sites and habitats identified as con¬ taining gondwanic elements today is like yesterday and such sites, even without detailed surveys for reservation are likely to retain significant Gondwanan biota The present Surface characteristics, drainage, topography and rainfall combine to determine the quality of any water within a catchment. Under high rainfall conditions, soils accumulate little salt while in low rainfall areas salt accumulates in the soil profile and is released as ground water discharge when the hydrological balance is disturbed e.g . following clearing. Saline ground waters are common in the semi-stripped etch plain to the north and east of the forest region and absent in the high rainfall areas to the west and south of the region. Thus three sorts of river water occurs, saline in those streams whose head waters drain the semi-stripped etch plain; streams which are saline in the head waters but fresh in the lower reaches where flow from fresh tributaries dilutes the saline waters from the upper reaches; and those, usually short streams, draining only high rainfall areas. The erosion by the western and southern streams has resulted in the present etched landscape of high topographic relief so characteristic of the high rainfall areas, particularly along the Darling Scarp and in the vicinity of Manjimup and Pemberton (Finkl & Churchward 1973). This new landscape contrasts markedly with the former subdued Gondwanan topography. To summarise the changes to the present; the latitude of the continent has changed, climate has altered and become more seasonal, and much of the old surface has been destroyed by erosion. To refer back to the intro¬ ductory analogy, today is not like yesterday. This would suggest simply that biotic relicts of ancient times will not have survived to the present! Yet collection of biotic material quickly shows that this is not so. Podocarpus is an obvious example; it has a long geological history from at least early Tertiary times and still occurs in other fragments of the former Gondwana. But there are countless other plant genera with equally ancient lin¬ eages e.g . Xylomclum, Adenanthos, Banksia and other Proteacea. Many terrestrial invertebrates show their Gondwanan origins through their affinities and relation¬ ships to elements of the other southern continents. So, despite the changes noted above there must be some places which are still like yesterday. Some of the sites can be characterised and thus offer a good guide for 302 Journal of the Royal Society of Western Australia, 79(4), December 1996 selection of areas having a high likelihood of containing Gondwanan elements and thus capable of retaining this element of biodiversity. Moreover, some of the unique vertebrate elements also occur in similar or the same sites e.g. freshwater fish (Hopper et al. 1996; Table 4). The low topography of the old Gondwanan surface and in places on the incipient etch plain results in poor drainage, ill defined water courses, and extensive areas of swamp land which is either permanentely wet or variously referred to as winter-wet or summer-dry swamps. These swamp lands provide one of the modern habitats which retain Gondwana-like characteristics. In drier sites invertebrate elements with gondwanic affinities have adapted by modifiying life histories e.g. aestivating or behaviourablly by burrowing. The chain of soil types (the catenas) of the valley sides, even in saline areas often have a series of summer-dry swamps each with its own characteristic fauna and flora. Better drained areas permit the growth of jarrah and karri which provide a closed canopy and thus another Gondwana-like habitat. Thus each land surface has its own hydrological regime and the potential to retain different Gondwanan elements. The retention of whole catenas with perched swamps and impeded ground water flow extending from granite crests to valley floors is important. Selection of Gondwanan sites So, what are the characteristics which could be used to identify sites for reservation so that Gondwanan elements might be conserved (Main & Main 1991; Hop¬ per et al. 1996)? They are; 1 unaffected by salinisation; 2 high rainfall areas with short summer drought; 3 topographically high south coastal areas subject to frequent mists, cloud and drizzle; 4 areas adjacent to granite rocks from which water is shed; 5 areas of impeded ground water flow so produc¬ ing winter-wet swamps; 6 streams with extensive fresh head water swamps and year round flow; 7 areas where vegetation can harvest water from fog or cloud by drip from leaves and stem flow e.g. tingle forest and south coast dunes and heath; 8 areas with southern or south-western aspect which are thus sheltered from summer insolation e.g. valley slopes and wet valley floors; and 9 areas of intact forest canopy under which the characteristic under storey shrubs and herbs oc¬ cur. Endemic elements The foregoing are not the only criteria for selecting elements of the biota for conservation. Numerous plants and animals have evolved since Gondwanan times. Such species will be found in sites whose characteristics have developed more recentlly i.e. are drier, better drained or more exposed, eucalypts and acacias are perhaps the most successful of these more recently evolved types. But, any sites where evolution and selection might be expected to produce unique endemic forms because they are different to 'typical' Gondwanan land forms should be considered when surveying for conservation sites e.g. the twig-lining habit of the spider Aganippe raphidiica (B Y Main 1976). Thus in selecting sites for reservation the following should be considered in addition to those in which old Gondwanan elements may occur; a. new topographic elements to which Gondwanan forms may have evolved adaptations and thus be¬ come unique endemic elements of the biota; and b. sites where Tertiary and more recent migrants to the south-western part of the continent have evolved to be typical elements of certain commu¬ nities or ecosystems. Implications A goal of retaining Gondwanan elements has impli¬ cations for management. The fossil record shows a long history of fire in Australia. However, many Gondwanan elements, especially wet-land forms, are more sensitive than later evolved forms to fire and so in site selection and management this factor needs to be considered. Field inspection of sites from which biotic elements with Gondwanan affinities have been collected suggest that in addition to fire, drainage, harvesting or removal of vegetation cover and salinisation are all inimical so management plans that recognise the need to retain spe¬ cial elements and the dangers to their persistance have the maximum chance of ensuring that tomorrow (the future) will be like today (the present). Recent developments demonstrate the readiness with which the tourism industry will embrace new reservations as a resource to be exploited. Such use has the potential to destroy the qualities which justified the initial reservation. Such an outcome can only be negated by detailed expensive management, or by having numerous large areas reserved, within which there is an adequate and significant replication of historically important habitats. Nevertheless it should always be borne in mind that reservations will only be part of a mosaic of land uses whether they be within native forests, plantation forests of eucalypts or conifers, or wheatfields. Their health and persistence will be inextricably linked to the whole matrix of reservations, how isolated the biota of each reservation may be and how the management of the non-reserved portion of the landscape impinges on the long term viability of the reservations i.e. the landscape rather than the individual reservation is the management unit. This last point is important because large size or an adequate surrounding buffer when areas are selected for reservation may be the only practical way of ensuring their long term viability. Conclusions The common practice when selecting areas for con¬ servation reserves is to interpret the broad patterns of 303 Journal of the Royal Society of Western Australia, 79(4), December 1996 the current situation without any time dimension. A fur¬ ther assumption is that what is present now will persist in the future despite interference and disturbance i.e. the future will be like the present. Elsewhere (Main 1996) I have expressed doubts about ecosystem stability and suggested a dynamic state as being a more realistic as¬ sumption. Furthermore, there is a strong possibility that increased amounts of greenhouse gases in the atmo¬ sphere will result in climatic changes which may impact on this nature conservation estate (Main 1993). Despite these possibilities, ecosystems are in a dynamic state on a trajectory determined by biological responses to envi¬ ronmental changes set in train as the continent moved from high latitude moist equable climates to warmer drier more seasonal ones. In this process the cool cli¬ mate moisture dependent elements have been restricted to refugia which are usually small scale, long undis¬ turbed, habitats which are easily overlooked in large scale forest harvesting. When selecting areas for reserva¬ tion to retain biodiversity we should realize that despite an absence of detailed taxonomic knowledge of the cool temperate Gondwanan elements of the biota we do have a basis for selecting reservations so that the possibility of retaining this element of the biodiversity is maximised. References Anon. 1995 National Forest Conservation Reserves. Common¬ wealth Proposed Criteria. A discussion Paper. Canberra. Finkl C W & Churchward H M 1973 The etched land surfaces of southwestern Australia. Journal of Geological Society of Aus¬ tralia 20:295-308. Ghassemi F Jakeman A J & Nix H A 1995 Salinisation of Land and Water Resources: Human Causes, Extent, Management and Case Studies. University of New South Wales Press, Sydney. Hocking R M & Cockbain A E 1990 The regolith. In: Geology and Mineral Resources of Western Australia. Geological Survey, Memoir 3:590-602. Hopper S D, Harvey M S, Chappill J A, Main A R & Main B Y 1996 The Western Australian biota as Gondwanan heritage - a review. In: Gonwanan Heritage: Past, Present and Future of the Western Australian Biota (ed S D Hopper, J A Chappill, M Harvey & A George). Surrey Beatty & Sons, Chipping Norton, 1-46. Main A R 1993 Restoration ecology and climatic change. In: Na¬ ture Conservation 3: Restoration of Fragmented Ecosystems (ed D A Saunders, R J Hobbs & P R Ehrlich). Surrey Beatty & Sons, Chipping Norton, 27-32. Main A R 1996 Keynote address: conservation. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (ed S D Hopper, J Chappill, M Harvey & A S George), Surrey Beatty & Sons, Chipping Norton, 104-108. Main B Y 1976 Spiders William Collins, Sydney. Main A R & Main B Y 1991 Report on the Southern Forest Region of Western Australia. Report. The Australian Heritage Commission, Canberra. 304 Journal of the Royal Society of Western Australia m tmm or vktowa ubrwy CONTENTS Mini-Symposium on Satellite Remote Sensing and its Applications in Western Australia Page Preface P Withers i Applications of satellite remote sensing to the marine environment in Western Australia A F Pearce & C Pattiaratchi 1-14 Applications of satellite remote sensing for mapping and monitoring land surface processes in Western Australia R C G Smith 15-28 A bibliography of research into satellite remote sensing of land, sea and atmosphere conducted in Western Australia R C G Smith & AF Pearce 29-39 Volume 80 Part 1 March 1997 Museum of Victoria 43646 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia Patron Her Majesty the Queen Vice-Patron His Excellency Major General Michael Jeffery AD MC Governor of Western Australia COUNCIL 1996-1997 President M G K Jones MA PhD Immediate Past President S Hopper BSc (Hons) PhD Senior Vice-President H Recher BSc PhD Junior Vice-President A George BA Hon Secretaries P Gardner BEng (Hons) GDipCSci DipEd V Hobbs BSc (Hons) PhD P Lavery BSc (Hons) PhD Hon Treasurer R Froend BSc (Hons) PhD Hon Editor P C Withers BSc (Hons) PhD Hon Journal Manager J E O'Shea BSc (Hons) PhD Hon Librarian M A Triffitt BA ALIA Members W A Cowling BAgricSci (Hons) PhD J Dodd BA Msc PhD D Gordon BSc (Hons) PhD K Rosman BSc (Hons) PhD V Semeniuk BSc (Hons) PhD L N Thomas BSc MSc PGDipEIA G G Thompson MEd PhD The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March - December) at 8 pm Kings Park Board offices. Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1996/1997 financial year are $40 for ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the membership Secretary, % W A Museum, Francis Street, Perth, WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The Journal is indexed and abstracted internationally. Cover design: Mangles' kangaroo paw ( Anigozanthos tnanglesii) and the numbat ( Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences ambraced by the Royal Society of Western Australia. (Artwork: Dr Jan Taylor). Journal of the Royal Society of Western Australia CONTENTS Page Recent Advances in Science in Western Australia 41 Diet of herbivorous marsupials in a Eucalyptus marginata forest 47 and their impact on the understorey vegetation KA Shepherd, GW Wardell-Johnson, WA Loneragan & DT Bell Dietary preferences of the black-gloved wallaby ( Macropus irma) 55 and the western grey kangaroo (M. fuliginosus) in Whiteman Park, Perth, Western Australia JM Wann & DT Bell Waychinicup Estuary, Western Australia: the influence of fresh- 63 water inputs on the benthic flora and fauna J Phillips & P Lavery Contributions of N H Speck to the biogeography of Proteaceae in 73 Western Australia N Gibson, GJ Keighery & B] Keighery Contents Volume 77 79 Contents Volume 78 81 Contents Volume 79 83 Volume 80 Part 2 June 1997 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia Patron Her Majesty the Queen Vice-Patron His Excellency Major General Michael Jeffery AD MC Governor of Western Australia COUNCIL 1996-1997 President M G K Jones MA PhD Immediate Past President S Hopper BSc (Hons) PhD Senior Vice-President H Recher BSc PhD Junior Vice-President A George BA Hon Secretaries P Gardner BEng (Hons) GDipCSci DipEd V Hobbs BSc (Hons) PhD P Lavery BSc (Hons) PhD Hon Treasurer R Froend BSc (Hons) PhD Hon Editor P C Withers BSc (Hons) PhD Hon Journal Manager J E O'Shea BSc (Hons) PhD Hon Librarian M A Triffitt BA ALIA Members W A Cowling BAgricSci (Hons) PhD J Dodd BA Msc PhD D Gordon BSc (Hons) PhD K Rosman BSc (Hons) PhD V Semeniuk BSc (Hons) PhD L N Thomas BSc MSc PGDipEIA G G Thompson MEd PhD The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March - December) at 8 pm Kings Park Board offices, Kings Park West Perth, WA 6005. Individual membership subscriptions for the 1996/1997 financial year are $40 for ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the membership Secretary, % W A Museum, Francis Street, Perth, WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The Journal is indexed and abstracted internationally. Cover design: Mangles' kangaroo paw (. Anigozanthos manglesii) and the numbat ( Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences ambraced by the Royal Society of Western Australia. (Artwork: Dr Jan Taylor). Journal of the Royal Society of Western Australia, 80(3), March 1997 Granite Outcrops Symposium September 14-15, 1996 sponsored by The Royal Society of Western Australia , The Gordon Reid Foundation for Conservation The University of Western Australia Page CONTENTS Preface S Hopper & P C Withers Granite Outcrops Registered Participants Geology of granite J S Myers Granite landforms E M Campbell Granite outcrops: A collective ecosystem B York Main Reproductive ecology of granite outcrop plants from the south-eastern United States R Wyatt The geobiological interface: Granitic outcrops as a selective force in mammalian evolution M A Mares Plants of Western Australian granite outcrops S D Hopper, A P Brown & N G Marchant Terrestrial fauna of granite outcrops in Western Australia P C Withers & D H Edward Invertebrates of temporary waters in gnammas on granite outcrops in Western Australia I A E Bayly Aboriginal people and granite domes P Bindon Water harvesting from granite outcrops in Western Australia I A F Laing & E J Hauck Management of granite rocks A R Main Geography, environment and flora of Mt Mulanje, Central Africa J Beard Inselberg vegetation and the biodiversity of granite outcrops S Porembski, R Seine & W Barthlott Remnant vegetation, priority flora and weed invasions at Yilliminning Rock P J Pigott & LW Sage Why is Sant alum spicatum common near granite rocks? JED Fox Rocks as museums of evolutionary processes J D Bussell & S H James The ant communities of Sanford Rock Nature Reserve, Westonia, Western Australia A I Doronilla & J E D Fox Acacia williamsiana (Fabaceae: Juliflorae): A new granitic outcrop species from northern New South Wales J T Hunter ii iii 87 101 113 123 131 141 159 167 173 181 185 189 193 201 209 221 231 235 i Journal of the Royal Society of Western Australia, 80(3), March 1997 Preface S D Hopper1 & P C Withers2 'Kings Park and Botanic Garden, West Perth WA 6005 -Department of Zoology, The University of Western Australia, Nedlands WA 6907 ^ “eQ °“tCrops SymP°sium was held at University of Western Australia on September 14 and lb, 19% It aimed to achieve a cross-disciplinary synthesis of current knowledge on granite outcrops, and to distill effective guidelines for conservation managers. About 15% of the earth's continents consists of granite, which may outcrop in the form of inselbergs ranges, ridges and isolated hills - that stand abruptly from surrounding terrain, like islands in a sea Most commonly, granite outcrops appear as dome-shaped hills with bare rock exposed over much of their surfaces. The water catchment so formed, combined with a diversity of microhabitats make granite outcrops havens for biodiversity world-wide. In Western Australia, granite outcrops form characteristic landforms across much of the gentlv undulating terrain of the State. These granite outcrops are of special interest: • geologically, being among the world's oldest rocks • hydrologically, providing a source of water in a dry landscape • biologically, as refuges rich in endemic plants and animals, and culturally, being vital for aboriginal, colonial and contemporary peoples alike. Despite their mtrmsic interest, world-wide occurrence, and special values in Western Australia, granite outcrops had yet to be the subject of a significant cross-disciplinary scientific meeting. For these reasons, the Royal Society of Western Australia decided to host a two-day symposium on granite outcrops, including a combination of invited reviews and contributed papers. Over 70 participants from local, national and international destinations contributed to a diverse and stimulating meeting. The symposium was significant in high-lighting ideas of global importance concerning the geology, landforms, biodiversity and human use of granite outcrops. Special research opportunities arise from the insularity and geographically compact nature of most outcrops. They offer model systems for exploring evolutionary and ecological processes, and for achieving conservation outcomes through the stewardship of informed local communities. Participants at the symposium resolved to meet again on publication of these proceedings at a management workshop aimed at helping conserve granite outcrop biodiversity and cultural heritage. In addition, at least two books proposed by invited reviewers are now well underway, receiving extra impetus from discussions at the 1996 symposium. These initiatives alone have affirmed the value and success of the Royal Society of Western Australia's initiative. Perusal of this volume highlights the diversity of scientific ideas and opportunities deserving future research on granite outcrops. We hope that the symposium will stimulate other contributions to a most deserving field of multi-disciplinary enquiry. As editors, we thank the Royal Society of Western Australia for hosting the symposium, the Gordon Reid Foundation for Conservation for a significant grant without which the symposium would not have been such a success, and The University of Western Australia for providing a suitable venue for the meetmg. Don Edward, our colleague on the organising committee, made an invaluable contribution for which we are grateful. ii Journal of the Royal Society m of Western Australia CONTENTS Page History and management of Culham Inlet, a coastal salt lake in south-western Australia. E P Hodgkin 239 Pre-contact human skeletal remains from Useless Loop, Western Australia. N G Jablonski & S Bowdler 249 Status of a shallow seagrass system, Geographe Bay, south¬ western Australia. K McMahon, E Young, S Montgomery, j Cosgrove, J Wilshaw & D Walker 255 Foraminifera from Exmouth Gulf, Western Australia. D W Haig 263 Abundance of arthropods in tree canopies of Banksia woodland on the Swan Coastal Plain. R A Tassone & ] D Majer 281 The Royal Society of Western Australia Medallists, 1997 E P Hodgkin & A J McComb 287 Obituaries: W H Cleverley, B J Grieve and C F H Jenkins 289 Al Volume 80 Part 4 December 1997 k ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia Patron Her Majesty the Queen Vice-Patron His Excellency Major General Michael Jeffery AD MC Governor of Western Australia COUNCIL 1997-1998 President M G K Jones MA PhD Immediate Past President S Hopper BSc (Hons) PhD Senior Vice-President H Recher BSc PhD Junior Vice-President A George BA Hon Secretaries P Gardner BEng (Hons) GDipCSci DipEd V Hobbs BSc (Hons) PhD M Lund BSc (Hons) PhD Hon Treasurer G G Thompson MEd PhD Hon Editor P C Withers BSc (Hons) PhD Hon Journal Manager J E O'Shea BSc (Hons) PhD Hon Librarian M A Triffitt BA ALIA Members L Broadhurst BSc (Hons) N Gibson BSc (Hons) PhD S Griffin BSc (Hons) M Harvey BSc (Hons) PhD L N Thomas BSc MSc PGDipEIA V Semeniuk BSc (Hons) PhD H Stace MSc PhD The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March - December) at 8 pm Kings Park Board offices, Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1998/1999 financial year are S40 for ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1998 calendar year. For membership forms, contact the membership Secretary, % W A Museum, Francis Street, Perth, WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The Journal is indexed and abstracted internationally. Cover design: Mangles' kangaroo paw ( Anigozanthos manglesii ) and the numbat (Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences ambraced by the Royal Society of Western Australia. (Artwork: Dr Jan Taylor). mma * vwo«« Journal of the Royal Society of Western Australia, 80:i, 1997 UmRY PREFACE Mini-Symposium on Satellite Remote Sensing and its Applications in Western Australia Satellite remote sensing is a rapidly-developing technology which has important roles in many aspects of science, industry and government. It is a particularly important technology in Western Australia with its considerable land mass, extensive coast-line and off-shore waters. Fortunately, the Leeuwin Centre for Earth Sensing Technologies at Floreat Park, and the affiliated institutions and researchers including CSIRO, DOLA, TAFE, Curtin University and World Geoscience Corporation, provide state-of-the-art facilities for local researchers. The Royal Society of Western Australia consequently sponsored a mini¬ symposium on aspects of satellite remote sensing with particular reference to its uses in Western Australia, by drawing upon the expertise of researchers at the Leeuwin Centre. The Mini-Symposium on Satellite Remote Sensing and its Applications in Western Australia was held as a monthly meeting of The Royal Society of Western Australia on Monday 18,h September, 1995, at the Leeuwin Centre for Earth Sensing Technologies. The meeting was convened by Dr Andy Gable (CSIRO Division of Exploration and Mining), who provided a brief overview and technical background to the different remote sensing technologies. Mr Alan Pearce (CSIRO Division of Oceanography) then presented several case studies concerning the uses of satellite remote sensing in the marine environment. Thermal satellite data provide sea-surface temperature imagery, which is being used in studies of the ocean currents off Western Australia and their role in recruitment to a number of local fisheries, including the western rock lobster, scallops in Shark Bay, and the salmon and pilchard fisheries along the south coast of Western Australia. The images also helped explain the movement of the oilslick from the disabled tanker Kirki in 1991. Use of medium resolution colour imagery for bathymetric and habitat studies in shallow coastal and bay waters, such as Broome, Shark Bay and Geographe Bay, and high resolution imagery for monitoring water quality in estuaries such as the Peel Harvey system, were described. Dr Richard Smith (WA Department of Land Administration) then discussed the uses of satellite remote sensing for land surface structures. Vegetation spectral properties for comparing different vegetation types over space and time can provide information on land productivity and weather events. For example, effects of land clearing on rainfall patterns, movements of flood waters, erosion, and bushfires can be monitored. Dr Ian Tapley (CSIRO Division of Exploration and Mining) described three applications of remote sensing in mineral exploration. SAR (synthetic aperture radar) sensing technology can provide information on surface topography and soil properties, and Landsat images reveal spectral properties of soil related to the surface regolith. Dr Gable then concluded the mini-symposium with a brief look at future technologies and their potential applications. This issue of The Journal of the Royal Society of Western Australia presents two papers based on this mini¬ symposium and provides a comprehensive bibliography of studies using satellite remote sensing technology with particular reference to Western Australia. The Royal Society of Western Australia thanks the contributors to the mini-symposium, and Alan Pearce and Richard Smith for preparing papers for this issue of the journal. Philip Withers Honorary Editor i Journal of the Royal Society of Western Australia, 80:1-14, 1997 Applications of satellite remote sensing to the marine environment in Western Australia A Pearce1 & C Pattiaratchi2 1 CSIRO Division of Marine Research, Marmion WA 6020: alan.pearce@marine.csiro.au 2 Centre for Water Research, University of Western Australia Nedlands, WA 6907: pattiara@cwr.uwa.edu.au Manuscript received February 1997 Abstract Satellite remote sensing is the only feasible means of monitoring on a regular basis the large expanse of oceanic surface waters off Western Australia, and marine scientists are increasingly applying this technology to studies of our coastal and ocean waters. Up to now, most use has been made of Landsat Multi-Spectral Scanner (MSS) and Thematic Mapper (TM) imagery for studies of estuarine and coastal features, and the NOAA Advanced Very High Resolution Radiometer (AVHRR) for sea-surface temperature and inferred ocean currents. Some preliminary work has also been undertaken on the chlorophyll distribution of our continental shelf waters using the Coastal Zone Colour Scanner (CZCS). Satellite remote sensing will continue to play an indispensible role in marine science with the availability of altimeters for directly monitoring sea- surface elevations (and hence ocean circulation), scatterometers for surface winds and waves and ocean colour sensors for studies of chlorophyll distributions and ocean productivity. Introduction Western Australia has a coastline of some 8230 km (excluding offshore islands, which increase the total coastline to about 12330 km; R Galloway personal com¬ munication) and the area of the continental shelf is about 400,000 km2 (Jarvis 1986). Because of the complexity of oceanic processes on a range of temporal and spatial scales as well as the expense and limited range of research vessels, the only effective way of monitoring such large areas on a regular basis is through remote sensing. For the past decade, marine scientists in Western Australia have effectively applied satellite remote sensing to study our ocean waters, using the imagery to complement oceanographic data collected by conventional (surface- based) methods and to obtain background information for areas where little other data are available. Satellite remote sensing has appreciably advanced our knowledge of surface oceanic features off Western Australia. Over the next decade, more accurate and higher resolution satellite sensors will play an increasingly important role in marine research, improving our knowledge of ocean circulation, mixing processes, upwelling systems, wave generation, plankton blooms and fish distributions. It must be remembered, however, that remote sensing effectively covers only the surface waters of the oceans, and therefore complements conventional oceanographic measurements of the subsurface structure. In contrast with the cool equatorward eastern bound¬ ary currents in the South Atlantic and South Pacific Oceans, the dominant ocean current along the Western Australian coast is the Leeuwin Current (Godfrey & Ridgway 1985; Pearce & Walker 1991; Smith et al. 1991), a warm poleward flow of tropical water (Fig 1). Also unlike the Benguela and Humboldt Current systems, © Royal Society of Western Australia 1997 40°S L~ — — - - - - * - 110° E 115° 120° 125° Figure 1. Map of Western Australia showing its marine regions. The cross-hatch area is a schematic representation of the Leeuwin Current, and the dotted line shows the approximate position of the edge of the continental shelf. 1 Journal of the Royal Society of Western Australia, 80(1), March 1997 there is no persistent upwelling of cool nutrient-rich sub¬ surface water onto the Western Australian continental shelf, and consequently sea temperatures off Western Australia are significantly higher than at the same latitudes off southwestern Africa and Chile (Pearce 1991). The Leeuwin Current appears to play an important role in the biogeography of the south-eastern Indian Ocean (Maxwell & Cresswell 1981; Morgan & Wells 1991; Hutchins & Pearce 1994) and to influence recruitment to some of the commercial fisheries of Western Australia (Pearce & Phillips 1988; Lenanton et al 1991; Fletcher et al. 1994; Caputi et al. 1995). Here we review applications of satellite remote sensing in Western Australian coastal and continental shelf waters, briefly outlining the most important results from published as well as unpublished research. Our aim is to demonstrate both the past use and future potential of marine remote sensing in our State and to illustrate how different remote sensing techniques have contributed to our knowledge of local oceanic processes. Technical details should be sought in the standard oceanographic remote sensing textbooks such as Robinson (1985) and Stewart (1985) and in the companion paper by Smith (1997). Further information on local marine applications can be gained from the extensive bibliography prepared by Smith & Pearce (1997). Satellite remote sensing It is less than four decades since the first satellite (the Russian Sputnik) was launched into space in October 1957, to be followed by many earth-observing satellites with a variety of specialised sensors (Fig 2). The ''modem era" of oceanographic remote sensing was perhaps bom in 1978 when three highly successful satellites were launched; the short-lived Seasat (with a wTide array of sensors including an altimeter, synthetic aperture radar SAR and scatterometer), Nimbus-7 (with the Coastal Zone Colour Scanner CZCS) and TIROS-N (carrying the Advanced Very High Resolution Radiometer AVHRR). Satellites may be broadly classified as geostationary (effectively fixed in space some 36000 km above the equator and rotating w'ith the earth) or polar-orbiting. 1970 1980 1990 Figure 2. Operational periods of selected satellites useful for marine remote sensing from 1970 to 1993. Dashed lines indicate satellites which are still functional but not currently operational. (See text for details.) There is a chain of geostationary meteorological satellites viewing almost the entire earth's surface at half-hourly intervals; these include the Japanese GMS satellite (located above the equator at 140° E to the north of Australia), the European Meteosat and Indian Insat. These carry both visible and infrared sensors and are largely used for weather monitoring although they have also been used for oceanographic purposes in some parts of the world. Polar-orbiters, on the other hand, orbit the earth a few hundred km over the poles with a revisit period of one to a fewr days, and generally have a higher spatial resolution (smaller pixel size on the surface) than the geostationary- series. Many of the polar-orbiting satellites have played a crucial role in oceanography, including the highly suc¬ cessful TIROS/ NO AA meteorological satellites (with altitudes around 830 to 870 km and orbital periods of about 102 minutes, first launched in the late 1970s and providing more useful imagery over the oceans than any other series of satellites) and the Landsat series, launched in June 1972 as the first operational high-resolution satellite monitoring the earth's surface and resources. Nimbus satellites generally carried experimental sensor packages which may have then become operational on other satellites. Seasat and Geosat were specialised satellites for monitor¬ ing sea-surface topography and roughness, as are the currently-operating Topex /Poseidon and the European Remote Sensing Satellite ERS-1. There are four basic properties of the ocean which can be measured using the visible, infrared and microwave portions of the electromagnetic spectrum. These are: • surface temperature, which reveals surface circulation patterns including thermal fronts and upwelling zones; • water colour, used to measure phytoplankton and other particulates in the water column as well as properties of the seabed, and so assist in the study of ocean fronts, surface circulation patterns, upwelling regions, dispersion of river and outfall effluent plumes, and seabed habitats in shallow wrater; • surface elevation, where microwave altimeters measure the distance between the satellite and the sea surface to study ocean circulation and tidal heights; and • surface roughness, showing surface waves and wind from backscatter caused by wind on the ocean surface. The first two properties can be measured using passive techniques in which electromagnetic radiation emitted by (or reflected from) the water surface and the atmosphere is received at the satellite sensor. The third and fourth properties are measured by the return of pulses of electro¬ magnetic energy that are actively transmitted from the satellite and the reflected /scattered signals are received back at the sensor. For larger-scale oceanic studies, such as mesoscale (10's to 100's of kilometres) ocean circulation and monitor¬ ing of upwelling zones, the polar orbiters with sensors such as the AVHRR and CZCS (pixel size of order 1 km) are adequate. For coastal waters, where smaller-scale processes are important, sensors such as the Landsat TM with pixel sizes of less than 100 m are appropriate. 2 Journal of the Royal Society of Western Australia, 78:33-38, 1995 We concentrate here on satellite remote sensing of sur¬ face temperature and ocean colour in Western Australian waters to illustrate the role that each has played (or is playing) in studying marine systems in this State. Satellite measurements of surface topography and roughness have not yet been used to any extent in our waters but will almost certainly become invaluable tools in the future. Sea surface temperature Satellite-derived sea-surface temperature (SST) measure¬ ments have been available for over two decades and have been widely used for the derivation of circulation patterns, structure of oceanic fronts, behaviour of eddies/ meanders and the location of upwelling zones. The temperature in the upper layer of the ocean is influenced by radiative processes and both incoming and outgoing surface heat fluxes. Vertical mixing usually ensures that the tempera¬ ture of the thin surface "skin" (from which the infrared radiation is emitted) is close to the "bulk" temperature of the upper few metres of water. Surface temperatures therefore generally reflect the subsurface thermal structure which is in turn related to ocean currents. SST can be estimated using thermal infrared sensors such as the AVHRR on the NO A A series of satellites, with a pixel size of 1.1 km and a swath width of about 2800 km (Fig 3) and the Thematic Mapper (TM) on Landsat 4 and 5, which has a single thermal infrared channel with 120 m pixels and a 185 km swath. Two NOAA satellites are maintained operationally at any time, and each passes over any particular place on the earth's surface twice per day. The AVHRR on satellites NOAA-7, -9, -11, -12 and -14 has 5 spectral bands; one in the visible region of the spectrum, one in the near-infra¬ red, a third in the mid-infrared, and two in the thermal infrared. The visible and near-infrared bands are useful in daylight for land/sea /cloud discrimination, and com¬ binations of these bands are used to derive vegetation indices (Smith 1997) and for cloud detection. Techniques of cloud detection are many and varied, but the simplest (and most popular) involve using threshold limits from both individual bands and combinations of bands (e.g. Saunders & Kriebel 1988a, b). Sea-surface temperatures are computed by combining the radiance temperatures derived from the two indi¬ vidual thermal bands to account for the varying amounts of water vapour in the atmosphere (Barton 1995). Pearce et al (1989) compared AVHRR-SSTs derived from a variety of SST algorithms with surface measurements off Perth and selected the McMillin & Crosby (1984) formulation as giving the most reliable results in this area; this was confirmed by some unpublished data from the Abrolhos Islands, although on occasion the difference between the AVHRR and in situ temperatures was found to exceed 2 °C. [More recent unpublished data indicates that better 0 20 EQUATOR &N LANDSAT STN ‘E SPRINGS \ INDIAN; »TR ALIA BRISBANE ADELAIDE. Typical Ground Swath NDi|| MELBOURNE^ Typical Grou^fiNfcji Is de KERGUELEN' NEW GUINEA^ T^ LU I INDONESIA Eiltpse of Image coverage \ Acquisition range ( N OCEAN / f / / TASMAN SEA MACQUARIE \% o Figure 3. Map of the approximate coverage of a typical swath of the NOAA/ AVHRR (broad shaded band) compared with that of Landsat (narrow band) (Carroll 1982). The stippled circle/ellipse indi¬ cates the total reception area for the AVHRR from the Perth receiving station. 3 Journal of the Royal Society of Western Australia, 80(1), March 1997 corrections may in fact be required near the edges of the swath, the so-called "scan-angle effect"]. Pellegrini & Penrose (1986) found that the McMillin & Crosby (1984) algorithm was also appropriate for tropical waters of the Northwest Shelf, but they suggested that scan-angle effects may become important towards the edges of images in such regions with high water-vapour loadings. Western Australian satellite users are fortunate that the first AVHRR receiving station in Australia was estab¬ lished in Perth in 1981 (Carroll 1982) as a joint project between CSIRO and the Western Australian Institute of Technology (now Curtin University). AVHRR satellite images from satellites NOAA-6 to NOAA-14 have been archived in Perth since late 1981, the longest AVHRR archive in Australia (Pearce 1989) which is being used extensively in studies of the Leeuwin Current and its interaction with inner-shelf waters. NOAA/AVHRR: Large-scale circulation and the Leeuwin Current Historically, the earliest use of SST imagery off West¬ ern Australia appears to have been a study of thermal conditions on the Northwest Shelf in 1966 by Szekielda & Mitchell (1972), who used High-Resolution InfraRed (HRIR) imagery from the Nimbus-2 satellite to derive the annual cycle of SST. They also pointed out the presence of anomalously warm patches of water in this region of high air-sea heat flux. Prata & Lynch (1985) and Lynch et al. (1986) further examined these patches (which can have a horizontal scale of 200 km and are up to 2° warmer than the surrounding water) and deduced that they can propagate southwestwards at over 1 ms1. This is too fast for normal advective processes, so it was con¬ cluded that the patches are possibly caused by spatial and temporal variations in localised heating /cooling of the near-surface waters. While earlier studies of ocean temperatures and salinities had shown a seasonal change in water properties off Western Australia (Rochford 1969; Gentilli 1972), current trajectories from free-drifting buoys as well as hydro- graphic measurements led to the identification and naming of the Leeuwin Current (Cresswell & Golding 1980). The advent of NOAA/AVHRR imagery received in the United States in the late 1970s provided the thermal resolution to confirm the strong southward transport down the west Australian coast during the autumn and winter months and the much weaker flow in summer, also revealing details of the complex eddy-like features associated with the Leeuwin Current (Fig 4; Legeckis & Cresswell 1981). Since the establishment of the NOAA receiver in Perth in 1981, locally-received AVHRR imagery has become a well-used tool in studies of the structure and variability of the Leeuwin Current and associated coastal waters off Western Australia. The surface temperatures show the Leeuwin Current as a strong current meandering south¬ wards from about Exmouth along the edge of the continental shelf; on reaching Cape Leeuwin, it curves eastwards and flows into the Great Australian Bight. Along both the west and south coasts, however, satellite (Prata et al 1986; Pearce & Griffiths 1991) and drifting buoy (Cresswell 1980) studies have shown that the Leeuwin Current is in fact a complex of mesoscale (order of 100 km) eddies and meanders interposed with along- Figure 4. Sea-surface temperature contours off Western Australia. A: April 1980, the tongue of warm water penetrating south¬ wards along the outer continental shelf showing a strongly- flowing Leeuwin Current. B: October 1979 with a weak current. Reprinted from Deep-Sea Research Volume 28, Legeckis R & Cresswell G R, Satellite observations of sea-surface temperature fronts off the coast of western and southern Australia, pages 297-306. Copyright 1981. With kind permission from Elsevier Science Ltd, The Boulevard, Langford lane, Kidlington OX5 1GB, UK. shore current jets. There are frequently a number of these wave-like meanders of different dimensions along the west coast, deflecting the Current offshore and representing an active transport of warm Leeuwin Current waters into the southeast Indian Ocean (Plate 1). These anti-cyclonic (anti-clockwise) meanders can grow to more than 200 km from the coast on occasion before apparently becoming unstable and either breaking off as free-standing eddies (Griffiths & Pearce 1985a) or simply dissipating. Between the meanders, strong alongshore current jets stream southwards along the shelfbreak and outer continental shelf. 4 Journal of the Royal Society of Western Australia, 80(1), March 1997 Plate 1. NOAA-AVHRR image of the Leeuwin Current between Shark Bay and Cape Leeuwin in June 1984. The image has been geometrically corrected and shows the brightness temperature in band 4, with warmest water (the Leeuwin Current) in red cooling through yellow and green to the coldest water in blue. Land is white and clouds are mottled blue/white. Image received by the Western Australian Satellite Technology and Applications Consortium. Plate 2. NOAA-AVHRR image of the cool Capes Current flowing northwards off south-western Australia in March 1991. Other details as in Plate 1. Plate 3. NOAA-AVHRR image of Shark Bay; January 1994, representing summer conditions. Other details as in Plate 1. Plate 4. NOAA-AVHRR image of Shark Bay; July 1993, for winter. Other details as in Plate 1. 5 Journal of the Royal Society of Western Australia, 80(1), March 1997 Plate 5. Bathymetry off Broome derived from a Landsat TM Band 1 image in December 1987. The image shows bottom stratigraphic features, the two features parallel to the coast in the middle of the picture being in water depths of 15 to 30 m. (Image courtesy of Peter Hick, CSIRO). Plate 6. Wake eddy behind Cape Voltaire (14° 16' S, 125c 35' E) in Admiralty Gulf, enhanced from a Landsat TM image. The horizontal scale of the eddy is about 1 km in a current of about 1.5 m s '. Plate 7. Landsat TM image of the Peel-Harvey estuary, showing the equivalent of Secchi disk depth on 24 January 1990, derived from a combination of Landsat TM bands 1 and 3. Reprinted by permission of the Publisher, from Lavery P S, Pattiaratchi C B, Wyllie A & Hick P T, Water quality monitoring in estuarine waters using the Landsat Thematic Mapper, Remote Sensing of Environment Volume 46 pages 268-280. Copyright 1993 by Elsevier Science Inc. Plate 8. Landsat MSS composite image (bands 4, 5, 7) of the Capes area on 14 November 1984 (courtesy Alex Wyllie and Robert Shaw, WA Department of Land Administration) showing internal waves near Cape Leeuwin. 6 Journal of the Royal Society of Western Australia, 80(1), March 1997 Closer inshore, there is a seasonally-varying north¬ wards counter-current along the continental shelf north of Rottnest Island (Cresswell et al. 1989). However, satellite images have suggested that a cold plume also penetrates northwards past Cape Leeuwin during the summer months (Phillips et al 1991; Cresswell & Peterson 1993). More detailed analysis of a series of AVHRR images enabled Pearce & Pattiaratchi (unpublished observations) to define and name the Capes Current, a nearshore current flowing northwards past Capes Leeuwin and Naturaliste and along the mid-continental shelf past Rottnest Island. In contrast with the winter situation where the Leeuwin Current is close inshore near Cape Leeuwin, the Leeuwin Current swings offshore in summer as the Capes Current penetrates northwards against the coast (Plate 2). This cool current (which may also involve localised upwelling; Gersbach et al. personal communication) is narrow and shallow, persisting between about October and March while the wind stress is strongly northward. After rounding Cape Leeuwin, the Leeuwin Current changes character. In contrast with the meandering behaviour along the west coast, where the current path as a whole deflects offshore, the current along the south coast tends to run along the outer continental shelf (with the strongest currents just beyond the shelfbreak) towards the Great Australian Bight. It is characterised by strong thermal fronts as the tropical water borders cold South¬ ern Ocean waters. Periodically, warm offshoots peel off to the south and temporarily disrupt the eastward flow (Griffiths & Pearce 1985b; Cresswell & Peterson 1993); these features frequently have the form of eddy pairs, with clearly-defined clockwise (cool) and anticlockwise (warm) circulation elements. Some offshoots continue growing southwards away from the coast and eventually separate from the Current to drift away as freely-rotating eddies 300 km or more offshore (Griffiths & Pearce 1985a). Using AVHRR imagery to complement simultaneous Acoustic Doppler Current Profiler (ADCP) measure¬ ments from a research vessel, the three-dimensional structure of the Leeuwin Current and some of the warm offshoots can be analysed. Cresswell & Peterson (1993) measured current velocities in some of the offshoots which were evident in satellite images, finding current speeds of up to 1.5 m s*1 (Fig 5). This work is a fine example of the value of linking conventional oceano¬ graphic measurements with satellite imagery. By extracting digital SST transects across the Leeuwin Current from the imagery, the surface expression of the current structure can be quantified in terms of the position of the peak temperature in the current and derivation of SST gradients (Prata & Wells 1990; Pearce & Griffiths 1991). The thermal topography varies seasonally and with latitude. During the summer months when the Leeuwin Current is weak, the thermal structure is ill- defined and SST gradients are small, but significantly increase while the current is strengthening in autumn, before decaying again in late spring (Pearce & Prata 1990). As the Leeuwin Current flows southwards into increasingly cool waters, the temperature differential across its boundary correspondingly increases. At the latitude of Shark Bay, this differential is typically about 1° C, whereas off Perth it is more like 2° to 3° C and along the western part of the south coast where the warm Leeuwin Current waters enter the Southern Ocean it can exceed 4° C (Godfrey et al. 1986; Prata & Wells 1990). 120°E 34 S — h37°S 120°E Figure 5. Surface currents in Leeuwin Current offshoots along the south coast near Albany. The shaded region represents warmer water from a NOAA-AVHRR image and the current vectors are from the ADCP (see text). The letters D to L are from the original paper and are not used here. Reprinted by permission of the Publisher from Cresswell G R & Peterson J L, The Leeuwin Current south of Western Australia, Australian Journal of Marine and Freshwater Research Volume 44 pages 285-303. Copyright 1993. 7 Temperature ‘C Temperature Journal of the Royal Society of Western Australia, 80(1), March 1997 — Oct — Nov . Dec Months — 111-E — 112-E 113-E ° 113*48' Figure 6. A to D Monthly mean SST transects on latitude 25° 20' S across the continental shelf between 111° E and the eastern shore of Shark Bay derived from AVHRR data 1988 to 1993. E: depicts the seasonal SST cycle at four selected longitudes representing offshore water (111° E), the Leeuwin Current (112° E), the entrance to Shark Bay (112° E) and the shallow eastern part of Shark Bay (113° 48' E). Prata et al. (1986) derived the trajectory of the Current by determining the position of the peak SST along each dataline, clearly illustrating the meandering of the flow down the coast. Another example of digital data extracted from AVHRR imagery is a series of monthly mean SST transects across the continental shelf at 25° 20' S and extending into Shark Bay, illustrating the seasonal variation in cross-shelf thermal structure. In the summer months (Fig 6A, Plate 3), the water near the eastern shore of Shark Bay is 3° to 4° C warmer than that outside the Bay. During autumn and into winter (Plate 4; Fig 6B, C), the shallow water in the bay cools as heat is lost to the atmosphere so that by June the eastern bay water is 4° C cooler than that in the Leeuwin Current. As spring extends into summer, the cross-shelf gradient switches again (Fig 6D). The transition months when the water is almost iso¬ thermal across the shelf and Bay are April and October. The seasonal variation is graphically shown in Figure 6E where both the amplitude and phase of the annual cycle change between offshore and the bay itself; in the Leeuwin Current (112° E) the water is warmest in May and coolest in September/October and the annual range is 4° C, whereas in the eastern bay (113° 48' E) the range is almost 9° C and the extremes are in March and July. 8 Journal of the Royal Society of Western Australia, 80(1), March 1997 NOAA/AVHRR for fisheries applications Because of anticipated links between water temperatures and fish catches (particularly for pelagic species such as tuna), one of the original motivations for establishing the NOAA remote sensing facility in Perth was to provide SST imagery to the fishing industry. A study examining the correlation between satellite-derived SSTs (provided in real-time to the fishing industry) and southern bluefin tuna catches along the southwestern coast of Australia (Myers & Hick 1990) found somewhat inconclusive results. This was largely because of the inadequate facilities available for image acquisition and processing at that time, persistent cloud over the fishing area near Albany, the general decline of the tuna stock (with consequent changes in the operation of the industry) and the practices of the fishermen themselves. Nevertheless, recent im¬ provements in both the reception and image processing facilities have enabled SST images to be made available to the fishing industry in real-time, and some fishing companies in Western Australia are now making use of the pictures to assist in fishing management decisions. To explain some of the established links between the flow of the Leeuwin Current (as indexed by annual mean coastal sea levels) and recruitment of the western rock lobster, Pearce & Phillips (1988) have suggested that the strong onshore flow in the southern part of mesoscale meanders may regionally boost the shoreward migration of the phyllosoma larvae. A particularly well-defined meander in an AVHRR image during the settlement period in October 1984 may have contributed to exceptionally high local settlement at Dongara at that time. Likewise, AVHRR images suggested that fluctuations in recruitment at Cape Mentelle (34° S) may be associated with the onshore/ offshore position of the Leeuwin Current in that area (Phillips et al 1991). The annual recruitment of tropical fish larvae at Rottnest Island in April has been attributed by Hutchins & Pearce (1994) to the southwards flow in the Leeuwin Current. Analysis of AVHRR imagery indicated that the arrival of larvae was associated with tongues of Leeuwin Current water penetrating across the shelf to bathe the Island, although the authors acknowledged that other factors (both oceanographic and biological) would certainly play a role as well. Fletcher et al. (1994) used AVHRR images to assist in the interpretation of pilchard larval and egg distributions along the south coast. Larval surveys during the July spawning period when the current flows most strongly showed that there was eastward advection of eggs and larvae along the shelf at between 30 and 40 km d 1 (35 to 45 cm s*1). During the December spawning, on the other hand, there was no indication of alongshore advection of eggs and larvae because the current is wTeaker at that time of year. AVHRR imagery also proved useful in explaining the movement of an oil slick from the tanker Kirki which lost its bow near Jurien in 1991, with fears that the oil spill could decimate the sensitive rock lobster nursery reefs. Analysis of SST images shortly after the disaster showed that a meander of the Leeuwin Current was responsible for carrying the oil offshore despite generally onshore winds at the time. Use of current patterns derived from the imagery in OSSM (the On-scene Oil Spill Model) en¬ abled the oil distribution to be reproduced in the model (Easton et al. 1992), proving the value of satellite imagery in marine disasters of this type. NOAA/AVHRR and submerged strandlines In a novel use of AVHRR thermal imagery, Tapley (1990) produced a relief-shaded night-time image of sur¬ face thermal boundaries for a coastal region off Broome. Oceanic eddies were revealed, but more surprisingly the image also showed the surface expression of submerged limestone strandlines consisting of pairs of narrow ridges some 300 to 500 m wide and rising 5 to 8 m above the seabed in water about 20 m deep. Tapley's interpretation was that the strandlines created vertical currents and both the upwelled cooler water and changed surface sea- state were evident as a thermal signature. The presence of the strandlines was verified by diver observations during the validation phase of the study. NOAA/AVHRR for coastal processes While the mesoscale meanders and eddies described above extend offshore (westwards), smaller eddies and billows with length scales between 5 and 35 km are also found along the inshore boundary of the current and penetrate across the continental shelf (Pearce & Griffiths 1991; Wyllie et al. 1992). Field studies of nearshore oceanographic processes off Perth using conventional oceanographic measurements complemented by NOAA- AVHRR and Landsat TM imagery have revealed some of the dynamics of the interaction between the Leeuwin Current and the nearshore waters (Mills et al 1992). Cross-shelf mixing processes result in active "flushing" of the inshore region and are believed to play a role in the dispersal of coastal pollutants. A composite TM image also showed discoloured water from the Peel-Harvey estuary and the Swan River moving northwards (Wyllie et al. 1992), illustrating the complementary roles of multi- spectral and multi-scale imagery in interpreting coastal and nearshore processes. Nevertheless, under appropriate conditions NOAA/ AVHRR imagery alone can reveal relatively small-scale localised features, such as upwelling around Rottnest Island caused by the interaction between flow and topography ("island wake effect"). During the summer months, the predominantly southerly winds drive the currents on the continental shelf northward past Rottnest Island, result¬ ing in a persistent cold water patch immediately to the north of the Island (Pattiaratchi, unpublished data). During winter, on the other hand, interaction between the south¬ ward flowing Leeuwin Current and Rottnest Island leads to flow separation with the formation of an eddy off the western tip of the Island. Another coastal application of full-resolution thermal AVHRR imagery involved a tidal jet from an estuary at Bunbury (Hearn & Pearce 1985), showing a plume of buoyant warm water extending some 5 km from the inlet, with a maximum thermal contrast of about 0.5° C; this is approaching the limits of the radiometric and spatial resolution of the AVHRR sensor. Estimates of the frontal propagation speed and entrainment coefficient using the image and a tidal model agreed with those obtained using surface measurements. 9 Journal of the Royal Society of Western Australia, 80(1), March 1997 Ocean Colour In contrast with thermal infrared remote sensing, comparatively few ocean colour data are available for Western Australian continental shelf waters, because the CZCS was operational only intermittently and because the visible band on the AVHRR has insufficient spectral resolution to adequately detect the small range of colour changes in the open ocean. For coastal waters, however, remote sensing in visible wavelengths has been used for monitoring chlorophyll (phytoplankton) concentrations, suspended sediments, and bottom type (in clear shallow water). Visible imagery is, of course, only useful during daylight hours and in cloud-free conditions. While thermal infrared energy is emitted from a thin surface skin, water-leaving radiance in the visible region of the spectrum indicates colour variations in the upper part of the water column, depending on the absorption, reflection and scattering of sunlight from particles in the water as well as on surface illumination, roughness, foam and surface films. Clearly, the spectral characteristics of a water body depend markedly on the mix of phytoplank¬ ton species present, the ages of the various components and the presence of inorganic particles. Because much of the radiance reaching the satellite is from scattering and reflection by the atmosphere, corrections must be applied to remove these atmospheric components from the water- derived upwelling signal. The colour sensors which have been most widely used are the Multi-Spectral Scanner (MSS) and the Thematic Mapper (TM) on the Landsat series of satellites, and the Coastal Zone Colour Scanner (CZCS) which operated between 1978 and 1986 on the experimental Nimbus-7 satellite. The CZCS was the first satellite sensor dedicated to monitoring phytoplankton and sediments in the near¬ surface waters of the ocean. The bands had narrow band- widths and high sensitivity tailored to the absorption spectrum of oceanic planktonic pigments, so that chlorophyll concentrations could be determined from the ratios of the radiances in different spectral bands, showing regional and temporal variations in phytoplankton concentration. The CZCS swath-width was about 1600 km and the pixel size 825 m. The Landsat sensors were primarily designed for land applications, but the TM (which has a spatial resolution of 30 m and swath width of 185 km) has proved useful in coastal and bathymetric studies. SPOT is a French satellite with even higher resolution visible sensors (but correspond¬ ingly narrower swath) covering similar wavelengths to the Landsat MSS. The Sea-viewing Wide Field-of-view Sensor (SeaWiFS, scheduled for launch in mid-1997) will provide colour imagery in 8 spectral bands to reveal regions of enhanced biological productivity such as upwelling zones. There have been many unfortunate delays in the launch of this sensor, which will have a swath-width of 2800 km and a resolution of 1.1 km (similar to the AVHRR) and is ex¬ pected to provide a major boost to coastal oceanic and fisheries-related studies. The Ocean Colour and Tem¬ perature Scanner (OCTS) was recently launched on the Japanese Advanced Earth Observing Satellite (ADEOS); it has 8 bands in the visible and near infra-red and 4 in the thermal region of the spectrum, providing coincident visible and thermal imagery with a spatial resolution of about 700 m. Landsat TM: Bathymetric and habitat mapping Because visible radiation can penetrate into the water column, the bathymetry and texture of the seabed can be determined under suitable conditions. The shorter wave¬ lengths at the blue-green end of the spectrum can pen¬ etrate deeper than the longer (red) wavelengths, so Landsat TM band 1 (0.45 to 0.50 pm) is suited to bathy¬ metric determination while band 2 (0.50 to 0.59 pm) can provide information on the characteristics of the water column. Light reflected from the seabed passes through the water column (where some absorption and scattering by phytoplankton and sediment particles occurs) and then through the atmosphere (with further absorption and scattering) before reaching the satellite sensor. Use of TM imagery for seabed mapping therefore requires the radiances from these component paths to be sepa¬ rated. Landsat TM imagery has been used to map the seabed in coastal waters off Broome, in Shark Bay, around the Houtman Abrolhos Islands, an area south of Geraldton, in Cockburn Sound and Geographe Bay (Fig 1). The study off Broome examined the relationship between reflectance and depth as well as the spectral reflectance characteristics of pearl oyster habitat (Hick & Scoones 1990). Comparison of data from TM band 1 with hydro- graphic charts showed that the radiance in this band is highly correlated with bottom depth down to depths approaching 30 m provided the seabed is reasonably uniform, the water clear and the sun high in the sky to provide good illumination conditions. Classification of the substrate in the imagery as reef, sand and seaweed agreed well with visual observations by divers. Plate 5 shows the bathymetric features off Broome, the visible bottom structures being ancient coastlines and wave-cut platforms in water depths exceeding 30 m. The problem of separating in-water and seabed reflectances has been addressed by Bierwirth et al. (1993) using Landsat TM bands 1, 2 and 3 to " unmix" these components in part of Shark Bay (Fig 1), thus providing information on the colour and structure of the seabed as well as estimating the water depth. By assuming a model of the attenuation of radiant flux with depth and the spatial uniformity of suspended materials in the water column, and then applying a constraint to the substrate reflectances for each pixel, both the water depth and bottom type (seagrass meadows, microbial mats and sand) could be determined from the multispectral data. The distribution of marine habitats at the Houtman Abrolhos Islands has been examined using Landsat TM (bands 1 to 3) and SPOT imagery (Anon 1993, 1994). A number of habitat categories (including sand shallows, coral, seagrass beds and plant-dominated reefs) with different spectral properties were identified from the imag¬ ery, aided by aerial photography and diving surveys. Although interpretation of the imagery was ambiguous in waters deeper than 10 to 15 m, it was concluded that the TM and SPOT imagery was useful in providing broad habitat distributions and some new features of the sea¬ bed were revealed. Landsat TM data were also used by Catalano et al (1991) in a study of environmental information for the management of coastal developments (specifically the tourism and fishing industries) along the west coast between 10 Journal of the Royal Society of Western Australia, 80(1), March 1997 Guilderton and Dongara, south of Geraldton (Fig 1). Mapping of coastal bathymetry to depths of about 10 m, and bottom habitat type was undertaken using Landsat TM bands 1 to 3. While some difficulties were encoun¬ tered in distinguishing seagrass beds from deeper water and reef systems, and attempts to derive the coastal water circulation were unsuccessful, it was concluded that the use of remotely sensed data showed great potential. However, existing data and/or fieldwork are required to validate interpretation of the satellite information. Following an earlier study by Gee & Forster (1984) in the moderately turbid waters of Cockburn Sound just south of Perth (Fig 1), Corner & Lodwick (1992) made use of residual modelling to enhance the accuracy of TM depth estimates in waters covering a depth range of 0.3 to 22 m. They found that, under some conditions, the depth estimates could be improved using a calibrated regression relationship between the TM estimates and the surveyed depths (from a chart). Similarly, seabed features in Geographe Bay (Fig 1) could be identified in TM band 1 images at depths down to 45 m under good conditions (Hick et al 1994) using in situ radiometric mea¬ surements to quantify the effects of suspended sediments and chlorophyll in the water column. The decline in the density of seagrass meadows with increasing depth was evident, as were areas of bare sand where storm events had completely eroded the seagrass. A wide-ranging description of underwater features (seagrass meadows, reefs, bare sand and deltaic areas) in Western Australia is being undertaken to prepare a national atlas to aid in coastal management (H Kirkman, personal communication). Landsat TM band 1 imagery is being used to prepare maps at a scale of 1:100,000, supported by ground truth observations to confirm features identi¬ fied in the imagery. Landsat TM: Island wakes Small-scale dispersion of coral spawn from reefs is assisted by turbulent wakes and eddies shed from the islands or reefs. Landsat TM imagery has been used to study these wakes behind islands and headlands in Admiralty Gulf, Northwestern Australia, where suspended sediment is visible as a tracer (Pattiaratchi 1994). The structure and behaviour of the wakes depends on an "island wake parameter" involving the size of the island, the regional current speed, the water depth and the vertical eddy viscosity coefficient. Analysis of island wakes and headland eddies identified from satellite data has shown good agreement between the observed wakes and those predicted using the island wake parameter. Plate 6 shows a headland eddy shed from Cape Voltaire in Admiralty Gulf observed in a Landsat TM image; the tidal range in this area is up to 6 m at spring tides, which results in maximum surface currents of 1.5 m s1 and recirculating eddies with dimensions consistent with the theory. Landsat TM: Water quality Routine monitoring of water quality parameters such as chlorophyll, suspended sediment, light attenuation, etc. in coastal waters may be undertaken using Landsat TM data, and techniques to undertake such studies have been developed for Western Australian estuaries and coastal waters by Lavery et al (1993) and Pattiaratchi et al. (1992). Estimates of the surface chlorophyll concentra¬ tion and Secchi disk depth (a measure of water clarity) were obtained using the satellite-received radiance cor¬ rected for atmospheric effects. The algorithms were developed using water quality data collected at the time of the satellite overpass. The study sites selected were the Peel-Harvey estuarine system (between Perth and Geographe Bay, Fig 1) and Cockburn Sound, two areas where degradation of water quality has been documented over the past few years. The distribution of Secchi disk depths in the Peel-Harvey system derived from a Landsat TM image during a flooding tide clearly showed the frontal structure of the clearer oceanic water (larger Secchi disk depths) and the more turbid estuarine waters (Plate 7). Landsat MSS: Internal waves An interesting application of Landsat MSS Band 6 imagery was the identification of internal waves on the Northwest Shelf (Baines 1981). Internal waves exist in the interior of the ocean and tend to form along the inter¬ face between layers of fluid of different densities. Associ¬ ated with these waves are surface bands of convergence and divergence (Robinson 1985) with differing surface roughness visible in satellite imagery. They generally lie parallel to bathymetric features. Baines (1981) interpreted the rough/smooth bands on the sea surface as indicating the presence of tidally-generated internal waves on the thermocline with wavelengths of 300 to 1000 m and propagation speeds of between 0.5 and 1 m s'1. A more recent (November 1984) Landsat MSS image of the area near Cape Leeuwin has revealed a series of internal waves presumably propagating shorewards from about the position of the continental shelf break, with a wavelength of about 3.5 km (Plate 8). CZCS: Coastal productivity The only application of CZCS imagery off Western Australia to date appears to be that of Pattiaratchi et al. (1990) who used chlorophyll concentrations derived from the CZCS colour bands and SST from the thermal band to examine the interaction between the Leeuwin Current and the continental shelf waters between Shark Bay and Perth (Fig 1). Chlorophyll concentrations were estimated using standard CZCS algorithms with the removal of atmospheric effects. Although the absolute chlorophyll concentrations were somewhat uncertain, it was clear that the continental shelf waters contain higher chlorophyll levels than in the Leeuwin Current (Pattiaratchi et al, unpublished obser¬ vations). Mesoscale meanders sweeping the Leeuwin Current onto the continental shelf were evident in the thermal images, matched by the chlorophyll which acted as a tracer indicating that the current was entraining the productive shelf waters and effectively exporting them offshore (Fig 7). Depending on the degree of entrainment, some offshore meanders /eddies were relatively rich in chlorophyll, with implications for enhanced biological production in these water bodies. SST and chlorophyll transects across the eddies clearly showed the entrain¬ ment of the chlorophyll-rich shelf waters into one of the eddies. 11 Journal of the Royal Society of Western Australia, 80(1), March 1997 28 28 30 32 Figure 7. Line sketch interpretations of the Leeuwin Current from (A) CZCS thermal infrared image (band 6) and (B) chlorophyll distribution on 10 May 1981, indicating entrainment of high-chlorophyll shelf waters by the Leeuwin Current. The warm Leeuwin Current is depicted as white in A, as are the relatively high-chlorophyll waters in B. The current arrows have been derived from the SST image A and then transferred to the chlorophyll image B. Features D, K and N-P represent eddies, and the main flow in the Leeuwin Current is along E-H-L-M-N-P-Q-R-S. The densely shaded areas with curly boundaries are clouds. Features A to T and the four transects (solid horizontal lines) are not discussed here. The future Understanding the ocean circulation is fundamental to almost all aspects of marine science including biology, fisheries management and research, shipping operations and climate research. There is a growing awareness of the important role played by the circulation in the life histories of many fish and hence in recruitment to com¬ mercial fisheries, and the near-realtime tracking of marine pollutants (for example the movement and dispersion of oil spills) by satellite can provide essential and timely information for fisheries managers and pollution authorities to make decisions about appropriate alleviation measures. Medium resolution thermal imagery (NOAA/AVHRR SST) will continue to provide useful and accessible datasets for monitoring surface features of the large-scale ocean circulation, but is increasingly being supplemented by ocean surface topography. Traditionally, oceanogra¬ phers have derived the ocean circulation by computing the dynamic height of the sea surface from temperature and salinity profiles, but measurement of sea surface elevation can now be made on a global scale using satellite altimeters such as Topex/ Poseidon to yield direct estimates of ocean currents (Fu et al 1994), with the advantage of not being restricted by cloud cover. The distance between the satellite and the sea surface is measured by the time lapse between radar pulses and the return echoes of a narrow beam transmitted vertically downward from the altimeter to the sea surface. The vertical resolution of altimetric measurements has steadily improved from about 1 m (Skylab) to about 5 cm (Topex; Fu et al. 1994) subject to corrections for uncertainties in the geoid, orbit error, atmospheric and ionospheric effects, errors related to sea state, and tides. Acquisition of global sea level data will also allow' monitoring of sealevel variability as a result of climate change. Complementing thermal satellite imagery with ADCP-measured currents has confirmed that, at least in subtropical regions writh strong boundary currents, surface temperatures generally indicate near¬ surface current patterns (Cresswell & Peterson 1993). Most of our present knowledge of global wind and wave conditions has been derived from shipping obser¬ vations. Satellites can provide wind and wave data on a regular basis over large areas, including remote and in¬ hospitable ocean regions, by recording the roughness of the ocean surface with instruments such as the altimeter, synthetic aperture radar (SAR) and scatterometer (Stewart 1985). The first satellite to carry all these instruments was the short-lived SeaSat (which operated for only about 3 months in 1978), and they are now on the currently- operational European ERS satellites. Satellite-derived wind and wave estimates from the present generation of satellite instruments are comparable with those using in situ measurements and are available for coastal wave climatologies and offshore maritime activities such as oil production, shipping and coastal engineering. Routine monitoring of sea state and winds off Western Australia from satellites is very likely in the near future. 12 Journal of the Royal Society of Western Australia, 80(1), March 1997 Following the important results derived from CZCS colour imagery, the availability of ocean colour data from SeaWiFS and OCTS will greatly improve our under¬ standing of regions where enhanced biological produc¬ tivity has implications for commercial fisheries operations. High resolution ocean colour imagery will also play an increasingly valuable role in monitoring variations in seabed topography, habitat type (in appropriately shallow waters), coastal pollution and natural tracers such as river plumes. There will be an ongoing need, however, for validation of remotely sensed data using conventional (surface- based) measurements, a typical example being the need for in situ measurements from surface vessels, buoys, etc. to validate satellite-derived surface ocean temperatures. This is particularly true in tropical areas of high atmo¬ spheric water vapour loading, where the accuracy of remotely-sensed SSTs is often poor and SST imagery may not be a reliable indicator of either absolute temperatures or circulation patterns. The advent of new satellites and sensors will provide increased opportunities for use of complementary sensors in different regions of the electromagnetic spectrum, for example the study of surface circulation patterns using both thermal infrared imagery (to show thermal fronts), altimeter data to derive current speeds and ocean colour imagery to track the movement of tracers such as sediments or plankton patches. Because of the vast expanse of ocean recently proclaimed as Australia's Exclusive Economic Zone (EEZ), satellite remote sensing provides the only feasible means of monitoring both physical and biological processes in our surface waters on a regular basis. Conventional oceanographic measurements are still necessary, however, for providing the full 3-dimensional description of the oceans, so satellite remote sensing must be recognised as only another tool available to the marine scientist. Acknowledgements. This paper has been partly based on information provided by M Grundlingh (CSIR South Africa) as well as by colleagues too numerous to mention individually. Satellite pictures were processed by A Way and C Bowron (CSIRO Division of Oceanography) from imagery supplied by the Western Australian Satellite Technology and Applications Consortium (WASTAC). Landsat TM images have been kindly supplied by P Hick (CSIRO), A Wyllie and R Shaw (WA Department of Land Administration). K Nardi (Abrolhos Islands Consultative Council) and J Stoddart (Marine Science Associates) gave permission to use reports prepared for the Mid-West Development Commission. C Burton under¬ took the analysis of SST transects off Shark Bay. References Anon 1993 Distribution of marine habitats at the Houtman Abrolhos Islands. I: the Easter Group. Report. Kinhill Engineers, for the Geraldton Mid-West Development Authority. Western Australia, 39 pp. Anon 1994 Distribution of marine habitats at the Houtman Abrolhos Islands. II: the Wallabi and Southern Groups. Report. Marine Science Associates for the Mid-West Devel¬ opment Commission. Perth, Western Australia, 11 pp. Baines P G 1981 Satellite observations of internal waves on the Australian North-West Shelf. Australian Journal of Marine and Freshwater Research 32:457-463. Barton I J 1995 Satellite-derived sea surface temperatures: current status. Journal of Geophysical Research 100:8777-8790. Bierwirth P N, Lee T & Burne R V 1993 Shallow sea-floor reflec¬ tance and water depth derived by unmixing multispectral imagery. Photogrammetric Engineering & Remote Sensing 59:331-338. Caputi N, Fletcher W J, Pearce A F & Chubb C F 1995 Effect of the Leeuwin Current on the recruitment of fish and inverte¬ brates along the Western Australian coast. Proceedings of the International Larval Fish Conference. Marine and Fresh¬ water Research 47:147-155. Carroll W 1982 The WAIT satellite receiving station. In: Work¬ shop on Applications of Environmental Satellites (eds F R Honey & W Carroll), Muresk, CSIRO, Paper 2.1.1, 6 pp. Catalano P, Wyllie A & Eliot I 1991 Development of a coastal planning information system: Guilderton to Dongara, Western Australia. Proceedings of the Conference on Remote Sensing and Geographical Information Systems for Coastal Catch¬ ment Management. Lismore, New South Wales, 205-222. Corner R J & Lodwick G D 1992 Spectral bathymetry using Landsat Thematic Mapper data in Cockbum Sound, Western Australia. Proceedings of the 6th Australasian Remote Sensing Conference Wellington, New Zealand 1:180-190. Cresswell G R 1980 Satellite-tracked buoys in the the eastern Indian Ocean. Proceedings of the 14,h International Sympo¬ sium on Remote Sensing of the Environment. Costa Rica, 531-541. Cresswell G R, Boland F M, Peterson J L & Wells G S 1989 Continental shelf currents near the Abrolhos Islands, Western Australia. Australian Journal of Marine and Fresh¬ water Research 40:113-128. Cresswell G R & Golding T J 1980 Observations of a south¬ flowing current in the southeastern Indian Ocean. Deep-Sea Research 27:449-466. Cresswell G R & Peterson J L 1993 The Leeuwin Current south of Western Australia. Australian Journal of Marine and Freshwater Research 44:285-303. Easton A, Pearce A F & Buchan S J 1992 Modelling the Kirki spill using OSSM. Proceedings of the Third National Scientific Support Coordinators Workshop. Fremantle, Western Aus¬ tralia, 93-96. Fletcher W J, Tregonning R J & Sant G J 1994 Interseasonal variation in the transport of pilchard eggs and larvae off southern Western Australia. Marine Ecology Progress Series 111:209-224. Fu L-L, Christensen E J, Yamarone C A, Lefebvre M, Menard Y, Dorrer M & Escudier P 1994 Topex /Poseidon mission over¬ view. Journal of Geophysical Research 99:24369-24381. Gee L & Forster B C 1984 Landsat data as an aid to hydro- graphic and bathymetric mappping in shallow seas. Proceed¬ ings of the 3rd Australasian Remote sensing Conference. Gold Coast, Queensland, 161-173. Gentilli J 1972 Thermal anomalies in the eastern Indian Ocean. Nature 238:93-95. Godfrey J S & Ridgway K R 1985 The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. Journal of Physical Oceanography 15:481-495. Godfrey J S, Vaudrey D J & Hahn S D 1986 Observations of the shelf-edge current south of Australia, winter 1982. Journal of Physical Oceanography 16:668-679. Griffiths R W & Pearce A F 1985a Satellite images of an unstable warm eddy derived from the Leeuwin Current. Deep-Sea Research 32:1371-1380. Griffiths R W & Pearce A F 1985b Instability and eddy pairs on the Leeuwin Current south of Australia. Deep-Sea Research 32:1511-1534. Hearn C J & Pearce A F 1985 NOAA satellite imagery and air¬ borne remote sensing of a small-scale tidal jet. Australian Journal of Marine & Freshwater Research 36:643-653. 13 Journal of the Royal Society of Western Australia, 80(1), March 1997 Hick P T & Scoones R 1990 Satellite-derived imagery for water penetration and its relevance to fisheries in the Kimberley region of Western Australia. CSIRO Exploration Geoscience Restricted Report 122R, 29 pp. Hick P T, Wyllie A & Kirkman H 1994 Marine substrate feature delineation from Thematic Mapper data: the Geographe Bay example. Proceedings of the 7th Australasian Remote Sensing Conference. Melbourne, Australia 3:79-84. Hutchins J B & Pearce A F 1994 Influence of the Leeuwin Current on recruitment of tropical reef fishes at Rottnest Island, Western Australia. Bulletin of Marine Science 54:245-255. Jarvis N 1986 Western Australia, an atlas of human endeavour. Western Australian Department of Lands and Surveys, Perth. Lavery P S, Pattiaratchi C B, Wyllie A & Hick P T 1993 Water quality monitoring in estuarine waters using the Landsat Thematic Mapper. Remote Sensing of Environment 46:268- 280. Legeckis R & Cresswell G R 1981 Satellite observations of sea- surface temperature fronts off the coast of western and southern Australia. Deep-Sea Research 28:297-306. Lenanton R C, Joll L, Penn J & Jones K 1991 The influence of the Leeuwin Current on coastal fisheries of Western Australia. Journal of the Royal Society of Western Australia 74:101-114. Lynch M J, Prata A J & Hunter J R 1986 Sea surface temperature anomalies off the North West Shelf of Western Australia. Proceedings of the Is* Australian AVHRR Conference. Perth, Western Australia, 259-268. Maxwell J G H & Cresswell G R 1981 Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Australian Journal of Marine and Freshwater Research 32:493-500. McMillin L M & Crosby D S 1984 Theory and validation of the multiple window sea surface temperature technique. Journal of Geophysical Research, 89:3655-3661. Mills D A, D'Adamo N, Wyllie A & Pearce A F 1992 The response of stratified shelf waters to the Leeuwin Current and wind forcing: winter observations off Perth, Western Australia. Proceedings of the 6th International Conference on the Physics of Estuaries and Coastal Seas, Margaret River, W. Australia; Mixing in Estuaries and Coastal Seas (ed C B Pattiaratchi). American Geophysical Union. Coastal and Estuarine Studies 50:5-28. Morgan G J & Wells F E 1991 Zoogeographic provinces of the Humboldt, Benguela and Leeuwin Current systems. Journal of the Royal Society of Western Australia 74:59-69. Myers D G & Hick P T 1990 An application of satellite-derived sea surface temperature data to the Australian fishing industry in near real-time. International Journal of Remote Sensing 11:2103-2112. Pattiaratchi C B 1994 Physical oceanographic aspects of the dis¬ persal of coral spawn slicks: a review. In: The bio-physics of marine larval dispersal (eds P W Sammarco & M L Heron). American Geophysical Union. Coastal and Estuarine Studies 45: 89-105. Pattiaratchi C B, Lavery P, Wyllie A & Hick P T 1992 Estimates of water quality in coastal waters using multi-date Landsat Thematic Mapper data. International Journal of Remote Sensing 15:1571-1584. Pattiaratchi C B, Parslow J, Pearce A F & Hick P T 1990 Appli¬ cation of Coastal Zone Colour Scanner (CZCS) imagery for productivity and circulation studies of the Leeuwin Current, Western Australia. Proceedings of the 5th Australasian Remote Sensing Conference. Perth Western Australia, 252-255. Pearce A F 1989 A catalogue of NOAA-AVHRR satellite imagery received in Perth, Western Australia, 1981-1987. CSIRO Marine Laboratories Report, CSIRO, Perth, 203, 36 pp. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. Journal of the Royal Society of Western Australia 74:35-45. Pearce A F & Griffiths R W 1991 The mesoscale structure of the Leeuwin Current: a comparison of laboratory models and satellite imagery. Journal of Geophysical Research 96:16739- 16757. Pearce A F & Phillips B F 1988 ENSO events, the Leeuwin Current and larval recruitment of the western rock lobster. Journal du Conseil 45:13-21. Pearce A F & Prata A J 1990 Sea-surface temperature transects across the Leeuwin Current. Proceedings of the 5th Australasian Remote Sensing Conference. Perth, Western Australia, 242-251. Pearce A F, Prata A J & Manning C R 1989 Comparison of NOAA/AVHRR-2 sea surface temperatures with surface measurements in coastal waters. International Journal of Remote Sensing 10:37-52. Pearce A F & Walker D I 1991 The Leeuwin Current. Journal of the Royal Society of Western Australia 74:140pp. Pellegrini J J & Penrose J D 1986 Comparison of ship-based and satellite AVHRR estimates of sea-surface temperature. Pro¬ ceedings of the 1st Australian AVHRR Conference. Perth, 269-279. Phillips B F, Pearce A F & Litchfield R T 1991 The Leeuwin Current and larval recruitment to the rock (spiny) lobster fishery off Western Australia. Journal of the Royal Society of Western Australia 74:93-100. Prata A J & Lynch M J 1985 Monitoring tropical sea-surface temperature structure from NOAA/AVHRR measurements. Proceedings of the 2nd Australian Conference on Tropical Meteorology. Perth, 70-72 (Abstract). Prata A J, Pearce A F, Wells J B & Carrier J M 1986 Satellite sea- surface temperature measurements of the Leeuwin Current. Proceedings of the 1st Australian AVHRR Conference. Perth, 237-247. Prata A J & Wells J B 1990 A satellite sensor image of the Leeuwin Current, Western Australia. International Journal of Remote Sensing 11:173-180. Robinson I S 1985 Satellite Oceanography: An Introduction for Oceanographers and Remote-Sensing Scientists. Ellis Horwood, Chichester. Rochford D J 1969 The seasonal interchange of high and low salinity surface waters off South-west Australia. Technical Paper 29, 8 pp. CSIRO Division of Fisheries and Oceanogra¬ phy, CSIRO, Sydney. Saunders R W & Kriebel K T 1988a An improved method for detecting clear sky and cloudy radiances from AVHRR data. International Journal of Remote Sensing 9:123-150. Saunders R W & Kriebel K T 1988b Errata. International Journal of Remote Sensing 9:1393-1394. Smith R C G 1997 Applications of satellite remote sensing for mapping and monitoring land surface processes in Western Australia. Journal of the Royal Society of Western Australia 80:15-28. Smith R C G & Pearce A F 1997 A bibliography of research into remote sensing of land, sea and atmosphere in Western Aus¬ tralia. Journal of the Royal Society of Western Australia 80:29-39. Smith R L, Huyer A, Godfrey J S & Church J A 1991 The Leeuwin Current off Western Australia, 1986-1987. Journal of Physical Oceanography 21:323-345. Stewart R H 1985 Methods of Satellite Oceanography. University of California Press, Berkeley. Szekielda K-H & Mitchell W F 1972 Oceanographic applications of color-enhanced satellite imageries. Remote Sensing of Environ¬ ment 2:71-76. Tapley I J 1990 The perception of submerged strandlines on night-thermal NOAA-AVHRR satellite imagery. Proceedings of the 5th Australasian Remote Sensing Conference. Perth, 2:815-821. Wyllie A, Buchanan A, D'Adamo N, Mills D A & Pearce A F 1992 The use of Landsat Thematic Mapper and NOAA AVHRR for environmental investigations of the Perth metro¬ politan coastal waters. Proceedings of the 6,h Australasian Remote Sensing Conference. Wellington, New Zealand 3:203- 208. 14 Journal of the Royal Society of Western Australia, 80:15-28, 1997 Applications of satellite remote sensing for mapping and monitoring land surface processes in Western Australia R C G Smith Remote Sensing Services, Department of Land Administration, Leeuwin Centre, 65 Brockway Road, Floreat WA 6014: ricliardsmith.dola@notes.dola.iva.gov.au Manuscript received February 1997 Abstract Visible and near infrared multi-spectral imagery of the earth from a civilian polar orbiting sun synchronous satellite first became available in 1971 and was first processed in Western Australia in 1974 for research into the broad area mapping of land cover types and geological structures. As other regions of the electromagnetic spectrum have become available on later satellites, applica¬ tions have expanded. By the 1990s, repeat observations by similar sensors over many years were being used successfully to monitor changes in land surface conditions over large areas. The integrated, synoptic and temporal views of the changing earth is providing scientists with more complete information into land surface changes and associated processes. Regular monitoring by satellite is enabling future outcomes to be forecast and the community to be better involved in solutions. This is illustrated by increased community support for the management of bush fires in the Kimberley region following the introduction of routine satellite monitoring. Future opportunities for expanding such applications are discussed. Introduction Western Australia has a sparse population of only 1.7 million, managing a vast area of 253 million ha bounded by over 12,000 km of coastline. As a result there is an increasing use of satellite remote sensing to map, monitor, manage and explore the natural resources of Western Australia. Satellite observations complement aerial photog¬ raphy, where large area coverage is needed, natural features change rapidly or geographic information exists beyond the visible and near infrared parts of the electromagnetic spectrum. Satellite usage is expanding with the number of hard copy high resolution satellite images produced by the Remote Sensing Services Branch, Department of Land Administration (RSS, DOLA) increasing from about 200 Figure 1. Number of high resolution satellite images produced as high quality photographic prints by Remote Sensing Services, DOLA. © Royal Society of Western Australia 1997 to 700 per year from 1984 to 1995 (Fig 1). The demand of the mineral exploration industry has probably caused a similar expansion in the private sector. Expanding demand for weather forecasting and for broad scale surveillance to monitor bush fires, drought and sea surface temperatures has contributed to significant growth in the reception and archiving of coarse resolution NOAA satellite data in Western Australia (Fig 2). 4500 4000 2 3500 X 3000 S. 2500 g 2000 ° 1500 | 1000 500 0 T - C\J CO CO CO CO CO CD CD CD CD LCD CD r- CO CD o T— CM CO LO CO CO 00 CO CO CD CD CD CD CD CD O) CD CD CD CD CD CD CD cd CD CD 1 t— ■*“ ▼— T— T“ i — T— i — 1 — Figure 2. Number of overpasses of the NOAA satellite received and archived in Western Australia. To outline how these satellite observations of the earth are impacting the quality of life in Westerna Australia, a gen¬ eral introduction to the principles of remote sensing is given, followed by a brief review of the application of these principles to monitoring and managing land surface processes. Complementary papers cover the applications of satellite remote sensing to oceanography (Pearce & Pattiaratchi 1997) and a bibliography of all publications on satellite remote sensing (research and development) conducted in Western Australia (Smith & Pearce 1997). Some basic texts on remote sensing are Richards (1986), 15 Journal of the Royal Society of Western Australia, 80(1), March 1997 Lillesand & Kiefer (1987), Legg (1992), Buitens & Clever (1993) and D'Souza et al. (1993). Satellite remote sensing The ability of people standing on land to observe the earth's surface is extremely confined by space and time, the oblique view angle, and the very narrow portion of the electromagnetic spectrum observed by the human eye. Earth observing satellites extend the human powers of observations by; • continuous acquisition of data from the nadir view; • regular revisit capabilities giving regular updates; • broad regional and global coverage; • wide coverage of the electromagnetic spectrum in the visible, infra-red, thermal and microwave bands; • a range of spatial resolutions (from 100 km to 10 metres); • ability to manipulate and enhance digital data; • ability to combine satellite digital data with other digital data; • cost effective coverage of regional, continental and global areas; • map-accurate data; • possibility of stereo viewing and digital elevation extraction; and • large archive of historical data for monitoring changes. Earth observations using satellites are limited by their fixed revisit times, the complexity of acquiring and pro¬ cessing the data and for the optical wavelengths by cloud cover. The electromagnetic spectrum Remote sensing measurements of the earth's surface are non-contact and non-invasive. They are based on measuring the interaction of the electromagnetic spectrum with the earth's surface. Passive remote sensing systems are the most common, where energy from the sun reaches the sensor by reflectance or emittance from the earth's surface. In active remote sensing systems the energy is provided by a radar or laser source from the remote sensing platform itself. The electromagnetic spectrum includes a wide range of energy, from X-rays through visible light to radio waves. Only a portion of the electromagnetic spectrum can be used for satellite remote sensing due to limits set by the physical interaction of the radiation with the sur¬ face or the earth's atmosphere. When the wavelength of the radiation diminishes to the same order as the spacing between the molecules in the surface, the radiation is not reflected but is absorbed. In addition, depending on the wavelength, the energy from the surface being sensed is scattered or absorbed by the atmosphere on the way back to the sensor. This restricts remote sensing at the lower end of the visible wavelengths. The upper limit is set by the need to have reasonably detailed images of the earth therefore use of long wavelength radio waves is imprac¬ tical. Having set the upper and lower limits, the useable electromagnetic spectrum is divided into six main sections by the absorption properties of the atmosphere (Fig 3). 100% Atmospheric Transmittance Visible Infrared Ultra-violet Near Infrared Thermal 10.0 pm 1mm 1cm lm Microwave Atmospheric windows available for remote sensing Figure 3. Portions of the electromagnetic spectrum used for remote sensing, showing the main subdivisions and atmospheric windows (modified from Legg 1992). 16 Journal of the Royal Society of Western Australia, 80(1), March 1997 Figure 4. Typical reflectance spectra of dense green vegetation, soil and pure water (Modified from Legg 1992) Ultraviolet. Atmospheric absorption is very strong in the ultra-violet, limiting satellite applications but applica¬ tions are feasible using active airborne systems. World Geoscience Corporation in Western Australia operate an Airborne Laser Fluorescence (ALF) instrument at a wave¬ length of 0.266 pm for oil exploration. The spectra of the fluorescence enables oil films caused by seeps below the sea bed to be identified. Using the same principles, ALF could probably be used for the measurement of algal blooms caused by phytoplankton. Visible. The largest atmospheric window and the highest amount of reflected solar energy is found in the visible wavelengths, and consequently it is the easiest to measure by satellite sensors in space. The first civilian earth observing satellite sensor, the Landsat Multispectral Scanner, had two of its four bands in the visible region with the third and fourth in the near infrared. Reflec¬ tance in the visible wavelength shows little spectral variation for rocks and soil (Fig 4) and is strongly affected by atmospheric scattering which causes a loss of dynamic range which limits contrast. Reflectance in the red part of the visible spectrum is useful for indicating the relative iron content of soils and all visible wavebands are useful for mapping surface geological structures (stratigraphy). Decreased reflectance in the visible part of the spectrum caused by chlorophyll absorption for photosynthesis is used for measurement of vegetation characteristics. Visible wavelengths penetrate water and are scattered back by suspended material and phytoplankton, whereas radia¬ tion in the infrared is almost totally absorbed by water (Fig 3). This enables visible wavelengths to be used for remote sensing studies of water quality and for the bathymetry of shallow coastal waters. However the energy levels are low and the dynamic range of sensors optimised for the land surface are often not well suited for remote sensing water characteristics unless the gain setting can be reset. Near infrared. The near infrared (NIR) portion of the spectrum has similar energy to the visible part and con¬ sequently with appropriate filters can be recorded with the same type of sensor. Less absorption and scattering by the atmosphere occurs in the NIR, resulting in a higher dynamic range and imagery of good contrast. Natural surfaces are rough, causing shadows. Conse¬ quently reflectance is a function of the sun and sensor view angles, giving rise to what is known as the bi-direc¬ tional reflectance function (BDRF). In addition reflectance from vegetated surfaces with partial canopy cover is affected by the reflectance of the soil background. The NIR in combination with a visible waveband can be used to estimate green vegetation cover (Fig 5). To minimise the effects of season, sun angle and soil background on reflectance when measuring vegetation it is common to use a ratio such as the Simple Ratio (SR) between NIR and red (RED) reflectances; SR= N1R/RED or the Normalised Difference Vegetation Index; NDVI = (NIR-RED)/ (NIR+RED) (Figure 5). Using the visible and near infrared wavebands, the first civilian remote sensing satellite Landsat 1 was launched in 1971 into a polar sun-synchronous orbit at an altitude of about 900 km (Table 1). Orbiting the earth every 100 minutes at 27,000 km h l, the MSS sensor scanned a swath width of 185 km and covered the whole earth's surface after 233 orbits over 16 days at a resolution of 80 m. The analog signal was quantised to 6 bits (64 levels) and transferred to earth at a rate of 15 Mb s'1. The spatial resolution of approximately 1 ha per pixel of Landsat MSS is relatively coarse compared with later sat¬ ellites (compare Plate 1 A with B,C,D). Mid infrared. Use of the mid infrared wavelengths did not become possible until the launch of the Thematic Mapper (TM ) sensor on board Landsat-4 (Table 2) in 1983. This spectral region has two atmospheric windows of high transmittance, one centred at 1.5 pm and the other at 2.2 pm (Fig 3). These wavebands are of impor¬ tance in geological and vegetation studies. The relatively low level of reflected radiation at MIR wavelengths requires more sophisticated sensors and cooling to achieve high signal to noise. Landsat-TM also provided 30 m resolution 17 Journal of the Royal Society of Western Australia, 80(1), March 1997 Figure 5. Hypothetical relationship between RED and NIR reflectances and the NDVI as a function of green biomass. Table 1 Characteristics of Landsat Multi-Spectral Scanner (MSS). Band Wavelength (pm) Description Applications 1 0.5-0. 6 Visible - green Water studies 2 0.6-0. 7 Visible - red NDVI, water, soils 3 0.7-0. 8 Red to Near Infrared NDVI, topographic mapping 4 0.8-1.10 Near Infrared NDVI, topographic mapping Table 2 Characteristics of Landsat-TM. Band Wavelength (pm) Waveband Applications 1 0.45-0.52 Visible Blue Water and urban studies 2 0.52-0.60 Visible Green Water and urban studies 3 0.63-0.69 Visible Red Water and vegetation studies 4 0.76-0.90 Near Infrared Vegetation and topography 5 1.55-1 .75 Mid Infrared Vegetation, especially forestry 6 10.4-12.5 Thermal Infrared Water, land surface temperatures and evapotranspiration 7 2.08-2.35 Mid Infrared Soils and minerals Table 3 Characteristics of the sensors on the SPOT satellite. Mu ltispectral Mode (XS) Panchromatic Mode (PAN) Spectral Bands (pm) 0.50-0.59 0.51-0.73 0.61-0.68 0.79-0.89 Pixel size 20 m 10 m Swath width 60 km 60 km Repeat cycle 21 days 21 days Steerable ± 27° 1-4 days revisit 1-4 days revisit 18 Journal of the Royal Society of Western Australia, 80(1), March 1997 for all bands with 120m for the thermal infrared. This improvement in resolution can be seen by comparing Plate IB with 1A from Landat-MSS. The Thematic Mapper was designed for mapping land cover themes and to im¬ prove precision the analogue signal was quantised to 8 bits (256 levels). The increased amount of data is trans¬ mitted to the earth at a rate of 85 Mb s'1, which resulted in expensive X band reception facilities being provided by the Commonwealth Government at Alice Springs in 1987 for local reception. Landsat-TM has been the main satellite sensor used in high resolution remote sensing in Western Australia since 1987. Band 7 is in a region where the phyllosilicates such as clay minerals show absorption peaks. Therefore it was included at the insistence of geologists and has proved of value in lithological mapping and in the discrimination of clay-rich alteration zones. Band 5 is sensitive to the surface water content, therefore it is responsive to the amount and condition of vegetation. The broad width of these two wavebands restricts the amount of mineralogical information that can be derived. A multispectra 1 capability in the MIR region is required to extract this information. The high resolution Landsat-TM data was later ex¬ ceeded by the 20 and 10 m resolutions of the French SPOT satellite launched in 1986 (Table 3; Plate 1C,D,). However, the higher cost and absence of routine coverage and lack of a mid infrared band has limited the application of SPOT multi-spectral data in Western Australia. The panchromatic data is often merged with Landsat-TM to improve the quality of the imagery for mapping surface features. SPOT with its steerable sensor enables stereo pairs to be acquired for digital elevation extraction. SPOT-4, due to be launched in 1997 (CEOS 1995), will have a mid-infrared band from 1.5-1 .7 pm and a wide field of view sensor, named the VEGETATION sensor with a swath width of 2,200 km and resolution of 1 km. An image based on a general enhancement of Landsat-TM using bands 7,4 and 1 is commonly produced to assist geological mapping by the Geological Survey of Western Australia. An example of this enhancement from inland Australia is given in Plate 2. Thermal infrared. This region includes two distinct at¬ mospheric windows (Fig 3), separated by a strong atmo¬ spheric absorption zone caused by water vapour. Remote sensing in this region is similar to the mid infrared with low energy levels associated with emitted radiation. Special detectors, cooling and optics are required to get a high signal to noise ratio. Development of civilian space-borne thermal infrared sensors was driven by meteorologists who required measurements of sea surface and cloud top temperatures. The first polar orbiting satellite to provide thermal infrared sensors was the NOAA-Advanced Very High Resolution Radiometer (Table 4). It was launched in 1979 following a successful prototype first launched in 1974. Designed for meteorological applications, NOAA- AVHRR has a wide field of view sensor covering a swath width of 2,800 km with a resolution of 1.1 km at nadir and transmitting at 665 kb s"1. Bands 1 and 2 are used for detection of clouds, vegetation measurement and detecting firescars. Band 3 is used for detecting fires, as at this wavelength the sensor does not saturate at high tempera¬ tures and has the sensitivity to detect gas flares from the oil platforms on the NW shelf. Bands 4 and 5 provide cloud, land and sea surface temperature measurements. Based on differential absorption by the atmosphere, use of the two bands (4 and 5) can be used to empirically correct for atmospheric effects. The NOAA-AVHRR data has high radiometric resolution with 10 bit quantisation (1024 levels) and on-board black body calibration of the thermal sensors. Lack of on-board calibration of bands 1 and 2 has made accurate long term measures of vegetation difficult due to changes in sensor response over time. These changes occur when new NOAA satellites are launched, sensors age and overpass times drift later causing decreasing sun illumination angles. NOAA-AVHRR is a American meteorological sensor whose data is free as part of a World Meteorological Agreement. Compared with Landsat-TM, NOAA- AVHRR provides high temporal resolution (daily vs 16 day coverage) but low spatial resolution (1 km vs 30 m), which keeps data handling manageable and makes eco¬ nomic local L-band reception (ca. $50,000) compared with X-band reception (ca. $2 million). This affordable capital cost has enabled operational use of NOAA-AVHRR to monitor seasonal vegetation growth, bush fire risk, fires, fire scar mapping and sea surface temperatures to be developed by RSS, DOLA. Operational monitoring of green vegetation cover is made using the NDVI (Plate 3) and achieved by combining successive NOAA overpasses over 14-16 day periods to produce cloud free images at the middle and end of each month. The wavebands in the thermal infrared region contain significant geological information. Temperature differ¬ ences between day and night often reveal buried geological features based on differences in specific heat. Differences in thermal emissivity are of value in lithological discrimi¬ nation, but requires a multispectral capability that will be available on a proposed Japanese/ American satellite. Table 4 Characteristics and applications of the NOAA-AVHRR satellite sensor Band Wavelength (jam) Waveband Applications 1 0.58-0.68 visible Surface features, cloud, vegetation, albedo 2 0.73-1.10 near infrared Water, vegetation, firescars, albedo 3 3.55-3.93 thermal Fires and volcanoes 4 5 10.5- 11.3 11.5- 12.5 thermal thermal Sea, land, cloud temperature and evaporation Sea, land, cloud temperature and evaporation Swath 2,800 km Scan angle ± 55.4°, Resolution at nadir 1.1 km Revisit 12 hours Repeat cycle 9.2 days 19 Journal of the Royal Society of Western Australia, 80(1), March 1997 Other wide field of view sensors (WiFs). The coarse resolution of the NOAA-AVHRR satellite restricts its use in agriculture as individual agricultural fields on the ground cannot be resolved. However WiFS sensors on the Russian RESURS-01 and Indian IRS-1C satellites (CEOS 1995) provide data in the visible and NIR wavebands at 170 m resolution that could be ideal for agricultural monitoring. In addition to the above satellites which are polar or¬ biting and sun synchronous. Western Australia is observed by the Japanese Geosynchronous Meteorological Satellite (GMS-5) which is situated in a geostationary orbit at 140 E, 36,000 km above the equator. The 36,000 km distance compared with the 700 to 900 km of the polar orbiting satellites is required to maintain the satellite in orbit while orbiting at the same rate as the earth. The GMS satellite has a Visible (0.5-0.75 pm) and Infrared (10.5- 12.5 pm) Spin Scan Radiometer (VISSR) which scans the earth from horizon to horizon every hour. The data is transmitted to Japan for processing before transmission to Melbourne, Australia for meteorological forecasting. Microwave. The microwave portion of the spectrum is divided into a series of bands known as C band (6 cm), S band (13 cm), L band ( 25 cm) and P band (68 cm). Passive remote sensing of microwave emissions from the earth's surface by the Special Sensor Microwave/Imager (SSM/I) on the DMSP (Defence Military Satellite Program) is used for ice monitoring and precipitation measurement at spatial resolutions of 10-100 km. Active remote sensing systems based on synthetic aperture radar (SAR) are used to get higher resolutions of 10-30 m. The shorter microwave wavelengths are more strongly absorbed by natural material and the longer wavelengths penetrate further into the soil and overburden. The amount of energy scattered back to the sensor depends on the dielectric constant of the surface and its surface roughness. A major application of SAR is in tropical areas wThere dense cloud can prevent observations by optical sensors. In these areas the main applications appear to be mapping geological structures and changes in land cover such as deforestation. Several satellites have been launched with SAR sensors, the main ones being the Japanese Earth Resources Satellite (JERS-1) with L band launched in 1994, the European Earth Re¬ sources Satellite (ERS-1) with C band launched in 1991 and the commercial Canadian RADARSAT-1 with C band, HH polarisation and variable look angle launched in 1995. Projected future uses for SAR data are monitoring of floods, ice fields and oil spillage dispersed on the water surface. To date no significant applications in Western Australia have emerged for SAR data. Other specialist radar sensors that might find oceano¬ graphic application in Western Australia are the altimeters and scatterometers. The altimeters on ERS-1 and TOPEX/ POSEIDON measure deviations of the sea's surface height from the theoretical geoid to give estimation of the gravity field below the ocean and ocean currents on the surface. The wind scatterometer on ERS-1 records the change in radar backscatter of the sea caused by the wind close to the surface. Wind direction is derived from the orientation of the backscatter relative to the orientation of the pulse of microwave radiation transmitted by the scatterometer. These data are becoming available opera¬ tionally and are being used in weather forecasting. NASA is using a wind scatterometer on Japan's Ad¬ vanced Earth Observing Satellite (ADEOS) to send wind data every 2 hours to US weather forecasters (Gibbs 1996). Scatterometer data can also be used to measure vegetation types (Long & Hardin 1994). Developments using airborne sensors The development and application of satellite sensors has been assisted by experience gained with airborne multispectral scanners such as the GEOSCAN which was built locally in Western Australia and widely used in mineral exploration and environmental monitoring. At a lower cost, a Digital Multi-Spectral Video (DMSV) which uses four CCD video cameras with filters giving wavebands typical of Landsat MSS for vegetation and water measurements has been developed locally. Dem¬ onstrating the potential value of SAR data from satellite, NASA has flown the Airborne Synthetic Airborne Radar (AIRSAR) instrument in Western Australia for experi¬ mental purposes. This instrument has P, L and C-band with four different polarisations (HH, HV, VH and W) and a duplicate set of sensors with L and C-bands at W polarisation (TOPSAR) for interferometry measurement of elevation. This multiple range of frequencies and polarisations will become available on later satellite SAR sensors. Use of satellite data for surface geological mapping complements a range of ground penetrating geophysical airborne instruments that are flown to map various below ground geological features. These instruments measure magnetics, conductivity (electro-magnetics) and gamma ray emissions (radiometrics). To aid geologic interpreta¬ tion of the distribution of weathering products in the landscape, it is common to merge airborne radiometric and Landsat-TM satellite data. Sensors of the future The application of satellite remote sensing is still young. Over the next 15 years another 80 missions are planned with over 200 sensors proposed (CEOS 1995). Capturing the benefits from these sensors will be a chal¬ lenge to researchers and applications scientists. Develop¬ ments in microelectronics creates new sensors, lower weight satellites, faster communications and more power¬ ful computers to process the data. Spatial resolution will increase from 10 m to <1 m creating digital photogram- metric applications. Hyperspectral data (e.g. 256 spectral bands with 0.08 pm bandwidth in the optical wave¬ lengths) will increase spectral resolution for enhanced applications in mineral detection and environmental monitoring. New radar satellites will have a wider range of microwave frequencies and polarisations. Without loss of spatial resolution, temporal resolution will be increased through the use of constellations of earth observing satel¬ lites to enable the monitoring of dynamic changes such as crop growth in agricultural fields and grass growth in rangelands. Capturing the economic and environmental benefits of these sensors for Western Australia will require a con¬ tinued growth in local research and development. Timely acquisition for monitoring may require the establishment of a local X-band reception capability. 20 Journal of the Royal Society of Western Australia, 80(1), March 1997 Development of terrestrial applications of remote sensing Given these technological opportunities, significant initiatives by Commonwealth and State Government agencies have helped develop applications of satellite remote sensing in Western Australia (Smith & Pearce 1997). The bibliography complied by Smith & Pearce (1997) indicates that the first recorded publication on ter¬ restrial applications was by Honey et al (1978) using Landsat MSS, for the classification of wetlands on the Swan coastal plain. This classification was based on the spectral differences of wetlands from the surrounding land. The next publication (Houghton 1979) reports the coordination of remote sensing activities which has been an important feature of the successful development of remote sensing in Western Australia where resources for the development of this new technology have been limited. Development of marine applications is covered by Pearce & Pattiaratchi (1997). Landcover type The bibliography records that landcover types mapped over the following 16 years using Landsat MSS and TM were mangroves (Honey & Hick 1981), tropical rainforests in the Kimberley (Kay et al. 1990), areas of remnant native vegetation in the agricultural area (Campbell & Wallace 1989), forest cover of water catchments (Wyllie & Barile 1990) and areas of gravel suitable for road building (Wyllie et al. 1992). The map¬ ping soon extended to geological structures using Landsat data (Smith & Green 1979) and subsequently NOAA-AVHRR (Honey & Tapley 1987; Tapley & Wilson 1986). There are many other significant applications to geology and exploration that are not covered in this review. The significance of this geological use is indicated by the fact that about 60% of all satellite data purchased in West¬ ern Australia is used for geological mapping or exploration. The outcomes appear not in scientific publications but as improved geological maps from the Western Australian Geological Survey and the finding of new mineral deposits by the exploration industry. High resolution satellite imagery is periodically or¬ dered by State agencies such as the Ministry of Planning, Department of Environmental Planning, Ministry of Transport, Department of Agriculture, Department of Conservation and Land Management and the Bush Fires Board for mapping urban development, status of wet¬ lands and new roads, status of National Parks and bush fire assessment. Land degradation Degradation has been the subject of a number of studies using classification or visual interpretation to map areas of agriculture affected by wind erosion (Carter & Houghton 1981), salinity (Currey et al 1981; Furby 1994; Wheaton 1992; Wheaton et al 1992) and waterlog¬ ging (Wallace & Wheaton 1990; McFarlane et al 1992) and areas of forest affected by dieback Phytophthera cinnamoni (Behn & Campbell 1992) and insect damage (Behn et al 1990). The success of the spectral technique for mapping salinity has been recently enhanced by inte¬ gration of satellite data with information derived from digital elevation data (Caccetta et al 1995a, b; Plate 4). Following the demonstration of the potential, the State Government plans to adopt this technology to map the extent of salinity and condition of remnant vegetation in the agricultural area to help combat these major environ¬ mental problems. The low albedo of native vegetation compared with agricultural crops (Smith et al 1992) means that spectral information in the visible, near infrared and mid infrared regions is ideally suited for mapping areas of remnant native vegetation in agricultural areas and water catchments (Wallace & Furby 1994). The first comprehensive mapping of changes in the area of native vegetation in the agricultural area from 1991-1991 using Landsat-TM data is currently being attempted. This work is being contracted by the Commonwealth Bureau of Resource Sciences to establish the sources of CO. causing the greenhouse effect. Water quality associated with land degradation has been successfully measured using spectral measures of water colour (Pattiaratchi et al 1991, 1992, 1994). Such measures are dependent on relatively smooth water to avoid the sunglint and turbidity caused by rough water. Sequential measures of sea surface water temperatures have been used to estimate currents in the Indian Ocean (McAtee 1992). Using the Coastal Zone Colour Scanner (CZCS) on the experimental Nimbus-7 satellite, measurement of ocean chlorophyll associated with land degradation has been demonstratedd by Pattiaratchi et al (1990). Future use of satellite measures of ocean chlorophyll in Western Aus¬ tralia awaits the anticipated launch of the Sea viewing Wide Field-of-view Sensor (SeaWiFS) in 1997 (Davies et al 1994). More detailed discussions of these applications in the marine environment is contained in Pearce & Pattiaratchi (1997). Atmosphere Operational use has been made of satellites in West¬ ern Australia to measure atmospheric conditions. The TOVS (Tiros Operational Vertical Sounder), a set of three instruments on the NOAA satellite, is used by the Bureau of Meteorology to measure longwave emittance in a large number of wavebands for a 60 km foot print. From the solution of the radiative transfer equation, the vertical distribution of temperature and moisture is derived for use in operational weather forecasts by the Bureau of Meteorology. Lynch & Marsden (1992) have also derived estimates of aerosol optical depths from NOAA-AVHRR data. NOAA-AVHRR also offers opportunity to measure cloud top temperatures and cloud climatology associated with changes in vegetation cover. For example, clearing of native vegetation and replacement by annual agricul¬ tural species has been observed to cause a decrease in formation of convective clouds (Lyons et al 1993; Lyons et al 1996). Further study of this process is occurring at Murdoch University using NOAA-AVHRR data. In Africa the duration of cloud top temperatures below a certain threshold determined from the European geosynchro¬ nous meteorological satellite is used to forecast the spatial distribution of rainfall in areas where rain gauges are sparse. Land surface processes Satellite imagery is used periodically by Remote Sens¬ ing Services, Department of Land Administration (RSS, 21 Journal of the Royal Society of Western Australia, 80(1), March 1997 DOLA) to analyse the impact of flooding events on rail, road and pipeline links to assist in assessing the suitability of these engineering structures as they cross the land¬ scape. These images have been enhanced using digital elevation models to create 3D perspectives of the distribu¬ tion of flood waters. These 3D perspectives are also widely sought after as an aid to geological interpretation (Davison 1992). The possibility of digital elevation ex¬ traction using stereo pairs from the SPOT satellite also exists but has not been used in Western Australia, where elevation information from aerial photography is readily available. Energy fluxes at the land surface over large areas can be studied by using NOAA-AVHRR data to help solve the energy balance equation of the surface. This approach has been used to study the impact of large scale clearing of native vegetation in south western Australia on sensible and latent heat fluxes (Lyons et al 1993, 1996). It was found that the higher albedo following clearing of land for agriculture caused lower sensible heat flux. This lower sensible heat flux causes reduced convection indi¬ cating a possible mechanism for the observed 20 to 30 % decline in rainfall that has followed large scale land clear¬ ing in south-western Australia. Monitoring and forecasting land cover changes: Accurate monitoring is a powerful information tool for increasing public awareness to achieve desired man¬ agement outcomes. Repeat observations by satellite sensors offer the possibility of monitoring changes in key envi¬ ronmental variables on a routine basis. Areas of bush fires, drought, salinisation, wind erosion, poor agricultural productivity, remnant native vegetation and plantations are examples of changes that can now be routinely moni¬ tored. From such changes the processes can be modelled, future outcomes forecast, public support enrolled and better management decisions taken. This forecasting application has been demonstrated by Caccetta et al. (1995) for salinity. Smith et al. (1995) for wheat yields, and Lyons et al. (1996) for the impact of clearing on rainfall. An example of monitoring and forecasting environ¬ mental change is the operational measurement of varia¬ tions in green vegetation cover at 14 to 16 day intervals across Western Australia using the NDVI from NOAA- AVHRR (Plate 3). Seasonal vegetation growth determines the outcome of later events such as fuel load build up Time from Germination - ► Figure 6. The hypothetical relationship between the seasonal growth of vegetative biomass of annual species and the result¬ ant output of grain, animal products and fuel load. that sustain fires, drought caused by feed deficit, crop yields and production of grazing animal, which can therefore be forecast (Fig 6). Monitoring of variations in seasonal vegetation growth of winter-growing Mediterranean annual species in the agricultural area of south-western Australia from NOAA-AVHRR data is shown in Figure 7. Every season, following the winter rains, there is a pronounced period of vegetation growth marked by a rise in the NDVI about May /June rising to a peak in September/ October (Fig 7). There is then a decline as the annual pastures and crops senesce. The variations in the peak NDVI across the grain belt of Western Australia are closely related to the grain yield of wheat (Fig 8). This close relationship between biomass around anthesis estimated by the NDVI in mid season and final grain yield can be used to forecast wheat yields in December (Fig 8). Similar relationships between pad- docks using NDVI calculated from Landsat-TM and SPOT XS data and wheat yields exist and are used to produce yield maps of farmer's fields (Stovold et al. 1996). Figure 7. Monthly NDVI time series from NOAA Global Area Coverage data from 1981-91 for a location in the central grainbelt of Western Australia. 22 Journal of the Royal Society of Western Australia, 80(1), March 1997 d Plate 1A,B,C,D. Comparison of images of the Port of Fremantle and the Swan River, Western Australia to indicate the change in spatial resolution from 1971 to 1986: (a) Landsat MSS (80m) (b) Landsat-TM (30m), (c) SPOT XS (20m) and (d) SPOT PAN (10m). 23 Journal of the Royal Society of Western Australia, 80(1), March 1997 Plate 2. An image map of sheet 2249 of the Australian 1:100,000 topographic series covering Mt Augustus produced from Landsat-TM bands 7,4,1 (Red, Green and Blue) produced for use in geological mapping. Mt Augustus, or Biirringurruah as it is known to the Wadjari Aboriginal people, is about 800 km north of Perth. It is one of the most spectacular solitary peaks in the world. It rises 717 metres above the surrounding plain and is about 8 km long. The dark coloured peak is evident in the lower right hand quarter of the image Plate 4. A typical application of Landsat-TM is to map areas of salinisation by classification. I his example is an area of the grain belt of Western Australia, 200 km south east of Perth covering an area of about 2,000 knv. The technique uses two successive Landsat-TM images during the winter period of crop growth and digital elevation data to classify areas affected by salinity. JANUARY APRIL MAY JULY Plate 3. A sequence of images of green vegetation cover of West¬ ern Australia based on the NDVI derived from successive over¬ passes of the NOAA-AVHRR sensor over a 14 to 15 day period using maximum value compositing. The area covered is some 252 million ha. Plate 5. Areas burnt by fires between April 1995 and March 1996 mapped from the NOAA-AVHRR sensor at 10 day inter¬ vals. 24 NDVI NDVI Journal of the Royal Society of Western Australia, 80(1), March 1997 Figure 8. Relationship between the mean NDVI in September 1995 from NOAA-AVHRR data and wheat yield (tonnes ha1) of 68 local government areas in the Western Australian grain belt (Smith et al. 1995). Figure 9. Monthly NDVI time series from NOAA-AVF1RR from 1981-91 for the Kimberley area of north¬ western Australia (note the change in scale from Fig 8). Months (Jan 1 981 )=0 Figure 10. NDVI time series from NOAA-AVHRR GAC data from the central desert area of Western Australia (note the change in scale from Figs 8 and 9). 25 Journal of the Royal Society of Western Australia, 80(1), March 1997 In north western Australia, which experiences summer monsoonal rains, there is a similar pattern of seasonal vegetation growth with a lower range in NDVI begin¬ ning in December and ceasing in about April (Fig 9). The seasonal distribution of this information is used by the Bush Fires Board in Western Australia to monitor the rapid build up in fuel load which provides the basis for extensive bush fires ignited by a variety of natural and human causes. It is also used to plan controlled burning by the Bush Fires Board. In comparison, the seasonal growth in the inland areas of Western Australia, with lower and more variable rain¬ fall, is much less and more varied (Fig 10). Comparison of current seasonal trends in the NDVI with the long term trend has been used to map areas affected by drought in pastoral areas (Cridland et al 1994). Near real-time mapping of bush fires for the Bush Fires Board covering the whole of Western Australia has been implemented (Smith et al 1996; Plate 5). The out¬ come of this fire monitoring in 1995/96 was the detection 1995/96 Figure 11. Daily number of suspected bushfires in Western Australia detected from the NOAA-AVHRR satellite in 1995/96 by Remote Sensing Services, DOLA, whose location was reported to the Bush Fires Board within 4 hours of the overpass. 1995/96 Figure 12. Area burnt by bushfires in Western Australia that was mapped from NOAA-AVHRR in 1995/96. Total area burnt is estimated at 20.6 million ha. 26 Journal of the Royal Society of Western Australia, 80(1), March 1997 of over 2,700 bush fires reported to the Bush Fires Board by fax or internet within 4 hours of the data being re¬ ceived (Fig 11). The mapping indicated a total area burnt of about 20.6 million hectares, most in the north west of the State (Fig 12). This is the first time that such compre¬ hensive information on bushfires has been available and is an example of the type of spatial and temporal geo¬ graphic information that can now be made available from satellites. The availability of these data on a routine basis in near real-time to the Bush Fires Board of Western Australia has had a significant impact on enrolling community support for the management of Bush Fires in the Kimberley. It has transformed a "we can do nothing" attitude to a "we can do". Instead of chasing bush fires, the Bush Fires Board reports that land managers are now able to see ahead and plan strategies that will provide more effective control and less risk. Conclusions Earth observations by satellite offers significant potential for using space to improve exploration and the manage¬ ment of Western Australia's renewable resources. After 20 years of research, techniques have progressed from the photo-interpretation of satellite imagery to the routine extraction of geographic information using automatic methods. The knowledge base for the much wider use of this technology has been created, but the challenge of widespread adoption remains. Land cover types can be mapped and the spatial vari¬ ability of certain key land surface processes measured from space. Monitoring changes in area of remnant native vegetation and salinity in the agricultural area using high resolution data, is now possible. Public access to such information provides a powerful means for Government to raise public awareness and support for extending the area of native vegetation. Application of NOAA-AVHRR to routinely monitor bush fires over West Australian has demonstrated that such systematic use of satellite infor¬ mation would improve the ability of Government to implement policies to combat environmental degradation. Increased spatial, spectral and temporal resolution of future sensors will expand the opportunities for exploration and monitoring the environment. Ongoing Government investment into methods for capturing information from these new sensors, including X band reception capability in Western Australia for timely coverage may be required. Continued CSIRO research to generate the core knowledge to enable geographic information to be extracted from these new forms of data will be important. The advent of ocean colour sensors and radar altim¬ eters and radar scatterometers as a complement to existing sea surface temperature data will greatly enhance our ability to manage Western Australia's surrounding oceans and coastal zone. We are at the beginning of the next era in earth obser¬ vations from space which could give rise to many new applications of satellite remote sensing to the successful exploration and the management of the renewable re¬ sources of Western Australia. Acknowledgements: The assistance and advice of members of the Leeuvvin Centre for Earth Sensing Technologies and the WA Satellite and Technology Applications Consortium, and in particular staff of Remote Sensing Services, DOLA (previously RSAC, DOLA), CSIRO, World Geo¬ sciences Corporation, Curtin University of Technology and Bureau of Meteorology is greatly appreciated. References Behn G A & Campbell N 1992 Dieback assessment, using multi- spectral data, over the Stirling Range National Park, Western Australia. Proceedings of the 6,h Australasian Remote Sensing Conference, Wellington, New Zealand, 1:69-178. Behn G A, Wallace J F & Flick P T 1990 Airborne multispectral scanner data for mapping insect damage in Jarrah forests of Western Australia. Proceedings of the 5th Australasian Remote Sensing Conference, Perth, 1032-1039. Buitens FI J & Clevers J C. P W, 1993 Land observation by Remote Sensing - Theory and Applications. Gordon & Breach Science Publishers. Caccetta P A, Campbell N A, West G A, Kiiveri FI T & Gahegan M 1995a Aspects of reasoning with uncertainty in an agricul¬ tural GIS environment. The New Review of Applied Expert System 1:161-177. Caccetta P A, Kiiveri H T, Evans F FI & Ferdowsian R 1995b A knowledge based approach to predicting salinity in the West¬ ern Australian wheatbelt. Proceedings of the 16th Asian Con¬ ference on Remote Sensing, Thailand, 13-3-1 to 13-3-6. Campbell N A & Wallace J F 1989 Cover class mapping in agricultual environments. Proceedings of the International Geoscience and Remote Sensing Symposium, 791-794. Carter D J & Houghton H J 1981 Remote sensing of wind erosion in croplands. In: Proceedings of Landsat-81 Conference, Canberra, 275-282. CEOS 1995 Committee on Earth Observations Satellites: Coordi¬ nation for the next decade (1995 CEOS Yearbook). Smith System Engineering Ltd, UK. pp. 133. Cridland S W, Burnside D G & Smith R C G 1994 Use by Man¬ agers in Rangeland Environments of Near Real-time Satellite Measurement of Seasonal Vegetation Response. Proceedings of the 7th Australasian Remote Sensing Conference, Melbourne, Victoria, Volume 2, 1134-1141. Currey D T, Wilson M A & O'Callaghan J F 1981 Mapping salinised land from Landsat. Proceedings of the 2nd Australasian Remote Sensing Conference, Canberra, 8:1-4. Davies J E, Fearns P, Lynch M J & Pearce A F 1994 SeaWiFS product development and validation program: Status Report. PORSEC '94: Proceedings of the Second Pacific Ocean Remote Sensing Conference, Melbourne, 267-274. Davison P J N 1992 Synthetic-stereo satellite imagery. Proceedings of the 6'1, Australasian Remote Sensing Conference, Wellington, New Zealand. 1, 430-433. D'Souza G, Belward A S & Malingreau J-P 1993 Advances in the Use of NOAA AVHRR Data for Land Applications. Kluwer Academic Publishers. Furby S L 1994 Discriminating between pasture and barley grass and saltbush using multi-temporal imagery. CSIRO Division of Mathematics and Statistics Technical Report, Perth. George R 1990 The 1989 saltland survey. J. Agriculture Western Australia 31:159-166. Gibbs W W 1996 Where the Wind Blows. Scientific American, December 1996:44. Honey F R & Hick P T 1981 Application of Landsat imagery to mangrove mapping in north-west Western Australia. Pro¬ ceedings of the 2nd Australasian Remote Sensing Conference, 9.4 1-5. Honey F R, Hick P T & Blatchford D R 1978 Multilevel inventory and monitoring of wetlands on the Swan Coastal Plain, Western Australia. Proceedings of the 12,h International Symposium on Remote Sensing of Environment. Manila Environmental Research Institute of Michigan. 2027-2044. 27 Journal of the Royal Society of Western Australia, 80(1), March 1997 Honey F R & Tapley I J 1987 Regional, structural and litho¬ graphic mapping using NOAA-AVHRR imagery. Proceed¬ ings of the 3rd Australasian Remote Sensing Conference, Gold Coast, Queensland, 44-47. Houghton H J 1979 Coordination of Landsat activities in West¬ ern Australia. Proceedings of the 1st Australasian Remote Sensing Conference, Sydney, 454-464. Kay R J, Hick P T & Houghton H J 1990 Mapping tropical rainforests in Western Australia using multi-scene and multi¬ temporal Landsat thematic mapper data. Asian-Pacific Remote Sensing Journal 3(l):23-33. Legg C A 1992 Remote Sensing and Geographic Information Systems - Geological Mapping, Mineral Exploration and Mining. John Wiley & Sons in association with Praxis Pub¬ lishing Chichester, UK. 166 pp. Lillesand TM & Kiefer 1987 Remote Sensing and Image Inter¬ pretation. John Wiley & Sons, New York. Long D G & Hardin P 1994 Vegetation studies of the Amazon Basin using Enhanced Resolution Seasat Scatterometer data. Insitute of Electrical and Electronic Engineers, Transactiosn Geoscience and Remote Sensing 32(2):449-460. Lynch MJ & Marsden AJ 1992 Mt Pinatubo impact on aerosol optical depths in southern latitudes. Proceedigs of 6th Australasian Remote Sensing Conference, Wellington, New Zealand, 3:213-216. Lyons T J, Schwerdtfeger P, Hacker J M, Foster I J, Smith R C G & Huang Xinmei 1993 Land-atmosphere interaction in a semiarid region: The bunny fence experiment. Bulletin of the American Meteorological Society 74:1327-1334. Lyons T J, Smith R C G & Huang Xinmei 1996 The impact of clearing for agriculture on the surface energy budget. Inter¬ national Journal of Climatology 16:551-558. McAtee B 1992 Ocean current estimation using remote sensing. Honours Thesis. Curtin University of Technology, Perth. McFarlane D J, Wheaton G A, Negus T R & Wallace J F 1992 Effects of waterlogging on crop and pasture production in the Upper Great Southern, Western Australia. Technical Bulletin 86. WA Department of Agriculture, pp 45. Pattiaratchi C B & Hick P T 1992 Oceanic chlorophyll report. CSIRO/UWA Joint Research Report CWR, Perth, 16 pp. Pattiaratchi C B, Lavery P S, Wyllie A & Hick P T 1991 Remote sensing of water quality in Cockburn Sound: development of multi-temporal algorithms. University of Western Australia, Centre for Water Research Report WP607CP, Perth, 48 pp. Pattiaratchi C B, Lavery P S, Wyllie A & Hick P T 1994 Estimates of water quality in coastal waters using multi-date Landsat Thematic Mapper Data. International Journal of Remote Sensing 15:1571-1584. Pattiaratchi C B, Lavery P, Wyllie A & Hick P T 1992 Multi-date algorithms for predicting surface water quality parameters in estuarine and coastal waters using Landsat TM data In: Remote sensing environmental monitoring and resource management: European 'Journal Space Year' Conference 1992 (ed T D Guyenne). J J Hunt Paris: European Space Agency 2, 709-713. Pattiaratchi C B, Parslow J, Pearce A F & Hick P T 1990 application of Coastal Zone Colour Scanner (CZCS) imagery for productivity and circulation studies of the Leeuwin Current, Western Australia. Proceedings of the 5,h Australasian Remote Sensing Conference, Perth, 252-256. Pearce AF & Pattiaratchi CB 1997 Applications of satellite remote sensing to the marine environment in Western Australia. Journal of the Royal Society of Western Australia 80:1-14. Richards J A 1986 Remote Sensing Digital Analysis - An Intro¬ duction. Springer-Verlag, Berlin. Roderick M L, Smith R C G & Cridland S 1996a The precision of the NDVI derived from AVHRR observations. Remote Sensing Environment 56: 57-65. Roderick M L, Smith R C G & Lodwick G D 1996b Calibrating long term AVHRR derived NDVI imagery. Remote Sensing Environment 58:1-12. Smith R C G, Xinmei Huang, Lyons T J, Hacker J M & Hick P T 1992 Change in land surface albedo and temperature in south Western Australia following the replacement of native perennial vegetation by agriculture: satellite observations. Paper 1AF-92-0117. Proceedings of the 43rd Congress of the International Astronautical Federation, Washington, D C. 10 pp. Smith R C G, Wallace J F, Hick P T, Gilmour R F, Belford R K, Portmann P A, Regan K L & Turner N C 1993 Potential of using field spectroscopy during early growth for ranking bio¬ mass in cereal breeding trials. Australian Journal of Agricul¬ tural Research 44:1713-1730. Smith R C G, Adams J, Stephens DJ & Hick P T 1995 Forecasting wheat yield in a Mediterranean-type environment from the NOAA satellite. Australian Journal of Agricultural Research 46:113-25. Smith R C G, Browne C E, Craig R L, Adams J & Steber M 1996 Satellite monitoring of bush fires in north Western Australia. Proceedings of the 8th Australasian Remote Sensing Conference, Canberra ACT. CD ROM Records 9036-9058. Smith RCG & Pearce AF 1997 Introduction to the bibliography of research into the remote sensing of land, sea and atmo¬ sphere in Western Australia. Journal of Royal Society of Western Australia 80:29-39. Smith R E & Green A A 1979 Application of Landsat to the interpretation of volcanic relations and metasomatic alter¬ ation, Maddina Volcanics, Western Australia. Proceedings of the 1st Australasian Remote Sensing Conference, Sydney, 277-294. Stovold R G H, Smith RCG, Allen A &: Evans F 1996 A high resolution satellite information product for farm and catch¬ ment management in Western Australia. Proceedings of the 8,h Australasian Remote Sensing Conference, Canberra, ACT. CD-ROM Records 4423-1478. Tapley I J & Wilson P 1986 NOAA-AVHRR Imagery for palaeodrainage and lineament mapping in South Australia's north-west geological provinces. Proceedings of the 1st Australian AVHRR Conference, Perth, Western Australia, 222-236. Wallace J F & Furby S L 1994 Assessment of change in remnant vegetation area and condition. Report. Land and Water Resources Research and Development Corporation, Perth, 27 pp. Wallace J F & Wheaton G A 1990 Mapping the extent of water¬ logged crops using satellite imagery. WA Journal of Agriculture 31:48-50. Wheaton G A, Wallace J F, McFarlane D J & Campbell N A 1992 Mapping salt-affected land in Western Australia. Proceed¬ ings of the 6rh Australasian Remote Sensing Conference, Wellington, New Zealand, 2:369-3 77. Wyllie A & Barile P 1990 Rapid clearing assessment using multi- spectral scanner data. Proceedings of the 5’1' Australasian Remote Sensing Conference, Perth, 1079-1082. Wyllie A, Buchanan A, Hick P T & Butkas F 1992 The use of Thematic Mapper to detect road building materials. Proceed¬ ings of the 6th Australasian Remote Sensing Conference, Wellington, New Zealand, 2:348-354. Xinmei Huang, Lyons T J, Smith RCG, Hacker J M Schwerdtfeger P 1993 Estimation of surface energy balance from radiant surface temperature and NOAA-AVHRR sensor reflectances over agricultural and native vegetation. Journal of Applied Meteorology 32:1441-1449. 28 Journal of the Royal Society of Western Australia, 80:29-39, 1997 A bibliography of research into satellite remote sensing of land, sea and atmosphere conducted in Western Australia RCG Smith1 & AF Pearce2 ’Remote Sensing Services, Department of Land Administration: richardsmith@notes. dola.wa.gov.au and 2CSIRO Division of Marine Research: alan.pearce@marine.csiro.au Leeuwin Centre for Earth Sensing Technologies, 65 Brockway Road, Floreat WA 6014 Manuscript received February 1997 Background to the bibliography This bibliography on satellite and related remote sensing research in Western Australia covers both geological mapping for exploration and the monitoring of land, water, oceans and atmosphere. Publications related to exploration commissioned by the private sector and not publicly available have not been cited. The number of publications over time (Figure 1) reveals progressive development with alternate year peaks in recent years associated with the biennial Australasian Remote Sensing Conference. The development of satellite remote sensing in West¬ ern Australia has been driven by the need of a small population to explore and manage efficiently a large and complex land surface. Western Australia's population of 1.7 million occupies a vast land area of 253 million hectares, with over 12,000 km of coastline and 40 million ha of continental shelf which covers part of Australia's exclusive economic zone. The economy is dependent on mining and other primary industries, while environmental quality requires preservation of vegetative cover for maintaining bio-diversity and avoidance of desertification. Most of the land (93%) is under the ownership of the State. This State-owned land is allocated to conservation and timber extraction (17%), arid deserts, road reserves and fore¬ shores (36%), pastoral grazing (37%), mining 3%, etc. Freehold land is only 7% and is used for agriculture (6%) and urban settlement (1%). Unfortunately, 167 years of European settlement has caused much of the land, rivers and coastal waters to suffer degradation from bush fires, salinisation, erosion, disease, acidification, feral animals, eutrophication, silt- ation, pollution and resource depletion. Such degradation ranges from minimal to severe, but its actual extent and rate of spread is often not known because of the diffi¬ culty of monitoring the trends and changes over vast 60 T Year Figure 1. Progress of publications in remote sensing in Western Australia © Royal Society of Western Australia 1997 29 Journal of the Royal Society of Western Australia, 80(1), March 1997 areas. Absence of accurate information results in the problems being poorly managed. Therefore the Government in Western Australia needs access to comprehensive infor¬ mation on trends in land, water, atmosphere and coastal conditions to fulfill its environmental responsibilities. On land there is evidence of increasing use being made of satellite remote sensing to provide this information. Over the vast expanse of ocean there is evidence of use by marine scientists of satellite remote sensing for bathym¬ etry, marine habitat mapping, estuarine water quality, ocean circulation and thermal structure, as well as by the fishing industry in fisheries operations. To maintain the continued discovery of new world class ore deposits under WA's deeply weathered regolith, new airborne geophysical and satellite remote sensing exploration techniques continue to be developed. Commonwealth and State Government initiatives to capture the benefits of satellite remote sensing in Western Australia The first operational earth resources satellite Landsat-1 was launched in 1971 and CSIRO formed a remote sens¬ ing group in Western Australia in 1974. CSIRO's initial emphasis was on spectral radiometry measurements in the visible and near infrared wavebands, which helped provide the bio-physical basis of multi-spectral remote sensing. Recognising the importance of multi-spectral data for mapping land surface features not captured by use of air photography, the WA Department of Lands and Surveys appointed remote sensing officers in 1975, who in 1983 formed the Remote Sensing Applications Centre (RSAC) within the Department of Lands and Sur¬ veys (later named Department of Land Administration, DOLA). Formation of RSAC focused remote sensing skills thus providing a resource base for the widespread use of this technology by a large number of Government agencies. In the mid 1980s the CSIRO remote sensing group ex¬ panded as more disciplines saw research opportunities. Statisticians developed statistical classification of multi- spectral satellite data for mapping different land cover types and conditions. Geologists developed new satellite based methods to assist in geological mapping. Physicists applied space borne techniques for measurement of sea surface temperature and land surface evaporation. Biologists used satellites to measure biophysical variables from space such as biomass, crop yields and algal blooms. In contrast to the State's focus of its remote sensing efforts within DOLA, CSIRO's multi-disciplinary group is dis¬ tributed across several Divisions and Institutes. Universities By the mid 1970s the first tertiary courses in remote sensing were being taught at the Western Australian In¬ stitute of Technology (later Curtin University of Technology) in the Schools of Electrical Engineering, Applied Physics, and Surveying and Land Information with assistance from CSIRO. One of the outcomes of this interaction was the development of an L band receiving station for the NOAA satellite. The first images were received in 1981, using a World War 2 bofors gun mount and oscilloscope to guide the receiving dish. This system was upgraded in 1987 and operations taken over by the WA Satellite Tech¬ nology and Applications Consortium (WASTAC) of the Bureau of Meteorology, DOLA, CSIRO and Curtin Uni¬ versity of Technology. Teaching of tertiary courses in satel¬ lite remote sensing now occurs in all WA Universities and at the College of Technical and Further Education (TAFE). The Leeuwin Centre for Earth Sensing Technologies Successful application of satellite remote sensing requires a diversity of skills. With scientists and funds scarce, early development of remote sensing in WA was marked by close collaboration between the different remote sensing groups within CSIRO, Curtin University, DOLA, Bureau of Meteorology and the private sector. In 1993, the im¬ portance of this collaboration was officially recognised when the Leeuwin Centre for Earth Sensing Technologies was built for $5 million by the State Government on the CSIRO Floreat campus. The building co-located CSIRO's Remote Sensing Group, DOLA's Remote Sensing Services, Curtin University's Department of Applied Physics Remote Sensing Group, TAFE and World Geoscience Corporation, the largest airborne geophysical exploration company in the world. The bibliography of papers, reports, theses and con¬ ference presentations listed here reveals a steady devel¬ opment in the application of satellite remote sensing to exploration and the management of Western Australia's renewable resources. The Leeuwin Centre brings together for the first time in Australia the three technologies of airborne geophysics, satellite remote sensing and geographic information systems. Co-location offers the opportunities to expand greatly the applications of satellite remote sensing through the synergy of the participating groups. After 20 years of development, WA now has for the first time the capability to map and monitor major environ¬ mental changes occurring in salinity, water quality, land productivity and native vegetation and use this informa¬ tion to enrol community support to tackle these problems. The challenge for these remote sensing groups is to create the synergy which will realise the full potential of the Leeuwin Centre and capture the benefits from the next generation of satellite sensors and associated technologies. 30 Journal of the Royal Society of Western Australia, 80(1), March 1997 Bibliography Satellite remote sensing of land, sea and atmosphere in Western Australia This bibliography covers all Western Australian satellite and airborne optical remote sensing zvork comprising published research , reports, theses and conference proceedings. Adams J 1994 Bushfire devastation on the Australian continent. International Journal of Remote Sensing 16:1573-1575. Adams J 1995 Wildfire monitoring in New South Wales, Australia. International Journal of Remote Sensing 15:3641-3642. Agar RA 1992 Geological applications and limitations of air¬ borne multi-spectral scanners. Their availability, cost-effec¬ tiveness and future. Proceedings of 6th Australasian Remote Sensing Conference. The Conference Committee, Wellington, New Zealand. Albertz J, Li Z & Zhang W 1994 Rectification of airborne line- scanner imagery utilising flight parameters. Proceedings of First International Airborne Remote Sensing Conference, Strasbourg, France, 447-456. Baines PG 1981 Satellite observations of internal waves on the Australian North-West Shelf. Australian Journal of Marine and Freshwater Research 32:457-463. Barton IJ & Penrose JD 1988 Experimental investigations of bulk/skin temperature differences in the open ocean. Abstracts of the 2nd Conference on Air-Sea Interaction and its Conse¬ quences. The Conference, Merimbula, New South Wales. Behn GA 1996 An evaluation of the efficiency of remote sensing and geographical information system technologies for mapping and monitoring dieback. MSc Thesis. Curtin University, Perth. Behn GA & Campbell N 1992 Dieback assessment, using multi- spectral data, over the Stirling Range National Park, Western Australia. Proceedings of 6,h Australasian Remote Sensing, 1:169-178. The Conference Committee, Wellington, New Zealand. Behn GA & Campbell N 1995 Dieback assessment from historical air photos. Proceedings of 9th Annual Symposium, 2:683-687. Vancouver, Canada. Behn GA, Wallace JF & Hick PT 1990 Airborne multispectral scanner data for mapping insect damage in jarrah forests of Western Australia. Proceedings of the 5th Australasian Remote Sensing Conference, 1032-1039. The Conference, Perth. Behn GA, Wallace JF & Hick PT 1990 Remote sensing of dieback on the south coast. CSIRO Division of Exploration Geo¬ science, Restricted Report. 10 pp. Conservation and Land Management, Land Information Branch, Perth. Bellairs SM, Hick PT, Harvey, E Turner NC, Belford RK & Smith RCG 1993 Measuring cereal crop biomass using field radi- ometry. Proceedings of the 10th Plant Breeding Conference, 107-108. Conference Organising Committee, Canberra. Bellairs SM, Turner NC, Hick PT & Smith RCG 1996 Plant and soil influences on estimating biomass of wheat in plant breeding plots using field spectral radiometers. Australian Journal of Agricultural Research 47:1017-1034 Bierwirth PN, Lee T & Bume RV 1991 Unmixing of shallow sea¬ floor reflectance and water depth using multispectral imagery. Proceedings of the Conference on Remote Sensing and Geo¬ graphical Information Systems for Coastal Catchment Manage¬ ment, 223-242. Conference Committee, Lismore, New South Wales. Bierwirth PN, Lee T & Bume RV 1992 Shallow water mapping via the separation of depth and substrate components from multispectral data: an example from Useless Inlet, Shark Bay, WA. Proceedings of the 6,h Australasian Remote Sensing Conference, 1:99-109. The Conference Committee, Wellington, New Zealand. Bierwirth PN, Lee T & Bume RV 1993 Shallow sea-floor reflec¬ tance and water depth derived by unmixing multispectral imagery. Photogrammetric Engineering & Remote Sensing 59:331-338. Boniface P & Craig R 1986 Geometric correction of NOAA- AVHRR. Proceedings of the T* Australian AVHRR Confer¬ ence, 299-306. CSIRO Division of Groundwater Research, Perth. Buchanan A, Bevan A & Shaw R 1992 Australian impact structures - a photographic catalogue. Proceedings of 6th Australasian Remote Sensing Conference, 1:238-2482-6. The Conference Committee, Wellington, New Zealand. Caccetta PA, Campbell NA & West GA 1995 A massively parallel implementation of an image classifier. Proceedings of the 16th Asian Conference on Remote Sensing, 5-5-1 to 5-5-6. Organising Committee of World Technology, Thailand. Caccetta P, Campbell NA, West GA, Kiiveri HT & Gahegan M 1995 Aspects of reasoning with uncertainty in an agricultural GIS environment. The New Review of Applied Expert Systems 1:161-177. Caccetta PA, Kiiveri HT, Evans FH & Ferdowsian R 1995 A knowledge based approach to predicting salinity in the West Australian wheatbelt. Proceedings of the 16th Asian Conference on Remote Sensing, 13-3-1 to 13-3-6. Organising Committee of World Technology, Thailand. Campbell NA, DeBoer ES & Hick PT 1987 Some observations on crop profile modelling. International Journal of Remote Sensing 8:193-201. Campbell NA & Furby SL 1994 Variable selection along canonical vectors. Australian Journal of Statistics 36:177-183. Campbell NA, Furby SL & Fergusson B 1994 Calibrating images from different dates. Report. 16 pp. Land & Water Resources Research & Development Corporation, Perth. Campbell NA, Honey FR, Hick PT & Carlton MDW 1982 Evalu¬ ation and development of techniques for crop inventory in the wheatbelt of Western Australia using satellite data. Pro¬ ceedings of the 16,h International Symposium on Remote Sensing of Environment, 607-616. Environmental Research Institute of Michigan, Ann Arbor, Michigan. Campbell NA, Honey FR, Hick PT & Carlton MDW 1982 Evalua¬ tion of the application of Landsat data to crop discrimination in Western Australia. Proceedings of the International Sym¬ posium on Machine Processing of Remotely Sensed Data, 88- 96. Purdue University Laboratory for Applications of Remote Sensing, West Lafayette, Indiana. Campbell NA & Wallace JF 1989 Statistical methods for cover class mapping using remotely sensed data. Proceedings of the International Geoscience and Remote Sensing Sympo¬ sium, 493-496. IG ARSS, Vancouver, Canada. Campbell NA & Wallace JF 1989 Cover class mapping in agri¬ cultural environments. Proceedings of the International Geo¬ science and Remote Sensing Symposium, 791-794. IGARSS, Vancouver, Canada. Carroll W 1982 The WAIT satellite receiving station. Conference on Applications of Environmental Satellites, 6 pp. CSIRO Division of Groundwater Research, Perth. Carroll W, Cargill RD & Honey FR 1981 A low cost NOAA/ TIROS AVHRR receiving station. Proceedings of the 15th International Symposium on Remote Sensing of Environment, 1253-1264. Environmental Research Institute of Michigan, Ann Arbor. 31 Journal of the Royal Society of Western Australia, 80(1), March 1997 Carter DJ & Houghton HJ 1981 Remote sensing of wind erosion in croplands. Proceedings of Landsat - 81 Conference, 275- 282. Organising Committee, Landsat 81, Canberra. Catalano P, Eliot I & Wyllie A 1992 Creation of a broad scale planning atlas from a Thematic Mapper base. Proceedings of 6th Australasian Remote Sensing Conference, 1:426-429. The Conference Committee, Wellington, New Zealand,. Catalano P, Wyllie A & Eliot I 1991 Development of a coastal planning information system: Guilderton to Dongara, West¬ ern Australia. Proceedings of the Conference on Remote Sensing and Geographical Information Systems for Coastal Catchment Management, 205-222. The Organising Commit¬ tee, Lismore, New South Wales. Comer RJ 1992 A potential method of soil spectral reflectance recovery for cropped agricultural areas. Proceedings of 6th Australasian Remote Sensing Conference, 3:80-89. The Con¬ ference Committee, Wellington, New Zealand. Comer RJ 1992 Passive spectral bathymetry using satellite remote sensing in Cockbum Sound, Western Australia. MSc Thesis. Curtin University of Technology, Perth. Craig RL, Evans FL, Smith RCG, Steber MT & Wyllie A 1995 The use of NOAA-AVHRR data for the detection of bushfires in Western Australia. Proceedings of the 2nd North Australian Remote Sensing and Geographic Information Systems Forum, 27-34. The Organising Committee, Darwin, NT. Cresswell GR & Peterson JL 1993 The Leeuwin Current south of Western Australia. Australian Journal of Marine and Fresh¬ water Research 44:285-303. Cridland SW, Burnside DG & Smith RCG 1994 Use by managers in rangeland environments of near real-time satellite measure¬ ment of seasonal vegetation response. Proceedings of 7th Australasian Remote Sensing Conference, 2:1134-1141. Remote Sensing and Photogrammetry Association of Australia, Melbourne. Cridland S, Smith RGC & Burnside D 1993 VEGETATION WATCH: Development for agriculture. Report. 16 pp. Western Australian Pastoral Board/Department of Land Administra¬ tion, Remote Sensing Applications Centre, Perth. Currey DT, Wilson MA & O'Callaghan JF 1981 Mapping salinised land from Landsat. Proceedings of the 2nd Australasian Remote Sensing Conference, 8:1-4. Organising Committee Landsat 81, Canberra. Davies JE, Feams P, Lynch MJ & Pearce AF 1994 SeaWiFs product development and validation program: Status Report. PORSEC '94: Proceedings of the Second Pacific Ocean Remote Sensing Conference, 267-274. Bureau of Meteorology Re¬ search Centre, Melbourne. Davies JE & Lynch MJ 1993 Satellite remote sensing of ocean colour: Laying out the problem. Abstract. 4,h Australian Institute of Physics, Jarrahdale. Davison PJN 1992 Synthetic-stereo satellite imagery. Proceedings of 6,h Australasian Remote Sensing Conference, 1:430-433. The Conference Committee, Wellington, New Zealand. Domenikiotis C, Lodwick GD & Wright GL 1994 Reconstruction of a remotely sensed road network using an expert system approach. Proceedings of 7"’ Australasian Remote Sensing Conference, 1:448-455. Remote Sensing and Photogrammetry Association of Australia, Melbourne. Easton A, Pearce AF & Buchan SJ 1992 Modelling the Kirki spill using OSSM. Proceedings of the Third National Scientific Support Coordinators Workshop, 93-96. Dept. Marine and Harbours, Perth. Evans FH, Caccetta PA, Ferdowsian R, Kiiveri HT & Campbell NA 1995 Predicting salinity in the upper Kent River Catch¬ ment. Report. 27pp. Land & Water Resources Research & Development Corporation, Perth. Evans FH, Ferdowsian R & Campbell NA 1996 Predicting salinity in the Wadjekanup and Byenup Hill Catchments. Report. 12pp. Land & Water Resources Research & Development Corporation, Perth. Evans F, Wyllie A & Kierath P 1994 Barrow Island engineering and rehabilitation asessment. Proceedings of 7,h Australasian Remote Sensing Conference, 2:1054-1059. Remote Sensing and Photogrammetry Association of Australia, Melbourne. Fletcher WJ & Tregonning RJ 1993 Distribution and timing of spawning by the Australian pilchard (Sardinops sagax neopilcfmrdus) off Albany, Western Australia. Australian Journal of Marine and Freshwater Research 43:1437-1449. Fletcher WJ, Tregonning RJ & Sant GJ 1994 Interseasonal varia¬ tion in the transport of pilchard eggs and larvae off southern Western Australia. Marine Ecology Progress Series 111:209- 224. Fletcher WJ, Tregonning RJ, Sant GJ, Blight SJ & Rossbach MH 1992 Investigation of the abundance and distribution of pil¬ chard eggs and larvae off southern Western Australia. Project 91/24 Report. 74pp. Fisheries Industry Research & Develop¬ ment Corporation, Perth. Furby SL 1995 Operational cross-calibration of AVHRR using robust regression procedures. Interim Report 1, Cross-Cali¬ bration of NOAA 14 to NOAA 11 a first attempt. Report. 21pp. Department of Land Administraton, Perth. Furby SL, Kiverii H& Campbell NA 1990 The analysis of high dimensional spectral curves. Proceedings of the 5th Australasian Remote Sensing Conference, 175-184. The Con¬ ference, Perth. Furby SL & Palmer M 1993 Automatic control point selection for rewarping NOAA images: Investigation of a simple corre¬ lation algorithm. Report. 20pp. Department of Land Admin¬ istration, Perth. Furby SL, Palmer MJ & Campbell NA 1996 Image calibration to like values. Proceedings of the 8th Australasian Remote Sensing Conference, 279-285. Remote Sensing and Photogrammetry Association of Australia, Canberra. Furby SL, Wallace JF, Caccetta PA & Wheaton GA 1995 Detecting and monitoring salt-affected land. Report. 71pp. Land and Water Resources Research and Development Corporation, Perth. Gardiner DB & Wright GL 1992 Land form mapping using remote sensing and GIS. Proceedings of 1st Australian Conference on Mapping and Charting, 103-112. The Conference, Adelaide. Gee L & Forster BC 1984 Landsat data as an aid to hydrographic and bathymetric mapping in shallow' seas. Proceedings of the 3rd Australasian Remote Sensing Conference, 161-173. Organising Committee, Brisbane. Gozzard JR & Tapley, IJ 1992 Landform and regolith mapping in the Lawlers district. Report 2: Terrain classification mapping. CSIRO IMEC Division of Exploration Geoscience Report 240R. Gozzard JR, Tapley IJ, Grunsky EC & Furby SL 1992 Regolith- landform mapping in the Lawlers district, Report 3: Integration of digital elevation data and Landsat thematic mapper data for improved terrain classification. CSIRO IMEC Division of Exploration Geoscience Report 241 R. Griffiths RW & Pearce AF 1985 Satellite images of an unstable warm eddy derived from the Leeuwin Current. Deep-Sea Research 32:1371-1380. Griffiths RW & Pearce AF 1985 Instability and eddy pairs on the Leeuwin Current south of Australia. Deep-Sea Research 32:1511-1534. Hardstaff S 1992 Mapping bushfires in the Kimberley Region using NOAA-AVHRR data, cartography. BSc Hons Thesis. Curtin University of Technology, Perth. Hearn CJ & Pearce AF 1985 NOAA satellite imagery and air¬ borne remote sensing of a small-scale tidal jet. Australian Journal of Marine & Freshwater Research 36:643-653. Hick PT 1979 Remote sensing techniques applied to an estuarine environment problem in Western Australia. Australian Journal of Instrumentation and Control 40:4-6. 32 Journal of the Royal Society of Western Australia, 80(1), March 1997 Hick PT 1981 Window mount for aerial photography. CSIRO Industrial Research News 144: 4pp. Hick PT 1987 Remote sensing of agricultural salinity. MSc Thesis. Curtin University of Technology, Perth. Hick PT 1994 Relative specifications of the Daedalus 1268 air¬ borne scanner and the SpecTerra Systems DMSV for ranger uranium. Position paper. 5 pp. CSIRO Minesite Rehabilita¬ tion Research Program. CSIRO Division of Water Research, Perth. Hick PT 1996 Spectral measurement of illumination sources at Thevenard Island:- a preliminary study of the probable effects of gas flares and oil production facility lights on Green turtles:- a subsequent revisit to measure a range of gas-flow rates. 18 pp. CSIRO Division of Soils, Perth. Hick PT 1996 High-resolution remote sensing for long-term monitoring of environmental effects of the operations of Robe River Iron Associates in the Pilbara, Western Australia. Report. 19 pp. CSIRO Minesite Rehabilitation Program, Perth. Hick PT 1996 Measuring the effects of the construction of Dampier Salt's bitterns channel on proximal mangroves using remotely-sensed data. Report. 22 pp. CSIRO Minesite Reha¬ bilitation Program, Perth. Hick PT, Davies JE & Russell WGR 1987 A portable field spectroradiometer for studies of dryland salinity. Proceed¬ ings of the 41800 t Au production and reserves). This giant gold deposit owes its size and unique position to a number of features of favourable host-rock composi¬ tion, thickness and mechanical properties, favourable geom¬ etry of its host units and hosting greenstone belts at favourable PT conditions in the greenschist facies that together with proximity to the regional scale Boulder- Lefroy Fault allowed highly focused fluid flow on both a deposit and district scale. Rajesh HM, Santosh M & Yoshida M 1996 The felsic magmatic province in East Gondwana - implications for pan-African tectonics [Review]. Journal of Southeast Asian Earth Sciences 14:275-291. Reason CJC 1996 Topography and the dynamical response to easterly flow in southern hemisphere subtropical west coast regions. Meteorology & Atmospheric Physics 61:187-199. Robertson 1DM 1996 Ferruginous lag geochemistry on the Yilgam Craton of Western Australia - practical aspects and limitations. Journal of Geochemical Exploration 57:139-151. Robinson BH, Chiarucci A, Brooks RR, Petit D, Kirkman JH, Gregg PEH & Dedominicis V 1997 The nickel hyperaccumulator plant Alyssum bertolonii as a potential agent for phytoremediation and phytomining of nickel. Journal of Geochemical Exploration 59:75-86. Seaman RS 1997 A comparison of some methods for reduction of pressure to sea level over Australia. Austra¬ lian Meteorological Magazine 46:15-25. Seddon G 1996 Thinking like a geologist: the culture of geology. Mawson Lecture 1996. Australian Journal of Earth Sciences 43:487-495. Geology is described by G Seddon of the Centre for studies in Australian Literature (the University of West¬ ern Australia) as having a crucial role in both the scientific and popular culture. His conclusions regarding the na¬ ture of geology include: (i) there is no hierarchy of the sciences; (ii) geology is a Romantic science rather than a Classical one; (iii) there is no such thing as the scientific method; (iv) geologists often attempt to reconcile con¬ flicting hypotheses; (v) geological phenomena are often of an almost irreducible complexity and their investigation is beset by problems of scale, both spatial and temporal; and (vi) the concept of universality has a distinctive appli¬ cation in geology. Among the non-professional uses of geology are: (i) human history is incomplete without environmental history; (ii) geology has application in en¬ vironmental planning and management; and (iii) awareness of the geology of a region enhances the sense of place. Seitz HM & Keays RR 1997 Platinum group element segre¬ gation and mineralization in a noritic ring complex formed from proterozoic siliceous high magnesium basalt magmas in the Vestfold Hills, Antarctica. Journal of Petrol¬ ogy 38:703-725. Semeniuk V 1996 Coastal forms and Quaternary processes along the arid Pilbara coast of northwestern Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 123: 49-84. Coastal landforms along the arid Pilbara coast formed during the Quaternary through the influence of ancestral landforms and fluvial and shoreline accretion, coastal erosion, and cementation. Climate also played a signifi¬ cant role with high evaporation rates coupled with limited rainfall, cyclonic storms and limited sediment delivery to the coastal zone. As a result, this arid coast is characterised by a range of features such as construction of arid-zone deltas, delta destruction and sediment redis¬ tribution during times of sediment depletion, cyclone- induced erosion and sedimentation, mangrove deposits, formation of salt flats, and precipitation and cementation to form beachrocks, high tidal crusts and gypsum pre¬ cipitates. Simonson BM & Hassler SW 1997 Revised correlations in the Early Precambrian Hamersley Basin based on a horizon of resedimented impact spherules. Australian Journal of Earth Sciences 44:37-48. Si verson M 1996 Lamniform sharks of the mid Cretaceous Alinga Formation and Beedagong claystone. Western Australia. Palaeontology 39:813-849. Smith RE 1996 Regolith research in support of mineral exploration in Australia. Journal of Geochemical Explora¬ tion 57:159-173. Sugitani K, Horiuchi Y, Adachi M & Sugisaki R 1996 Anomalously low Aty3/Ti02 values for Archean cherts from the Pilbara Block, Western Australia - possible 42 Journal of the Royal Society of Western Australia, 80(2), October 1997 evidence for extensive chemical weathering on the early earth. Precambrian Research 80:49-76. Sundaralingam K 1997 Shear velocity structure beneath the Western Australian region. Australian Journal of Earth Sciences 44:69-75. Sylvester P, Campbell IH & Bowyer DA 1997 Niobium/ uranium evidence for early formation of the continental crust. Science 275:521-523. Tompkins LA, Groves DI, Windrim DP, Jablonski W & Griffin WL 1997 Petrology, mineral chemistry, and exploration significance of Fe-sulfides from the metal dispersion halo surrounding the Cadjebut Zn-Pb mvt deposit. Western Australia. Applied Geochemistry 12:37. Vanemden B, Thomber MR, Graham J & Lincoln FJ 1997 The incorporation of actinides in monazite and xenotime from Placer deposits in Western Australia. Canadian Miner¬ alogist 35:95-104. Vincent P 1996 Rillenkarren in the British Isles. Zeitschrift fur Geomorphologie 40:487-497. Williams GE 1997 Precambrian length of day and the validity of tidal rhythmite paleotidal values. Geophysical Research Letters 24:421-424. Wilson TJ, Grunow AM & Hanson RE 1997 Gondwana assembly - the view from southern Africa and east Gondwana. Journal of Geodynamics 23:263-286. Witt WK, Swager CP & Nelson DR 1996 Ar-40/Ar-39 and U-Pb age constraints on the timing of gold mineralization in the Kalgoorlie gold field, Western Australia - a discussion. Economic Geology & the Bulletin of the Society of Eco¬ nomic Geologists 91:792-795. Woodhead JD & Hergt JM 1997 Application of the double spike technique to Pb-isotope geochronology. Chemical Geology 138:311-321. Yeats CJ, McNaughton NJ & Groves DI 1996 Shrimp U-Pb geochronological constraints on Archean volcanic-hosted massive sulfide and lode gold mineralization at Mount Gibson, Yilgarn Craton, Western Australia. Economic Geology & the Bulletin of the Society of Economic Geolo¬ gists 91:1354-1371. Zhou HW 1996 A high-resolution p wave model for the top 1200 km of the mantle. Journal of Geophysical Research- Solid Earth 101(B12):27791-27810. Life Sciences Abensperg-Traun M, Arnold GW, Steven DE, Smith GT, Atkins L, Viveen JJ & Gutter M 1996 Biodiversity indicators in semi-arid, agricultural Western Australia. Pacific Conservation Biology 2:375-389. A cooperative study by researchers from CSIRO Divi¬ sion of Wildlife Ecology (Midland) and the Agricultural University, Wagingen (Netherlands), investigated biodiversity indicators in semi-arid, agricultural Western Australia. Remnant area, vegeta tional structural diversity, species richness of plants, lizards and terrestrial arthropods, and the relative abundance of arthropod species were examined as indicators of faunal richness in two contrasting vegetation types, gimlet woodland and shrublands. No indicator variables effectively predicted total faunal richness for either vegetation type, but veg¬ etation structural diversity and plant richness explained a high percentage of variation in the richness of lizards, scorpions, termites and beetles in woodlands. The rich¬ ness of the shrubland fauna was poorly predicted by all indicator variables. Differences in the predictive values of vegetation structure and plant richness between the two vegetation types was partly due to the spatial heteroge¬ neity of biotic richness, and perhaps the scale of structure measurements. Abensperg-Traun M, Steven D & Atkins L 1996 The influence of plant diversity on the resilience of harvester termites to fire. Pacific Conservation Biology 2:279-285. Researchers from CSIRO Division of Wildlife Ecology (Midland) describe how the harvester termites of floristi- cally-rich mallee-heath of southern Western Australia appear to be resilient to high-intensity fire, in contrast with harvest¬ ers in floristically-simple, intensely-burnt spinifex grass¬ land of tropical Western Australia. It appears that high floristic diversity enhances the resilience of harvester ter¬ mites to fire. Although the death of spinifex and associ¬ ated harvester termites after fire may be atypical, the temporary local extinction of harvester termites might not be exceptional, particularly if the fire occurs with drought or high grazing pressure. Agboma PC, Jones MGK, Peltonensainio P, Rita H & Pehu E 1997 Exogenous glycinebetaine enhances grain yield of maize, sorghum and wheat grown under two supplemen¬ tary watering regimes. Journal of Agronomy & Crop Science-Zeitschrift fur Acker und Pflanzenbau 178:29-37. Akilan K, Marshall JK, Morgan AL, Farrell RCC & Bell DT 1997 Restoration of catchment water balance - responses of clonal river red gum (Eucalyptus camaldulensis) to water¬ logging. Restoration Ecology 5:101-108. Amaoka K, Arai M & Gomon MF. 1997 A new species of Arnoglossus (Pleuronectiformes, Bothidae) from the south¬ western coast of Australia. Ichthyological Research 44:131-136. Anderson PK. 1997 Shark Bay dugongs in summer. 1. Lek mating. Behaviour 134:433-462. Bailey MC & Hamilton DP 1997 Wind induced sediment resuspension - a lake-wide model. Ecological Modelling 99:217-228. Bray RA & Cribb TH 1997 Lepocreadiid (Digenea) species from members of the marine teleost family Cheilodactylidae from south-western Australia, including four new genera and five new species. Systematic Parasi¬ tology 37:27-45. Carnegie AJ, Keane PJ & Podger FD 1997 The impact of three species of Mycosplmerella newly recorded on eucalyp¬ tus in western australia. Australasian Plant Pathology 26:71-77. Chapman A & Lane JAK 1997 Waterfowl usage of wet¬ lands in the south-east arid interior of Western Australia 1992-93. Emu 97:51-59. Cheeseman JM, Herendeen LB, Cheeseman AT & Clough BF 1997 Photosynthesis and photoprotection in mangroves under field conditions. Plant Cell & Environment 20:579- 588. Dawson TJ & Ellis BA 1996 Diets of mammalian herbivores in Australian arid, hilly shrublands: seasonal effects on overlap between euros (hill kangaroos), sheep and feral goats, and on dietary niche breadths and electivities. Journal of Arid Environments 34:491-506. 43 Journal of the Royal Society of Western Australia, 80(2), October 1997 Two researchers from the University of new South Wales describe the results of a 12 year study on the diets of euros ( Macropus robustus), domestic sheep ( Ovis aries) and feral goats ( Capra hircus) in hilly shrubland of south¬ ern Australia. The diet of euros was based around grasses whereas grasses were important in the diet of sheep in wetter conditions but shrubs were important in the dry season. Feral goats had a broad diet, but a high prefer¬ ence for browse. Dietary niche breadths and electivities indicate only limited competition between the euros, sheep and goats. Eldridge MDB & Pearson DJ. 1997 Chromosomal rearrange¬ ments in rock wallabies, Petrogale (Marsupialia, Macropodidae).9. Further g-banding studies of the Petrogale lateralis complex - P. lateralis pearsoni, the west Kimberley race, and a population heterozygous for a cen¬ tric fusion. Genome 40:84-90. Field LH & Bailey WJ 1997 Sound production in primitive Orthoptera from Western Australia: sounds used in defence and social communication in Ametrus sp. and Hadrogryllus sp. (Gryllacrididae: Orthoptera). Journal of Natural History 31:1127-1141. Researchers from the University of Canterbury (New Zealand) and the University of Western Australia have collaborated to describe sound production in two undescribed Western Australian orthopteran insects (Gryllacrididae). Sound is used for both defence, pro¬ duced by femoro-tergal stridulation, and intra-specific signalling by drumming on the substrate. The evolution of these calling behaviours is discussed with reference to the other primitive ensiferan family (Stenopelmayidae) known to produce both tergo-stemal defensive stridula¬ tion and femoral drumming. Kang HJ & Fenical W 1997 Aplidiamine, a unique zwitteri- onic benzyl hydroxyadenine from the Western Australian marine ascidian Aplidiopsis sp. Tetrahedron Letters 38:941-944. Kang HJ & Fenical W 1997 Ningalins a-d - novel aromatic alkaloids from a Western Australian ascidian of the genus Didemnum. Journal of Organic Chemistry 62:3254-3262. Loi A, Cocks PS, Howieson JG & Carr SJ 1997 Morphological characterization of mediterranean populations of Biserrula pelecinus L. Plant Breeding 116:171-176. Newell GR 1997 The abundance of ground-dwelling inver¬ tebrates in a Victorian forest affected by dieback ( Phytophthora cinnamomi) disease. Australian Journal of Ecology 22:206-217. Parsons KE 1997 Role of dispersal ability in the phenotypic differentiation and plasticity of two marine gastropods. 1. shape. Oecologia 110:461-471. Postmaster A, Sivasithamparam K & Turner DW 1997 Enumeration and identity of microorganisms isolated from the surface of banana fruits at three developmental stages. Scientia Horticulturae 69:189-197. Rahman MH & Hossain I Moslehuddin 1997 Nutritional evaluation of sweet lupin ( Lupinus angustifolius) - net protein utilization (npu), nitrogen balance and fraction¬ ation studies. British Journal of Nutrition 77:443-457. Regan KL, Siddique KHM, Tennant D & Abrecht DG 1997 Grain yield and water use efficiency of early maturing wheat in low rainfall mediterranean environments. Aus¬ tralian Journal of Agricultural Research 48:595-603. Robinson BH, Chiarucci A, Brooks RR, Petit D, Kirkman JH, Gregg PEH & Dedominicis V 1997 The nickel hyperaccumulator plant Alyssum bertolonii as a potential agent for phytoremediation and phytomining of nickel. Journal of Geochemical Exploration 59:75-86. Siverson M 1996 Lamniform sharks of the mid Cretaceous Alinga Formation and Beedagong claystone. Western Australia. Palaeontology 39:813-849. Smolker R, Richards A, Connor R, Mann J & Berggren P 1997 Sponge carrying by dolphins (Delphinidae, Tursiops sp.) - a foraging specialization involving tool use. Ethology 103:454-465. Speers DJ & Jelfs J 1997 Typing of Neisseria meningitidis by restriction analysis of the amplified pora gene. Pathology 29:201-205. Steinbomer ST, Wabnitz PA, Waugh RJ, Bowie JH, Gao CW, Tyler MJ& Wallace JC 1996 The structures of new peptides from the Australian red tree frog Litoria rubella - the skin peptide profile as a probe for the study of evolutionary trends of amphibians. Australian Journal of Chemistry 49:955-963. The Zoological Catalogue of Australia. Volume 28. Strepsiptera (TR New), Mecoptera (KJ Lambkin), Neuroptera (TR New), Siphonaptera (AA Calder). The Australian Biological Resources Study, Canberra. CSIRO Publishing, Melbourne. Volume 28 of the outstanding monograph series by the Australian Biological Resources Study, The Zoological Catalogue of Australia, presents informative family intro¬ ductions with generic and species synonymies, taxonomic information, status, distribution and ecological information, and bibliographies. The insect groups included in Volume 28 are Strepsiptera (42 species with a unique parasitoid existence), Mecoptera (30 endemic species that are im¬ portant predators in natural systems and of biogeo¬ graphic significance), Neuroptera (over 600 species that are important as biological control agents, with high endemism and fascinating archaic lineages), and Sipho¬ naptera (80 valid species-group names with an impact on human health and disease, with high endemism and also exotics). Thompson GG & Withers PC 1997 Comparative morphology of Western Australian varanid lizards (Squamata: Varanidae). Physiological Zoology 233:127-152. This cooperative study by researchers from Edith Cowan University and the University of Western Australia considers intra-specific and interspecific changes in body shape with size, for 17 species of Western Australian goannas. Although goannas are generally considered to be conservative in shape, many body morphometric measurements were non-isometric, indicating significant variation in shape. Canonical variate analysis clearly differentiated the two subgenera of goannas (Varanus and Odatria ) and species were generally sexually dimorphic. The morphological variation among the 17 goanna species was associated with foraging mode and ecology. Thompson GG & Withers PC 1997 Standard and maximal metabolic rates of goannas (Squamata: Varanidae). Physi¬ ological Zoology 70:307-323. Measurement of the standard and maximal metabolic rates of various Western Australian goannas, ranging in body mass from 10 to 3750 grams, by researchers from Edith Cowan University and the University of Western Australia indicated an unusual allometric effect of mass. 44 Journal of the Royal Society of Western Australia, 80(2), October 1997 Standard metabolic rate was proportional to mass0-90 10 1 with considerable variation between species, whereas maximal metabolic rate scales with mass1179. Conse¬ quently, the factorial metabolic scope is much greater for small goannas compared with large goannas. Inter-specific variation in metabolic physiology had some ecological correlates. Widely-foraging Varanus tristis and V. eremias had a high standard metabolic rate. Arboreal V. caudolineatus, V. gilleni and V. tristis, had a high maximal metabolic rate. Tong SM 1997 Heterotrophic flagellates from the water column in Shark Bay, Western Australia. Marine Biology 128:517- 536. The diversity of heterotrophic flagellates in the water column at Shark Bay was examined by a researcher from the University of Sydney, and found to include 41 species of apusomonads, cercomonads, choanoflagellates, cryptomonads, euglenids, heteroloboseids, stramenopiles, and others of uncertain taxonomy. Three quarters of the species had not been previously reported from southern subtropical regions. It appears that many heterotrophic flagellates have a cosmopolitan distribution, and the bio¬ geography of these Shark Bay species is discussed with regard to studies in other localities. Wallace CC 1997 New species and new records of recently described species of the coral genus Acropora (Scleractinia, Astrocoeniina, Acroporidae) from Indonesia. Zoological Journal of the Linnean Society 120:27-50. Williams ST & Benzie JAH 1997 Indo-west Pacific patterns of genetic differentiation in the high-dispersal starfish Linckia laevigata. Molecular Ecology 6:559-573. Physical Sciences Barton A 1997 States of Matter, States of Mind. IOP Publishing, USA. This easy-to-read book introduces the structure and property of matter, how the physical world is put together and stays together. It explains some of the intricate details and some of the grand schemes of life, the universe, and everything, by making analogies with everyday experi¬ ences. Ranging from fundamental ideas and fundamental particles to the makeup of the universe, the contents of this book are understandable to readers with an inquiring mind but no scientific background, [available in Australia from DA Books: service@dadirect.com.au] Clare BW & Supuran CT 1997 Carbonic anhydrase inhibitors. Part 41. Quantitative structure-activity correlations involv¬ ing kinetic rate constants of 20 sulfonamide inhibitors from a non-generic series. European Journal of Medical Chemistry 32:311-319. This collaborative research between Murdoch University and Universita Delgi Studi di Firenze (Italy) presents a quantitative structure-activity relationship for 20 sulfona¬ mide inhibitors of carbonic anhydrase. These drugs, which are not of a classical congeneric series as the only common factor is the sulfonamide group, have as important local factors the Millikan charge on atoms of the sulfona¬ mide group, and as global factors the size and shape of the molecule, calculated frontier orbital energies, and lipophilicity. The equilibrium constant and kinetic associa¬ tion rate were well correlated, but the kinetic dissociation rate constant was not. Hefter G & Marcus Y 1997 A critical review of methods for obtaining ionic volumes in solution. Journal of Solution Chemistry 26:249-. These researchers critically review the various methods for obtaining individual, or " absolute'7, ionic standard partial molar volumes from whole elctrolyte data in aque¬ ous and non-aqueous solutions. After revealing a number of undetected errors in previous analyses, it is shown that the reported agreement amongst the various methods in aqueous solution is largely fortuitous. Although all methods are unsatisfactory to varying degrees, the reference elec¬ trolyte approach, using an electrolyte such as tetraphenylarsonium, is the least objectionable. The research¬ ers recommend that, at present, a difference between the standard partial molal volumes for Ph,P* and BPh4‘ of 2 cm3 mol'1 be used in all solvents at 25 °C. Kang HJ & Fenical W 1997 Aplidiamine, a unique zwitteri- onic benzyl hydroxyadenine from the Western Australian marine ascidian Aplidiopsis sp. Tetrahedron Letters 38:941-944. Kang HJ & Fenical W 1997 Ningalins a-d - novel aromatic alkaloids from a Western Australian ascidian of the genus Didemnum. Journal of Organic Chemistry 62:3254-3262. Obsil M, Majer V, Grolier J-PE & Hefter G 1996 Volumetric properties of, and ion-pairing in, aqueous solutions of alkali-metal sulfates under superambient conditions. Journal of the Chemical Society, Faraday Transactions 92:4445-4451. Apparent molar volumes for Na2S04(flf/) and K2SOA{acj) have been obtained by collaborative densitometric re¬ search between Murdoch University and Universite Blaise Pascal chemists, and from selected experimental data from the literature, at temperatures and pressures up to 573 K and 30 Mpa. Using the Pitzer ion-interaction model to correlate the data, the researchers obtained recom¬ mended values as a function of temperature and partial molar volumes at infinite dilution. Ion pairing was found to be significant for Na2S04(aq) and K2SO 4(acj) at higher temperatures. Ralph DE & Stevenson JM 1995 The role of bacteria in well clogging. Water Research 29:365-369. These researchers from Murdoch University studied the effect of ochreous sludge from blocked irrigation bores on the oxidation rate of soluble Fe(II). In comparison with sterile flasks of Fe(Il) solution which showed the expected first order decline in Fe(II) concentration, flasks innoculated with ochreous sludge had significantly enhanced rates of Fe(II) oxidation, with the greatest increase at pH 8.5. Robinson BH, Chiarucci A, Brooks RR, Petit D, Kirkman JH, Gregg PEH & Dedominicis V 1997 The nickel hyperaccumulator plant Alyssum bertolonii as a potential agent for phytoremediation and phytomining of nickel. Journal of Geochemical Exploration 59:75-86. Steinbomer ST, Wabnitz PA, Waugh RJ, Bowie JH, Gao CW, Tyler MJ& Wallace JC 1996 The structures of new peptides from the Australian red tree frog Litoria rubella - the skin peptide profile as a probe for the study of evolutionary trends of amphibians. Australian Journal of Chemistry 49:955-963. Stelbovics AT & Berge L 1997 Nonuniqueness in the close¬ coupling method for e-He scattering. Physical Review A 55:1028-. 45 Journal of the Royal Society of Western Australia, 80(2), October 1997 These researchers from Murdoch University present a general analysis of the close-coupling equations for e-He scattering and show why close-coupling expansion gives rise to nonunique solutions and are construct equations with a unique solution. The analyses concentrate on two models of target states, a frozen-core model with one electron restricted to the Is He* orbital, and a model using full configuration-interaction target states. Thurgate SM 1996 Auger photoelectron coincidence experi¬ ments from solids. Journal of Electron Spectroscopy and Related Phenomena 81:1-31. [review] Webb J, St Pierre T G, Tran KC, Chua-anusom W, Macey DJ & Pootrakul P 1996 Biologically significant iron (III) oxyhydroxy polymers: Mossbauer spectroscopic study of ferritin and hemosiderin in pancreas tissue of b-thalas- semia /hemoglobin E disease. Inorganica Chimica Acta 243:121-125. Researchers from Murdoch University and Mahidol University (Thailand) used Mossbauer spectra of ferritin to show that it contained iron cores based on the ferrihydrite structure consistent with previous electron diffration data. Cores in the crude hemosiderin are gener¬ ally of this type, but some have a defective goethite struc¬ ture. Note from the Hon Editor: This column helps to link the various disciplines and inform others of the broad spectrum of achievements of WA scientists (or others writing about WA). References are abstracted from Current Contents by searching for Western Australia in the title and abstract. Other contributions to "Recent Advances in Science in Western Australia" are welcome, and may include pa¬ pers that have caught your attention or that you believe may interest other scientists in Western Australia and abroad. They are usually papers in refereed journals, or books, chapters and reviews. Abstracts from conference proceedings will not be accepted. Please submit either a reprint of the paper, or a short (2-3 sentences) summary of a recent paper together with a copy of the authors' names and addresses, to the Hon Editor or a member of the Publications Committee: Dr P C Withers (Hon Editor), Department of Zoology, University of Western Australia, Nedlands WA 6907 Dr S D Hopper (Life Sciences), Kings Park and Botanic Garden, West Perth WA 6004 Dr A E Cockbain (Earth Sciences), PO Box 8114, Angelo Street, South Perth WA 6151 Assoc Prof G Hefter (Physical Sciences), Mathemati¬ cal and Physical Sciences, Murdoch University, Murdoch WA 6150 Final choice of articles is at the discretion of the Hon¬ orary Editor. "Letters to the Editor" concerning scientific issues of relevance to this journal are also published, at the discre¬ tion of the Hon Editor. Please submit a word processing disk with letters, and suggest potential reviewers or respondents to your letter. P C Withers, Honorary Editor , Journal of the Royal Society of Western Australia. c/0 Western Australian Museum, Francis Street, Perth WA 6000 46 Journal of the Royal Society of Western Australia, 80:47-54, 1997 Diet of herbivorous marsupials in a Eucalyptus marginata forest and their impact on the understorey vegetation K A Shepherd1, G W Wardell-Johnson23, W A Loneragan1 & D T Bell1 ’Department of Botany, The University of Western Australia, Nedlands WA 6907: email wal@cyllene.uwa.edu.au 2Manjimup Research Centre, Department of Conservation and Land Management, Brain Street, Manjimup WA 6258 3Current address: Botany Department, University of Namibia, Windhoek, Namibia: email gwardell@unam.na Received April 1996; accepted February 1997 Abstract Wire exclosures were established to exclude herbivores in remnant jarrah (Eucalyptus marginata) forest vegetation within the Perup Nature Reserve, south-western Australia, and were assessed after a 10 year period. Significantly higher plant cover values were recorded for a number of species protected from herbivory inside the exclosures compared to outside. Sub-shrubs and vines showed the greatest increase in vegetation cover within exclosures. Bossiaea ornata, Billardiera vari [folia, Opercularia hispidula, Logania serpyllifolia and Tetrarrhena laevis were most favoured by herbivore exclusion in terms of relative abundance. Faecal analysis confirmed that plant species having the greatest decrease in cover outside wire exclosures were consumed by one or more of the five predominant marsupial herbivores of the Perup forest. Faecal material collected over three sampling periods during 1992 revealed a total of 42 different plant species. The largest of the herbivores, the western grey kangaroo ( Macropus fuliginosus), consumed the greatest diversity of plants (32 species in total), while the black-gloved wallaby (Macropus irtna) and the tammar wallaby ( Macropus eu genii) grazed 21 and 25 different species, respectively. The common brush-tail possum (Trichosurus vulpecula) consumed leaves from the two dominant trees of the region Eucalyptus marginata and E. calophylla , and four understorey species including Eqjtomena cunninghamii and Hakea lissocatyriw. Faecal samples of the western ring¬ tail possum ( Pseudocheirus occidental is) contained only forest canopy species. This study has implications for appropriate flora and fauna management of nature reserves. Possible competition for plant resources was indicated by diet overlap. However, due to the polyphagous nature of these particular herbivores and an ability to shift resource preferences, competitive limitations of particular food resource species in the Perup Nature Reserve are un¬ likely. As herbivores have been shown to reduce plant cover in the area and, therefore, the rate of build up of fire fuels, their population management and potential impact in the Perup forest are important considerations for fire management plans in the forest. Introduction The Perup Nature Reserve (40000 ha; 34° 16" S, 116° 36' E) is located 45 km east of Manjimup in the south¬ west of Western Australia (Fig 1). The Perup forest is partially contiguous with other state forest lands, but is primarily surrounded by cleared farmlands. The Perup Nature Reserve is designated as a fauna management priority area (MPA) due to the documented presence of 21 species of native mammals, including at least five gazetted as rare or endangered (Christensen et al 1985; Wardell-Johnson & Nichols 1991). A knowledge of the range of plant life forms, leaf characteristics and species consumed by a diverse com¬ munity of herbivores was considered desirable for greater understanding of individual dietary require¬ ments, herbivore interactions and potential impact on vegetation. This would provide a basis for management © Royal Society of Western Australia 1997 decisions on fuel reduction burns, habitat management fires and predator control. This study had four aims. Firstly, to determine floris- tic differences between herbivore exclosures and adjacent open sites in the Perup Nature Reserve. Secondly, to document food resources utilised by the five major native herbivorous marsupials found within the area and deter¬ mine possible factors influencing diet choice. Thirdly, to compare the findings from this study with other dietary studies of marsupial herbivores in reserves around the state and, finally, to assess possible management impli¬ cations for the Perup Nature Reserve. Materials and Methods Study site The topographic relief of the Perup Nature Reserve includes low undulating ridges separated by broad, flat valleys with seasonal swamps and streams in the lower 47 Journal of the Royal Society of Western Australia, 80(2), October 1997 Figure 1. Location of the study site in the Perup Nature Reserve. regions (Christensen 1980). The vegetation of the area is predominantly open jarrah (Eucalyptus marginata) along the ridges on lateritic soils, while marri (Eucalyptus calophylla) occurs more commonly in the valleys on the sandier soils. The understorey strata of the ridges and slopes are dominated by a diverse mixture of low (< 1 m) xeric shrubs with the common species being Hakea lissocarpha, Leucopogon capitellalus and Bossiaea ornata. Tall thickets (> lm) form in some areas following a fire. These thickets are generally dominated by single species, either Gastrolobium bilobum, G. spinosum, Melaleuca viminea or Acacia pukhella (Christensen 1977; Buchanan & Wardell- Johnson 1990). The Perup Nature Reserve was burnt frequently from the late 1930's to the mid 1960's. Cyclic fuel-reduction burning that is dependent on litter build-ups was initi¬ ated in the 1960's under the control of the Forest Depart¬ ment and subsequently the Department of Conservation and Land Management. More recently, it was recognised that a combination of intense autumn fires lit under dry soil conditions and milder spring fires was required to maintain the monospecific thickets and to stimulate the regrowth of the legume plant species (Christensen & Maisey 1987). In the presence of the European fox (Vulpes vulpes), these G. bilobum thickets are essential as refugia for the tammar population within the Perup forest (Christensen 1980). Marsupial herbivore species This study assessed the plant species consumed by five marsupial herbivores; the western grey kangaroo (Macropus fuliginosus), the black-gloved wallaby (Macropus irma), the tammar wallaby (Macropus eugenii), the common brush-tail possum (Trichosurus vulpecula) and the western ring-tail possum (Pseudocheirus occidentalis). The western grey kangaroo (30-50 kg) has a wide distribution, extending from Western Australia across the continent to Victoria and central New South Wales (Poole 1983). The black-gloved wallaby is found only in the south-west of Western Australia (Christensen 1983). Although this large wallaby (7-9 kg) now occurs only in isolated populations, it is found in reasonable numbers within its natural range. The tammar wallaby, a smaller marsupial herbivore (5. 5-7.5 kg), was once dis¬ tributed over a diverse array of habitats prior to European settlement. This species is now gazetted as threatened and is restricted to a number of small isolated popula¬ tions on coastal islands (Bell et al. 1987; Poole el al. 1991) and on the mainland of the south-west of Western Aus¬ tralia. The largest population on the mainland occurs in the Perup forest near dense Melaleuca viminea and Gastrolobium bilobum thickets which provide protection from predators such as foxes and feral cats (Christensen 1992). Populations of the common brush-tail possum occur in the forested regions of Western Australia and its dis¬ tribution extends to eastern Australia (How 1983). The range of the western ring-tail possum is restricted to a few remnant jarrah forest and near-coastal Agonis flexuosa woodlands of the south-west of Western Australia (Jones et al. 1994a). Exclosure methods In autumn of 1981 following a prescribed burn, six wire-mesh exclosures were constructed along the eastern side of a 100 m section of the Glendale 2 Road, 1.5 km south of De Landgraft Road in the Perup Nature Reserve (Inions et al. 1989). The exclosures were positioned along the upper to mid-slope of a broad valley within open jarrah /marri forest. The exclosures were approximately 1.5 m in height and 100 m2 in area. They exclude all ground-dwelling mammals and it is unlikely that the arboreal possums gained access to the exclosed areas by traversing across the canopies of adjacent trees due to low canopy continuity at Perup. The possums are more likely to descend to ground level and walk to adjacent trees (Jones et al. 1994a). The area east of the Glendale 2 Road was subsequently burnt in 1984/1985 and the area to the west of the road was burnt in 1985/1986. In the winter of 1991, species cover within the six exclosures was determined along five, 10m transects spaced lm apart using the point interception method (Mueller- Dombois & Ellenberg 1974). Pins were placed at 20 cm intervals along the transects (250 points per plot). Using the same methods, plant species cover values were also obtained for 19 open areas surrounding the exclosures. Mean cover values for the plant species inside and out¬ side the wire exclosures were compared after arcsin transformation using a two-tailed Student's t-test. Leaf epidermal material A reference collection of prepared epidermal tissue slides of 67 species from the study site facilitated the microscopic identification of plant fragments recovered from faecal material. Epidermal vouchers were prepared using a modification of the acid digestion methods outlined 48 Journal of the Royal Society of Western Australia, 80(2), October 1997 by Storr (1961) and Halford et al. (1984a). Leaf blades were covered in a 50% glacial acetic acid solution and heated in a water bath at 80 °C for 24-72 h. The fragments were then rinsed in water and any remaining fibrous tissue was removed. The cleared leaf fragments were then dehydrated through a series of ethanol solutions (increasing in concentration to 95%) and stained by immersion in 0.5% gentian violet in 95% alcohol for 48 h. The fragments were washed in 95% alcohol and mounted on slides using Eukit® permamount. Diagnostic line drawings of the epidermal tissue patterns, especially the guard cell complexes, were produced for each species as a further aid to identification. Faecal material Faecal pellets for each of the five common marsupial species in the area are distinctive in shape and size and can be easily separated by sight (P Christensen & C Wheeler, pers comm). Pellets were collected from a mid¬ slope area approximately 150 m by 500 m in the jarrah/ marri open forest surrounding the exclosures. This area extended from the vicinity of the exclosures, down a westward facing slope into a low-lying gully. In contrast to the upper and mid-slope areas, large trees were absent in the gully being replaced by a dense thicket of Gastrolobium bilobum approximately 2 m in height. Fresh faecal pellets were collected over three sampling periods from the autumn through spring period of 1992. The plant species consumed by the marsupial herbivores were determined by comparison of epidermal fragments in faecal material to the plant voucher specimens. How¬ ever, a more rigorous digestion technique was required for preparation of the faecal pellets than for the plant vouchers. Single, air-dried faecal pellets were frag¬ mented, placed in test tubes and covered with equal parts of 10% chromic acid and 10% nitric acid. The test tubes were heated in an acid digestion block at 80-100 GC within a fume hood, simmering the contents for 15- 20 min. After maceration the contents were cooled to room temperature, and filtered with several washes of dilute potassium hydroxide solution and distilled water. The filtrate was sieved through a 0.5 mm sieve, collected and stained with 0.5% gentian violet in 95% alcohol. The fragmented plant material was scanned using the dis¬ secting microscope at 12.5-50x magnification, mounted on slides with Eukit and compared with the voucher col¬ lection for species identification. As an index of dietary preference, the percentage occurrence of a plant species was determined by noting the proportion of the sampled pellets that included a particular plant species. Results A combined total of 73 plant taxa were recorded inside the exclosures and in the adjacent vegetation at the Perup Nature Reserve study site (Table 1). Nineteen species had significantly different and greater percentage cover values inside the wire exclosures. These included Bossiaea ornata, Billardiera variifolia, Opercularia hispuiula, Logania serpyllifolia, Telrarrhena laevis, Scaevola striata and Billardiera floribunda. Total plant cover outside the wire exclosures was only 58% of that recorded inside. In contrast, Macrozamia riedlei, Eucalyptus marginata and Hibbertia commutata had significantly greater cover values outside the exclosures. Of the species showing a significant lower cover outside exclosures, 42% were small sub-shrub species, semi-woody or fibrous herbaceous perennials less than 0.5 m in height, and 26% were perennial vines and twining plants (Table 2). Table 1 Plant cover values comparing wire exclosure treatments and adjacent open areas. * indicates significant difference by two- tailed t-test (<0.05) following arcsin transformation. Plant classes: C, cycad; M, monocotyledon; D, dicotyledon. Life form: Tree, plant over 2 m height; Shrub, woody plant 0.5 - 2.0 m in height; Sub-shrub, semi- woody or fibrous herbaceous perennial < 0.5 m; Vine; twining perennial; Herb, mesophytic herbaceous perennial or annual. Species Percentage Cover Inside Outside Plant Class Life Form Bossiaea ornata 24.6 4.0* D Shrub Billardiera variifolia 13.8 1.2* D Vine Leucopogon capitellatus 12.5 15.4 D Sub-shrub Opercularia hispidula 10.6 0.6* D Sub-shrub Logania serpyllifolia 9.8 0.5* D Shrub Tetrarrhena laevis 9.7 0.2* M Sub-shrub Hakea lissocarpha 8.0 11.5 D Shrub Xanthorrhoea gracilis 7.7 6.4 M Sub-shrub Drosera macrantha 6.9 4.0 D Herb Eucalyptus calophylla 6.6 5.8 D Tree Scaevola striata 6.6 1.1* D Sub-shrub Billardiera floribunda 6.2 0.2* D Vine Tricon/ne elatior 5.6 2.8 M Sub-shrub Clematis pubescens 5.1 0.9* D Vine Tetratheca affinis 5.0 0.3* D Sub-shrub Chamaescilla corymbosa 4.8 1.4* M Sub-shrub Lomaruira caespitosa 4.7 0.7* M Sub-shrub Leucopogon verticillatus 4.3 4.4 D Shrub Macrozamia riedlei 4.1 9.7* C Shrub Eucalyptus marginata 3.9 11.4* D Tree H i bbert ia a mplexica u l is 3.5 4.5 D Shrub Danthonia pilosa 3.4 1.8 M Herb Lagen iphora h uegel i i 3.2 1.3 D Herb Cassytha sp 3.0 * p o D Vine Gastrolobium bilobum 2.8 4.4 D Shrub Pterostylis nana 2.7 0.4* M Herb Leptomeria cu n n ingham ii 2.5 0.5* D Shrub Thysanotus patersonii 2.4 0.0* D Vine Grass A 2.1 1.3 M Herb Xanthosia atkinsoniana 2.0 1.1 D Sub-shrub Agrostocrinum scabrum 1.9 0.0* M Herb Lomandra preissii 1.8 3.0 M Sub-shrub Loxocarya fasciculata 1.8 1.2 M Sub-shrub Acacia pulchella 1.5 1.1 D Shrub Acianthus reniformis 1.5 3.1 D Herb Amphipogon amphipogonoides 1.5 0.0* M Herb Lomandra Integra 1.5 0.5 M Sub-shrub Xanthorrhoea preissii 1.2 0.3 M Shrub Caladenia repens 1.2 0.7 M Herb Comesperma confertum 1.2 0.0* D Sub-shrub Leucopogon propinquus 1.2 0.9 D Sub-shrub Lomandra sp #4 1.2 0.6 M Sub-shrub Stackhousia monogxjna 1.2 0.0* D Sub-shrub Boronia spat hu lata 0.6 0.9 D Shrub Pimelea rosea 0.6 0.4 D Shrub Xanthosia Candida 0.6 0.4 M Sub-shrub 49 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 1 (continued) Species Percentage Cover Inside Outside Plant Class Life Form Acacia saligna 0.0 0.1 D Shrub Astroloma ciliatum 0.0 0.1 D Shrub Craspedia uniflora 0.0 1.0 M Herb Danthonia setacea 0.0 0.3 M Herb Daviesia preissii 0.0 0.7 D Shrub Hibbertia cunninghamii 0.0 1.9 D Shrub Hibbcrtia montana 0.0 6.7* D Shrub Hibbertia racemosa 0.0 0.3 D Shrub Leucopogon australis 0.0 1.0 D Shrub Lomandra sericea 0.0 0.2 M Sub-shrub Lomandra sonderi 0.0 0.2 M Sub-shrub Lomandra sp #1 0.0 0.7 M Sub-shrub Lomandra sp #2 0.0 0.2 M Sub-shrub Loxocarya flexuosa 0.0 0.2 M Sub-shrub Melaleuca viminea 0.0 0.4 D Shrub Neurachtie alopecuroidea 0.0 0.2 M Herb Patersonia Occident alis 0.0 0.8 M Sub-shrub Pentapeltis peligera 0.0 0.1 D Herb Persoonia longifolia 0.0 0.4 D Tree Phyllanthus calycinus 0.0 0.2 D Shrub Stipa sp 0.0 0.3 M Herb Stylidium adnatum 0.0 0.1 D Herb Tetrariopsis octandra 0.0 0.2 M Sub-shrub Trachymene pilosa 0.0 0.6 D Herb Tremandra diffusa 0.0 0.4 D Sub-shrub Trymalium ledifolium 0.0 0.3 D Shrub Dicotyledon #1 0.0 0.3 D Shrub Table 2 Plant life forms within the Perup study area (n = 73) and the percentage of life forms of the species showing potential of being selected by herbivores by recording a significantly lower cover outside the exclosures (n = 19). All Species Reduced Value Species Sub-Shrub 36 42 Shrub 31 16 Vine 7 26 Herb 22 16 Tree 4 0 Forty-two plant species in total were identified from the macropod faecal pellets, including both monocotyle- donous and dicotyledonous species of varying life forms, together with 12 unidentified species (Table 3). Faeces of the largest of the herbivores, the western grey kangaroo, had 32 species in total, with a range of six to 13 different species found per faecal pellet (Fig 2). Analysis of the individual faecal pellets indicated that the sedge Lepidosperma tenue was an important dietary component for the western grey kangaroo, occurring in 89% of the pellets sampled (Table 3). Other species frequently con¬ sumed (occurring in < 67% of pellets sampled) included Danthonia setacea, an unknown monocotyledon (#1), Gastrolobium bilobum and a Loxocarya species. Sixteen species were selected relatively infrequently, occurring in only 11% of the pellets sampled. Table 3 Frequency of occurrence of plants recovered from herbivore faecal pellets collected in the Perup Nature Reserve. The plant species are ordered in approximate order of overall percentage. Western grey kangaroo Black- gloved wallaby Tammar wallaby Brush- tail possum Ring- tail possum Monocotyledon #1 78 86 50 Danthonia setacea 78 43 80 Gastrolobium bilobum 67 43 50 40 Eucalyptus marginata 14 17 60 100 Hakea lissocarpha 56 43 17 60 Loxocarya sp 67 14 50 Bossiaea ornata 33 86 Leucopogon capilellatus 56 29 33 Lepidosperma tenue 89 17 Acacia pulchella 11 43 50 Eucalyptus calophylla 11 67 20 10 Lep tomeria cu n n inghamii 17 80 Monocotyledon #2 11 43 33 Dicotyledon #6 14 67 Dicotyledon #1 44 29 Opercularia hispidula 11 29 33 Juncus pallidus 22 14 33 Cassytha sp 29 33 Lomandra sericea 11 14 33 Leucopogon vert i cilia tus 22 14 17 Tetraria octandra 50 Monocotyledon #3 14 33 Monocotyledon #4 11 33 Boronia spathulata 11 29 Daviesia preissii 33 Lepidosperma angustatum 33 Lomandra sonderii 33 Dicotyledon #5 33 Tetrarrhena laevis 33 Billardiera variifolia 11 20 Melaleuca viminea 14 14 Monocotyledon #5 11 17 Neurachne alopecuroidea 22 Cen ta u ri u m erythraea 17 Dicotyledon #7 14 Astroloma ciliatum 11 Kennedia carinata 11 Lomandra sp #3 11 Dicotyledon #2 11 Dicotyledon #3 11 Dicotyledon #4 11 Dicotyledon #8 11 Twenty-one plant species were identified from the black-gloved wallaby pellets (Table 3), including four to nine plant species per sample (Fig 2). Bossiaea ornata and the unknown monocotyledon (#1) were frequent con¬ stituents of the black-gloved wallaby diet, being found in 86% of the pellets sampled. Other species were much less frequently consumed by wallabies, each species occur¬ ring in less than 43% of the pellets sampled. 50 Journal of the Royal Society of Western Australia, 80(2), October 1997 c o 3 C T O u. Uh 100 90 80 70 60 50 40 30 20 10 0 Ringtail possum S Brushtail Possum □ Tammar wallaby (2 Blackglove wallaby B Western grey kangaroo 2 3 4 5 6 7 8 9 10 11 12 13 Number of plants per faecal pellet Figure 2. Frequency of plant species in sampled faecal pellets for the ring-tail possum, brush-tail possum, tammar wallaby, black-gloved wallaby and western grey kangaroo collected from the study site in the Perup Nature Reserve. The diet of the tammar wallaby included 24 plant spe¬ cies in total (Table 3) with as many as 8-12 species recov¬ ered from single faecal pellets (Fig 2). One tammar faecal pellet, however, contained only two species. Plants most frequently consumed by this marsupial included Danihonia setacea, Eucalyptus calophylla and the unknown dicotyledon #6 recovered from 67-80% of the pellets sampled. Two other plant species frequently consumed, occurring in 50% of the pellets sampled, included the un¬ known monocot (#1) and the shrub Gastrolobium bilobum. Of the plant species found in the faecal pellets, only 28% were consumed by all three macropods. The plant types ranged from the small restionaceous Loxocarya spp and the coarse, herbaceous sub-shrub Opercularia hispidula, to larger plants, such as / uncus pallidus. Species that have tough leaf spines as seen in Hakca lissocarpha and Leucopogon capitellatus, or stem spines as seen in Acacia pulchella, were also frequently grazed. The arboreal possums appeared to consume fewer species than the macropods; also, fewer plant species were recorded per pellet during any one feeding period (Fig 2). The brush-tail possum diet consisted of six species (Table 3) with between one and four species per pellet (Fig 2). The most frequently consumed species included the understorey shrub Leptomeria citnninghamii and the two dominant tree species. Eucalyptus marginata and £. calophylla. Other understorey plants found in faecal material included the tall shrubs, H. lissocarpha and G. bilobum, and the vine Billardiera variifolia. Analysis of faecal pellets of western ring-tail possum revealed mainly E. marginata leaves, although fragments of E. calophylla were also retrieved from a single faecal sample. In addition to leaf epidermal fragments, pieces of seeds, bark and small insect larvae were also recov¬ ered from sampled pellets. Discussion After a ten-year period of herbivore exclusion, there was a floristic difference between the vegetation within the exclosures and that of adjacent open sites in the Perup Nature Reserve. A number of species of various life forms had a significantly decreased vegetative cover when exposed to herbivory. Total plant cover was also much reduced outside the wire exclosures. Dietary studies assist in understanding how herbi¬ vores utilise and subsequently influence their environ¬ ment. Faecal analysis is an indirect but accurate means of determining the specific food resources consumed by herbivores (Scotcher 1979; Johnson & Pearson 1981; Holechek el al. 1982). The presence of cuticular material of plant species showing a decrease in vegetative cover in pellets collected from the five native herbivores within the Perup Nature Reserve, suggest that these herbivores have a significant impact on a broad range of species. The Perup Nature Reserve is an important faunal reserve and an understanding of the resources required by the resident herbivores is essential to ensure appropriate management practices. The consumption of such a diverse array of species by the predominant herbivores may minimise the problem of overgrazing of any particular species. The ability of animal populations to selectively control relative densities of plant populations has been observed (Brown & Stuth 1993), but the complete elimi¬ nation of a species or particular life forms from a natural ecosystem is unlikely. Murden & Risenhoover (1993), using additions of a known high-quality food supplement to assess the effects of forage quantity and quality on pat¬ terns of resource use, found that deer and goats contin¬ ued to feed on a range of native plant species despite the presence of material of enhanced nutritional value. Their data suggest that normal behavioural patterns in large, free-ranging herbivores promote a polyphagous resource base, and thus reduce the probability of excessive utilisation pressure on any particular forage species. Although the marsupial herbivores of the Perup Nature Reserve showed particular preferences, the diversity of plant re¬ sources consumed indicates an ability to shift to other species when one plant species becomes less available. Increased emphasis on the land management practices designed preferentially for maintenance of the herbivores 51 Journal of the Royal Society of Western Australia, 80(2), October 1997 in this Reserve, specifically maintaining stable popula¬ tion numbers, should not adversely affect the survival of particular plant species. A comparison of plant species consumed by different herbivores potentially utilising the same food resource may give some indication of the interactions between herbivores and possible influences of diet choice. The overlap of plant species consumed by the large macropods indicates that there may be competition for resources in the Perup Nature Reserve. However, the herbivores tended to consume plants that were common in the understorey and, therefore, these are unlikely to become limiting. The polyphagous nature of these herbi¬ vores also allows them to graze a number of plant species at any one time, further limiting competition for specific plant resources. The choice of plant species consumed may be influ¬ enced by a number of factors. Many researchers have tried to determine what specifically influences diet choice as it is unknown whether animals select items on the basis of nutritional content, avoidance of toxic com¬ pounds or physical deterrents, such as leaf and stem spines. The grazing of a diverse number of species as seen in the Perup marsupial herbivores may preclude problems of nutritional deficiency. Also, by consuming small amounts of plant material relatively infrequently, inherent multifunction oxidate detoxification systems may break down any novel toxins present (Dawson 1989). The herbivores of the Perup Nature Reserve do not appear to be deterred by the physiological and mor¬ phological defence characteristics that have evolved in Australian plants. These herbivores consumed species that have sharp stem spines ( Acacia pulchella) and leaf spines ( Hakea lissocarpha) or high concentrations of toxic compounds ( Gastrolobium bilobum and Eucalyptus sp) (Freeland & Janzen 1974; Hume et al. 1984; Owen-Smith & Cooper 1987; Robbins et al. 1987). Gastrolobium bilobum contains up to 2600 mg kg'1 of the toxic compound sodium fluoroacetate (King et al. 1981; Twigg & King 1991). The western grey kangaroo and the tammar are fluoroacetate- tolerant (Twigg & King 1991) and this study suggests that the black-gloved wallaby also has a high tolerance as this herbivore was found to commonly consume Gastrolobium bilobum. Other factors such as inherent differences in body size may also influence diet choice as may behavioural learn¬ ing by the grazers. Body size may also have influenced food choice in the Perup Nature Reserve accounting for differences in the total number of species consumed by the macropods and the possums. Small animals require more energy relative to body mass than do larger animals (Freudenberger et al. 1989). This means that larger herbi¬ vores generally have a diet of lower nutritive value and higher fibre content than do smaller herbivores. The faecal material of the western grey kangaroo collected in the Perup Nature Reserve contained a higher number of species per pellet compared to any of the other marsupial herbivores. The kangaroo is larger in size and may retain digested material for longer periods of time, thus increas¬ ing species richness within the digestive system. Smaller animals such as the brush-tail and ring-tail possums would be expected to have high metabolic rates and nutri¬ tional requirements; however, both species consumed mostly Eucalyptus leaves which have a low nutritive value and contain toxic secondary compounds (Hume et al. 1984). The presence of several understorey species in the faeces of the common brush-tail possum indicates that this herbivore does not rely solely on the eucalypt canopy leaves as their only food resource. Both possum species may also supplement their diet with seeds, flow¬ ers and insect larvae. Ring- tail possums were found to have a very low field metabolic rate, much lower than expected for their body size (Hume et al. 1984). Due to the low metabolic rate these possums have lower total energy requirements, reflected in their decreased mobil¬ ity compared to the more active brush-tail possums. The stomach of the western ring-tail possum is also well adapted to a folivorous diet, as it is simple in structure with a well developed caecum. This possum has a long gut retention time for its size (38-39 hours) allowing in¬ creased fermentation of the digestible material consumed. Low field metabolic rates of the ring-tail possum are re¬ flected in low consumption rates, further minimising the consumption of the toxic compounds in the Eucalyptus foliage. The differences in diet preference between individuals and between different species may also reflect behavioural learning and experience. Animals sample and continue to consume a range of species which they can tolerate while others may have encountered and sub¬ sequently utilise, a different range of plants (Bartmann & Carpenter 1982; Provenza & Balph 1987, 1988; Gillingham & Bunnell 1989). The results of these analy¬ ses allow a direct comparison of the food resources utilised by the macropod species within the Perup Nature Reserve to other populations found in the state. Of major interest is the consumption of predominantly dicotyle¬ donous species by the western grey kangaroos. This is in contrast to the mixed diet of both dicotyledon and mono¬ cotyledon species recorded for populations of western grey kangaroos in pasture dominated landscape near Bakers Hill, Western Australia (Halford et al. 1984b; Bell 1993) and consumption of mainly monocotyledonous herbs in a Banksia woodland near Perth (Algar 1986). These differences in diet preference most likely reflect a difference in the predominance of these two subclasses as an available food source in the three study areas. The dietary findings for the black-gloved wallaby in the Perup Nature Reserve were consistent with those of Algar (1986) who found a wide range of species in the diet, a majority of which were dicotyledons. He also re¬ corded the consumption of tough sedge-like monocotyle¬ dons of the Banksia woodland habitat. The tammar wallabies on Garden Island consumed a number of species, but preferred grasses or new shoots that appear following a burn (Bell et al. 1987). Christensen (1977) suggested that tammar wallabies in the Perup region rely almost entirely on the grasses asso¬ ciated with Gastrolobium bilobum thickets. It is evident from this study, however, that these wallabies consume a much wider range of species than those immediately associated with their daytime refugia. Thus, tammars require the maintenance of the more diverse vegetation of the upper slopes as well as the Gastrolobium bilobum thickets in the low lying areas. The results from a dietary analysis by Jones et al. (1994b) of western ring-tail possum populations at Perup 52 Journal of the Royal Society of Western Australia, 80(2), October 1997 and in isolated refugia of coastal vegetation in the south west of Western Australia indicated that the canopy foli¬ age formed the main component of the possum diet. Pellets collected from regions of the Perup Nature Reserve away from the current study by Jones et ai (199b) also con¬ sisted of material of the two common canopy species. Eucalyptus marginata and E. calophylla. Analysis of pellets collected from coastal populations found that the common canopy species Agonis flexuosa dominated the possum diet selection (Inions et al. 1989). The presence of a number of unknown plant species in the faecal material of all the macropods indicates that their foraging range extends beyond that of the immedi¬ ate sampling site. Further study is essential to ensure establishment of appropriate management practices in Perup Nature Reserve to allow maintenance of marsupial herbivore population densities while considering the im¬ pact of grazing on surrounding vegetation. As all the marsupial herbivores utilise a number of common species and none of the macropods consume any single limited resource, it is unlikely that one group will out-compete another. Fire management in native forests requires a knowledge of the accumulation of understorey vegetative material and leaf litter within specified areas. A measured decrease in the vegetative cover outside the exclosures independent of fire confirms the impact of herbivores on the vegetation of Perup Nature Reserve. The continued maintenance of stable herbivore populations will have implications for the reduction of plant cover and leaf litter buildup and may have benefits in allowing slightly increased fire intervals in the prescription burning regime for this forest. Thus an understanding of herbivore population numbers over time is essential as a marked increase in numbers can have an equally significant impact on the vegetation as would a sharp decline. These are important areas for further research and must be considered in the long-term management of this reserve. Acknowledgments: Collection of data from the exclosures and surround¬ ing vegetation was assisted by students in the 1992 course in Synecology. Scat identification was assisted by P Christensen and C Wheeler, Manjimup Research Centre, Department of Conservation and Land Manage¬ ment (CALM). Use of the Perup Forest Ecology Centre was provided by CALM during the period of the study. Special thanks also go to C Wilkins for her encouragement and invaluable comments. References Algar D 1986 An ecological study of macropod marsupial species on a reserve. PhD Thesis. University of Western Australia, Perth. Bartmann R M & Carpenter L H 1982 Effects of foraging experi¬ ence on food selectivity of tame mule deer. Journal of Wild¬ life Management 46:813-818. Bell D T 1993 Plant community structure in south western Aus¬ tralia and aspects of herbivory, seed dispersal and pollination. In: Plant-Animal Interactions in Mediterranean-Type Eco¬ systems (eds M Arianoutsou & R H Groves). Kluwer Academic Publishers, Dordrecht, Netherlands, 63-70. Bell D T, Moredoundt J C & Loneragan W A 1987 Grazing pres¬ sure by the tammar (Macropus eugcnii Desm.) on the vegeta¬ tion of Garden Island, Western Australia, and the potential impact on food reserves of a controlled burning regime. Journal of the Royal Society of Western Australia 69:89-94. Brown J R & Stuth J W 1993 How herbivory affects grazing tolerant and sensitive grasses in a central Texas grassland: integrating plant responses across hierarchical levels. Oikos 67:291-298. Buchanan A P & Wardell-Johnson G 1990 Managing animal habitat using remote sensing and geographical information systems. Proceedings of the Fifth Australasian Remote Sensing Conference, Perth 2:1022-1031. Christensen P E S 1977 The biology of Bettongia penicillata Cray, 1837, and Macropus eugenii (Desmarest, 1817) in relation to fire. PhD Thesis. University of Western Australia, Perth. Christensen P E S 1980 The biology of Bettongia penicillata Gray, 1837, and Macropus eugenii (Desmarest, 1817) in relation to fire. Forests Department of Western Australia, Perth, Bulletin 91. Christensen P 1983 Western brush wallaby. In: The Complete Book of Australian Mammals (ed R Strahan). Angus and Robertson, Sydney, 235. Christensen P 1992 The Karri Forest. Department of Conservation and Land Management, Perth. Christensen P, Annels A, Liddelow G & Skinner P 1985 Verte¬ brate fauna in the southern forest of Western Australia, a survey. Forests Department of Western Australia, Perth, Bul¬ letin 94. Christensen P & Maisey K 1987 The use of fire as a management tool in fauna conservation reserves. In: Nature Conservation: The Role of Remnants of Native Vegetation (eds D A Saunders, G W Arnold, A A Burbidge & A J M Hopkins). Surrey Beatty, Sydney, 323-329. Dawson T J 1989 Diets of macropodid marsupials: general patterns and environmental influences. In: Kangaroos, Wallabies and Rat-kangaroos (eds G Grigg, P Jarman & I Hume). Surrey Beatty & Sons, Sydney, 129-142 Freeland W J & Janzen D H 1974 Strategies in herbivory by mammals: the role of plant secondary compounds. American Naturalist 198:269-89. Freudenberger D O, Wallis I R & Hume I D 1989 Digestive adaptations of kangaroos, wallabies and rat-kangaroos. In: Kangaroos, Wallabies and Rat-kangaroos (eds G Grigg, P Jarman & 1 Hume). Surrey Beatty & Sons, Sydney, 179-187. Gillingham M P & Bunnell F L 1989 Effects of learning on food selection and searching behaviour of deer. Canadian Journal of Zoology 67:24-32. Halford D A, Bell D T & Loneragan W A 1984a Epidermal char¬ acteristics of some Western Australian wandoo-woodland species for studies of herbivore diets. Journal of the Royal Society of Western Australia 66:111-118. Halford D A, Bell D T & Loneragan W A 1984b Diet of the western grey kangaroo ( Macropus fuliginosus Desm.) in a mixed pasture-woodland habitat of Western Australia. Journal of the Royal Society of Western Australia 66:119-128. Holechek J L, Martin V & Pieper R D 1982 Botanical composition determination of range herbivore diets: A review. Journal of Range Management 35:309-315. How R A 1983 Common brush tail possum. In: The Complete Book of Australian Mammals (ed R Strahan). Angus and Robertson, Sydney, 147-148. Hume I D, Foley W J & Chilcott M J 1984 Physiological mecha¬ nisms of foliage digestion in the greater glider and ringtail possum (Marsupialia: Pseudocheiridae). In: Possums and Gliders (eds A P Smith & l D Hume). Australian Mammal Society, Sydney, 247-251. Inions G B, Tanton M T & Davey S M 1989 The effect of fire on the availability of hollows in trees used by the common brushtail possum, Trichosurus vulpecula Kerr, 1872 and ring¬ tail possum, Pseudocheirus peregrinus Boddarts, 1785. Australian Wildlife Research 16:449-458. Johnson M K & Pearson H A 1981 Esophageal, fecal and exclosure estimates of cattle diets on a long leaf pine- bluestem range. Journal of Range Management 34:232-235. 53 Journal of the Royal Society of Western Australia, 80(2), October 1997 Jones B A, How R A & Kitchener D J 1994a A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae). I. Distri¬ bution and habitat. Wildlife Research 21:175-187. Jones B A, How R A & Kitchener D J 1994b A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae). II. Popula¬ tion studies. Wildlife Research 21:189-201. King D R, Oliver A J & Mead R J 1981 Bettongia and fluoroac- etate: a role for 1080 in fauna management. Australian Wild¬ life Research 8:529-536. Mueller-Dombois D & Ellenberg H 1974 Aims and Methods of Vegetation Ecology. John Wiley and Sons, New York. Murden S B & Risenhoover K L 1993 Effects of habitat enrich¬ ment on patterns of diet selection. Ecological Applications 3:497-505. Owen-Smith N & Cooper S M 1987 Palatability of woody plants to browsing ruminants in a south African savanna. Ecology 68:319-331. Poole W E 1983 Western grey kangaroo. In: The Complete Book of Australian Mammals (ed R Strahan). Angus and Robertson, Sydney, 248-249. Poole W E, Wood J T & Simms N G 1991 Distribution of the tammar, Macropus eugetiii, and the relationships of populations as determined by cranial morphometries. Wildlife Research 18:625-639. Provenza F D & Balph D F 1987 Diet learning by domestic rumi¬ nants: theory, evidence and practical implications. Applied Animal Behaviour Science 18:211-232. Provenza F D & Balph D F 1988 Development of dietary choice in livestock on rangelands and its implications for manage¬ ment. Journal of Animal Science 66: 2356-2368. Robbins C T, Hanley, T A, Hagerman A E, Hjeljord O, Baker D L, Schwartz C C & Mautz W W 1987 Role of tannins in defending plants against ruminants: reduction in protein availability. Ecology 68:98-10 7. Scotcher J S B 1979 A review of faecal analysis techniques for determining the diet of wild grazing herbivores. Proceedings of the Grassland Society of Southern Africa 14: 131-136. Storr G M 1961 Microscopic analysis of faeces, a technique for ascertaining the diet of herbivorous mammals. Australian Journal of Biological Sciences 14:157-164. Twigg L E & King D R 1991 The importance of fluoroacetate- bearing vegetation on native Australian fauna: a review. Oikos 61:412-430. Wardell-Johnson G & Nichols O 1991 Forest wildlife and habitat management in the southwestern Australia: knowledge, re¬ search and direction. In: Conservation of Australia's Forest Fauna (ed D Lunney). Royal Zoological Society of NSW, Mosman, 161-192. 54 Journal of the Royal Society of Western Australia, 80:55-62, 1997 Dietary preferences of the black-gloved wallaby ( Macropus irma) and the western grey kangaroo (M. fuliginosus ) in Whiteman Park, Perth, Western Australia J M Wann & D T Bell Department of Botany, The University of Western Australia, Nedlands WA 6907: email dbell@cyllene.mva.edu.au Manuscript received April 1996; accepted February 1997 Abstract Epidermal tissue trace analyses of faecal material indicated that the native banksia woodlands and adjacent managed grass areas of the Perth metropolitan Whiteman Park provided a varied diet for the black-gloved wallaby ( Macropus irma) and the western grey kangaroo (M. fuligmosus). Black-gloved wallabies fed on a range of species, with a total of 29 included in the diet. Cynodon dactylon, the dominant grass of the lawn areas, Carpobrotus edulis , a succulent species in roadside disturbed sites, and the native cycad, Macrozamia riedlei , generally of the native woodland, were the three species most frequently consumed. The diet of the western grey kangaroo included 25 species, with two grasses, Cynodon dactylon and Ehrharta calycina, the most frequently consumed. Western grey kangaroos showed a preference for monocotyledons, but dicotyledons also were common in the diet. Comparisons of thirteen chemical constituents and five morphological features of chosen and avoided plant species did not explain diet choice. A wide overlap of consumed plant species between the two herbivores indi¬ cated that competition for food resources was possible. However, the polyphagous diet for both species indicates that the ability to switch diet preferences probably precludes the competitive exclusion of one by the other. Analysis of the epidermal material of the stomach, the small intestine, and the large intestine and colon indicated that the digestive system of the black-gloved wallaby did not completely remove plant epidermal traces of any of the ingested species. Therefore, the acid digestion tech¬ nique used for faecal pellet analysis potentially indicates all the species consumed by the animal. Retention of the mosaic of natural banksia woodland cover with adjacent areas of watered lawn should benefit the limited population of black-gloved wallabies in Whiteman Park. At present the population of western grey kangaroos is not large enough to consume enough material to affect either the population of black-gloved wallabies or the structural elements of the vegetation of Whiteman Park. Introduction Management programs are becoming increasingly im¬ portant in the maintenance of wild animal populations. Land clearing and expansion of urban areas has resulted in small isolated remnants of bushland. Such 'remnant islands' are difficult for mammal populations to recolonise. Whiteman Park, in north-east metropolitan Perth, is such an 'island'. Whiteman Park is presently inhabited by two Macropodidae species, the black-gloved wallaby ( Macropus irma ) and the western grey kangaroo (M. fuliginosus). The population of black-gloved walla¬ bies in the Park has been estimated at 40 animals (Arnold et al. 1991). The population of western grey kangaroos is much higher, at between 550 (Arnold et al. 1991) and 1000 animals (H Gratte, pers comm). Maintenance of these animals in the park is a priority, but little is known of their dietary preferences or the potential that one species might require the same resources and exclude the other. Because of the low numbers of black-gloved wallabies in Whiteman Park, a non-destructive method of deter¬ mining their dietary preferences was required. Procedures used to estimate the botanical composition of herbivore © Royal Society of Western Australia 1997 diets include direct observation, utilisation techniques, analysis of mouth contents, fistula techniques, stomach analysis, and faecal analysis. Certain procedures are more useful than others, but each has important limita¬ tions. In this instance, faecal analysis was the preferred technique for these nocturnal macropods that often show severe stress when trapped. One road-killed black-gloved wallaby, however, afforded the opportunity to analyse digestive system contents and to provide a 'control' to the faecal pellet dietary samples. The objectives of this study were to determine the plant species consumed by the black-gloved wallaby and the western grey kangaroo, to assess characteristics of favoured plants, and to determine the potential for re¬ source competition between these two marsupial herbi¬ vores in the confined area of Whiteman Park. Materials and Methods Study site Whiteman Park, comprising 2600 ha of natural bush and pasture, is situated approximately 20 km north-east of the central business district of Perth. The park is situated 55 Journal of the Royal Society of Western Australia, 80(2), October 1997 on rolling sand dunes of the Bassendean soil association (Bettenay et al. 1960) over a portion of the important ground-water aquifer, the Gnangara Mound, that serves the metropolitan water supply (Anon 1992). The vegeta¬ tion is a mixture of low woodland of Banksia attenuata, B. menziesii and Eucalyptus todtiana on upper slopes, grad¬ ing through a taller woodland dominated by E. marginata, E. calophylla and B. ilicifolia to moist lowlands where a woodland of Melaleuca preissiana and B. littoralis occurs (Anon 1989). Stream margins in the southern end of the Park harbour an open forest dominated by E. rudis and M. rhaphiophylla. Mammal study species The black-gloved wallaby is a small wallaby (up to 9 kg), which is reported to graze rather than browse, and prefers a habitat of open forest or woodland (Christensen 1983). Due to its trap-shy nature and difficulty in handling, the black-gloved wallaby has rarely been studied. How¬ ever, Shepherd et al. (1997) documented that the black- gloved wallabies in the Perup region of the southern jarrah forest consume at least 21 different plant species. Christensen (1983) also reported that they appear to be able to survive without free water. The black-gloved wallaby adds to the conservation value of Whiteman Park, especially since it is the nearest relative to the now extinct toolache wallaby (Macropus greyi). Although still common in a number of jarrah forest regions, the black- gloved wallaby is rare in the greater Perth urban region. The western grey kangaroo is larger than the black- gloved wallaby, with males reaching 54 kg (Poole 1983). The western grey kangaroo both grazes and browses, and its diet has previously been shown to include a wide range of plant species and life-form types (Halford et al. 1984b; Priddel 1986; Bell 1994). Western grey kangaroos in a region of mixed wandoo-pasture country near Bakers Hill fed both on pasture species and a range of native plant species with some indication that nitrogen-rich legume species were selected in percentages greater than the percentages found as vegetative cover (Halford et al. 1984b). Western grey kangaroos in the Perup region of the southern jarrah forest consumed at least 32 different plant species, and marsupial herbivores appear to have the capacity to significantly reduce plant cover by their grazing pressure (Shepherd et al. 1997). Plant resources The detailed documentation of available plant re¬ sources for the two herbivore species was carried out in an area of approximately 700 ha in the northern region of Whiteman Park, where both marsupial herbivores have been observed. Detailed plant resource investigations were centred on three sites and covered a range of habitats. Site 1 was directly north and north east of the archery range parking lot. Site 2 included the archery range and the bushland to its east. Site 3 was approximately 300 m north-east of the International Trap and Skeet Shooting Complex. At each study site, percentage cover for each plant species was estimated for ten 2 m x 2 m quadrats placed along a randomly selected 20 m transect. Mean percentage cover and frequency of occurrence of each plant species were determined by combining the data for all three study transects, as the home ranges of both species would be larger than any one single site. Fresh material was collected for moisture content, ash content and morphological characteristics for all plant species. It was stored in plastic bags inside an ice-filled cooler chest and returned to the laboratory within 2-3 h. Approximately 1 to 5 g of fresh leaf material was weighed, placed in paper bags, oven dried at 60 °C for 48 h, and re-weighed to determine percentage moisture content. Ash content was determined using dried plant material fired in a muffle furnace at 550 °C for 2 h. Dried leaf material from 82 plant species was analysed for 13 chemical constituents using flame spectrophotometry (CSBP & Farmers Ltd, Bayswater, Western Australia). Morphological characteristics of leaves from each plant species in the study area were determined from dried specimens. In plants with leaves absent or reduced, the morphological characteristics of stems or branches were determined. Five characteristics thought to influence feeding choice were summarised into the following cat¬ egories; 1) apex type described the shape of the apices of leaves or branches /stems; 2) apex hardness separated all spine-tipped apices from soft-tipped, rounded, square-tipped and in¬ dented-tipped apices; 3) glands were noted as either present or absent; 4) leaf consistency scored as succulent, mesophyll, semi-sclerophyll or sclerophyll; and 5) both adaxial and abaxial tomentosity of leaves scored as either low, medium or highly pubescent or non-hairy. These characteristics of plant species subsequently confirmed by faecal analysis as food resource species were compared to those of species not in faecal pellets using unpaired, two-tailed t-tests and coded chi-square tests. Such tests identify deviations from random and, therefore, can be used to determine if diet selection is occurring. Epidermal reference collection All plant species were collected at the study sites dur¬ ing the early autumn. Leaf material was used for the epidermal tissue reference collection for most species (see Halford et al. 1984a). For plants with phyllodes, cladodes or very reduced leaves, the petioles, stems, or branch materials (the materials that are potentially available as food for herbivores), were used for the epidermal refer¬ ence collection. Two methods were used to separate the epidermal material from the underlying tissues. The first, a modifi¬ cation of Jain (1976), involved placing 5 mm x 5 mm squares, or whole small leaves, in glass vials with 50% glacial acetic acid. The vials were placed in a water bath at 80 °C for 24-48 h, depending on the sclerophyllous nature of the material. On removal from the water bath, the plant material was washed with water and the remain¬ ing fibrous tissue was removed under a dissecting micro¬ scope using tweezers, dissecting needles and scalpel blades. The epidermal material was then dehydrated through a series of ethanol solutions to 95% and stained with 0.5% gentian violet in 95% ethanol for 48 h. The stained material was then rinsed in absolute alcohol and mounted in Eukitt®. The second technique, generally used for thicker material, employed a method similar to 56 Journal of the Royal Society of Western Australia, 80(2), October 1997 that used by Storr (1961). Leaf material was placed in vials and covered with 10% chromic acid and 10% nitric acid, and boiled in an acid digestion heating block at 115 °C for 20-25 min. The tissue was then rinsed in 0.1 M KOH and the epidermal material was peeled off using tweezers and dissecting needles. The epidermal material was then dehydrated, stained and mounted as before. Epidermal preparations were drawn to provide a vi¬ sual record of the important characteristics of each plant species of the area. Faecal sample collection and preparation There was little difficulty in discriminating between the faecal pellets of black-gloved wallabies and western grey kangaroos. Faecal pellets of the former were smaller, round to oval, and pinched at the ends. Pellets of the latter were larger and at times almost cubic in shape. Fresh pellets were collected on a single day dur¬ ing each of the four seasons from the three vegetation study sites. Where there was more than one pellet in a deposit, only one was collected. For each season, one pellet from each study site for each of the herbivores was randomly selected for frag¬ ment identification. Pellets were thoroughly dried, bro¬ ken apart using tweezers, placed in a test tube and cov¬ ered with 20 ml of equal parts 10% chromic acid and 10% nitric acid. The test tubes were placed in the acid diges¬ tion block in a fume hood and boiled at 115 °C for 20-25 min. The material was allowed to cool to room tempera¬ ture before being filtered through a buchner funnel and Whatman No 10 filter paper with several washes of 0.1 M KOH. The filtrate was then collected and stained with 0.5% gentian violet in 95% alcohol for 48 h. After stain¬ ing, the filtrate was passed through a 0.5 mm sieve and the remaining material was washed several times with 70% alcohol to remove excess stain. The tissue was placed in 95% alcohol in a glass petri dish and viewed through a dissection microscope at 6x to 30x magnifica¬ tion. Different species could generally be identified using the dissection microscope at its highest magnification. In some cases, a fragment was washed in absolute alcohol, mounted and viewed under a binocular microscope. These epidermal fragments from faeces were then com¬ pared to the drawings of the epidermal reference collec¬ tion and identification was confirmed by comparing the slide of the faecal epidermal fragment to the original plant voucher collection slide. Once a species had been positively identified, the proportion of each plant species in each pellet was subjectively determined as rare, com¬ mon or abundant. Black-gloved wallaby digestive tract contents The carcass of an adult male black-gloved wallaby, killed on the road near the western boundary of Whiteman Park, was refrigerated within 5 h of its death. After 36 h in refrigeration, the digestive tract was removed and the stomach, small intestine and large intestine contents were separated. The large intestine and rectum contents could not be completely separated and were analysed together. The digestive contents were washed with water using a fine sieve to remove the digestive acids before being stored in alcohol. Epidermal tissue of the constitu¬ ent plant species was obtained, stained and mounted us¬ ing the methods described above. Results Plant resources available for herbivory A total of 73 species were encountered in the transect samples, with mean cover ranging from more than 3% for Patersonia occidentalis to 0.01% for a number of species, such as Adenanthos cygnorum, Arnocrinum preissii, Danthonia setacea and Oxylobium capitatum, that were encountered only once (Table 1). The more common species included Patersonia occidentalis, Xanthorrhoea preissii, Beaufortia elegans, Leucopogon conostephioides, Hypocalymma angustifolium and Alexgeorgea nitens. Of the total species, 88% had a mean cover value less than 1%. No species was found in all 30 quadrats. Gladiolus caryophyllaceus , Lyginia barbata , Hibbertia subvaginata and Patersonia occidentalis had the highest frequency values, ranging from 85% to 62%. Of the total species, 24% were found in only a single quadrat. Black-gloved wallaby digestive tract contents A total of eight species (seven dicotyledons and one monocotyledon) were identified in the stomach and in¬ testinal tract of the road-killed black-gloved wallaby Table 1 Mean percentage cover in quadrats and percentage frequency of plant species in the 30 quadrats sampled in Whiteman Park (see text for quadrat locations). Species Percentage Cover Percentage Frequency Patersonia occidentalis 3.03 62.5 Xanthorrhoea preissii 2.95 35.0 Beaufortia elegans 2.20 32.5 Leucopogon conostephioides 2.15 37.5 Hypocalymma angustifolium 2.02 15.0 Alexgeorgea nitens 1.86 60.0 Eremaea pauciflora 1.70 22.5 Lyginia barbata 1.29 70.0 Calytrix angulata 1.25 32.5 Hibbertia subvaginata 0.93 70.0 Stirlingia latifolia 0.92 27.5 Lechenaultia floribunda 0.70 15.0 Schoenus curvifolius 0.49 40.0 Dampiera linearis 0.49 30.0 Hibbertia hypericoides 0.45 22.5 Hemiandra linearis 0.45 12.5 Conostephium pendulum 0.44 30.0 Dasypogon bromeliifolius 0.42 20.0 Scholtzia involucrata 0.36 22.5 Gompholobium tomentosum 0.36 20.0 Stylidium repens 0.35 32.5 Acacia pulchella 0.33 20.0 Hypocalymma robustum 0.31 17.5 Calytrix flavescens 0.30 22.5 Allocasuarina fraseriana 0.30 5.0 Banksia attenuata 0.25 7.5 moss 0.24 27.5 Loxocarya flexuosa 0.24 20.0 Stipa sp 0.20 15.0 Banksia ilicifolia 0.20 7.5 Melaleuca preissiana 0.18 2.5 Trachymene pilosa 0.15 32.5 Hibbertia huegelii 0.15 12.5 57 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 1 (continued) Species Percentage Cover Percentage Frequency Gladiolus caryophyllaceus 0.13 85.0 Petrophile linearis 0.13 12.5 Airia sp 0.12 5.0 Bossiaea eriocarpa 0.09 17.5 Waitzia podolepis 0.09 17.5 Acacia sessilis 0.09 15.0 Jacksonia floribunda 0.08 10.0 Leucopogon polymorphus 0.08 10.0 Briza maxima 0.08 5.0 orchid spp 0.07 20.0 Eriostemon spicatus 0.06 17.5 Banksia menziesii 0.06 15.0 Drosera erythrorhiza 0.05 25.0 Pimelea sulphurea 0.05 5.0 Persoonia saccata 0.05 5.0 Daviesia triflora 0.05 2.5 Scaevola paludosa 0.05 2.5 Acacia stenoptera 0.03 2.5 Burtonia scabra 0.03 2.5 Calectasia cyanea 0.03 2.5 Hibbertia racemosa 0.03 2.5 Regelia ciliata 0.03 2.5 Ehrharta calycina 0.02 20.0 Haemodorum spicatum 0.02 17.5 Drosera macrantha 0.01 10.0 Ursinia anthemoides 0.01 7.5 Allocasuarina humilis 0.01 5.0 Hypochaeris glabra 0.01 5.0 Lepidosperma angustatum 0.01 5.0 Stylidium brunonianum 0.01 5.0 Thysanotus sp 0.01 5.0 Adenanthos cygnorum 0.01 5.0 Anizoganthos humilis 0.01 2.5 Amocrinum preissii 0.01 2.5 Banksia grandis 0.01 2.5 Danthonia setacea 0.01 2.5 Lomandra hermaphrodita 0.01 2.5 Oxylobium capitatum 0.01 2.5 Tricoryne elatior 0.01 2.5 rhizomatous monocotyledon 0.01 2.5 Table 2 Plant species identified in the stomach, small intestine and large intestine of a road-killed black-gloved wallaby. The subjective scale of fragment abundance in the sam pies is rare (R), common (C), abundant (A) and not recorded (x). Unknown dicotyledons #8 and #9 were not included in the voucher collection from the study site. Small Large Species Stomach Intestine Intestine Adenanthos cygnorum R R R Amocrinum preisii C C C Eriostemon spicatus C C C Hovea trisperma C c C Nuytsia floribunda R R R Leucopogon conos tephioides R R X dicotyledon #8 C C C dicotyledon #9 C C C Total Species Identified 8 8 7 (Table 2). Seven plant species were identified from all three parts of the digestive tract. Leucopogon conostephioides was present in very small quantities in the stomach and small intestine, but was not identified recorded for the large intestine. The similarity of species content throughout the digestive tract confirmed that faecal pellet analysis provides a sufficiently accurate indication of these plant species. Black-gloved wallaby faecal pellets Twenty-nine plant species were recovered from the black-gloved wallaby faeces collected at Whiteman Park (Table 3). On average, six to seven plant species were found in each pellet, and season had no effect on the number of plant species consumed. Of the 29 species con¬ sumed, 21 were positively identified using the epidermal reference collection. The remaining eight (seven dicotyle¬ don species and one monocotyledon species) were not able to be identified to species. Three of the 21 identified species consumed were the succulent exotic Carpobrotus edulis and the two introduced grasses Cynodon dactylon and Ehrharta calycina. Carpobrotus edulis was especially common in all faecal pellets during summer and autumn, but was absent from the pellets in winter and spring. Cynodon dactylon is the grass used in the managed and watered lawn areas of the archery and skeet shooting ranges. Ehrharta calycina is especially common in dis¬ turbed areas along the roads and parking lots. The most common native species in the black-gloved wallaby diet was Macrozamia riedlei, occurring in 50% of the faecal pellets analysed. Tricoryne elatior was found in 42% of the pellets, while Leucopogon conostephioides , Nuytsia floribunda and the unknowm dicotyledon #1 were found in 33% of the pellets examined. Eleven species, including Patersonia occidentals , Conostephium pendulum and Mesomelaena stygia, were found in only a single pellet. Carpobrotus edulis , Cynodon dactylon , Nuytsia floribnnda and unknown dicotyledon #1 wTere the only species classed subjectively as abundant in one or more faecal pellets. The remaining 25 species were classified as either common or rare in faecal pellets. All but one of the species consumed ( Macrozamia riedlei) belong to the class Magnoliopsida. Of these, 64% were dicotyledons and 36% were monocotyledons. This compares to the 69% dicotyledons:31% monocotyledons proportions noted for the ratio found in the vegetation survey. Western grey kangaroo faecal pellets Twenty-five species were represented in the faecal pel¬ let samples of the western grey kangaroo of Whiteman Park (Table 4). An average of six plant species was found per pellet over the entire year, with no obvious seasonal pattern observed in the number of species in each sample. Of the 25 species identified in the faecal material, 20 species were identified using the epidermal reference collection. The same three exotic species consumed by the black-gloved wallabies were found in the pellets of the western grey kangaroos. Carpobrotus edulis was found in 42% of the pellets examined, and was common to abundant in most of the pellets in all seasons except spring. Cynodon dactylon was the most frequently consumed exotic species, occurring in 58% of the faecal pellets. The tufted perennial Alexgeorgca nitens was the most commonly identified native species, occurring in 75% of the samples. Other native species commonly identified in 58 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 3 Plant species identified in black-gloved wallaby faecal pellets collected during the four seasonal sampling periods. Three pellets were analysed during each sampling period. The subjective scale of fragment abundance in the pellet is rare (R), common (C) and abundant (A). Percentage of occurrence (%) for all sample pellets (n=12) through the year, and the total species richness in each sample pellet, are shown. Plant Species Site 1 Summer 2 3 1 Autumn 2 3 Winter 1 2 3 1 Spring 2 3 % Acacia stenoptera R R 17 Alexgeorgea nitens C 8 Amocrinum preissii R 8 Beaufortia elegans R C R 25 Carpobrotus edulis A A C R C C 50 Conostephium pendulum R 8 Corynotheca micrantha C R 17 Cynodon dactylon C C A A C A C C C C 83 Dampiera linearis R R 17 Danthonia setacea R R 17 Ehrharta calycina R R 17 Eucalyptus marginata R 8 Leucopogon conostephioides R R C C 33 Leucopogon sp A C C C 25 Lysinema ciliatum R 8 Macrozamia riedlei R R C R c R 50 Mesomalaena stygia R 8 Nuytsia floribunda A R A R 33 Oxylobium capitatum R R 17 Patersonia occidentalis R 8 Tricoryne elatior C C C C C 42 dicotyledon #1 A A A C 33 dicotyledon #2 C R R 25 dicotyledon #3 R 8 dicotyledon #4 C 8 dicotyledon #5 C 8 dicotyledon #6 R R 17 dicotyledon #7 R 8 monocotyledon #1 C C C 25 Total Species Richness 6 5 7 5 7 9 7 8 4 5 8 5 the western grey kangaroo faecal material included Leucopogon sp A, Leucopogon conostephioides, Danthonia setacea, Corynotheca micrantha and Dampiera linearis. Of the 24 angiosperm species in the diet, 58% were dicotyle¬ donous species and 42% were monocotyledonous species. As for the black-gloved wallaby, the faecal pellets of the western grey kangaroo tended to contain a number of species of plants with few being in abundant concentra¬ tions. Diet overlap Both the black-gloved wallaby and the western grey kangaroo were versatile feeders, consuming a wide range of plant species. There was considerable overlap in the feeding preferences, with 18 species consumed by both macropods. Species that frequently appeared in the fae¬ cal pellets of both animals included Carpobrotus edulis, Cynodon dactylon, Leucopogon conostephioides and Leucopogon sp A. Some species were in a large proportion of the faecal pellets from one macropodid, but only in one or two pellets from the other species. For example, Alexgeorgea nitens occurred in 75% of the western grey kangaroo faecal pellets but only 8% of the black-gloved wallaby faecal pellets. Other species such as Ehrharta calycina, Oxylobium capitatum and Patersonia occidentals, were in only one or two faecal pellets from both macropods. Food resource characteristics The plant species consumed by black-gloved wallabies had a mean field cover of 0.43% compared to a mean cover of 0.24% for those not consumed, but the difference was not significant (Table 5). Similarly, mean percentage covers of the consumed and non-consumed species were not significantly different for the diet selected by the western grey kangaroos. Of the thirteen chemical con¬ stituents considered, there were no significant differences observed between the mean content in plants eaten and those not eaten by the black-gloved wallaby for any of the comparisons. Results were similar for the western grey kangaroo, except nitrate in dietary resources was significantly higher in the plants chosen compared to the plants avoided. Moisture content was not significantly different for plant species consumed and avoided, for either macropod. The ash content of the plants chosen and avoided were also not significantly different for the diets of either species. 59 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 4 Plant species identified in western grey kangaroo faecal pellets collected during the analysed during each sampling period. The subjective scale of fragment abundance (A). Percentage of occurrence (%) for all sample pellets (n=12) through the year, and shown. four seasonal sampling periods. Three pellets were in the pellet is rare (R), common (C) and abundant the total species richness in each sample pellet, are Plant Species Site 1 Summer 2 3 Autumn 1 2 3 Winter 1 2 3 1 Spring 2 3 % Adenanthos cygnorum R C R 25 Alexgeorgea nitens C C C R C C C C C 75 Bossiaea eriocarpa R R 17 Carpobrotus edulis C c C C A 42 Conostephium pendulum C C 17 Corynotheca rnicrantha R R C R 33 Cynodon dactylon c C c C A c C 58 Dampiera linearis R R R c 33 Danthonia selacea C C c C C 42 Ehrharta calycina R A 17 Jacksonia furcellata R 8 Leucopogon conostephioides C C C C 33 Leucopogon sp A C C R C C R 50 Loxocarya flexuosa R 8 Macrozarnia riedlei R C 17 Nuytsia floribunda c 8 Oxylobium capitatum C R 17 Patersonia occidentalis R 8 Stipa sp A R 8 Tricoryne elatior C 8 dicotyledon #1 c C 17 dicotyledon #4 R 8 dicotyledon #7 C C c C 33 dicotyledon #8 R 8 monocotyledon #2 R 8 Total Species Richness 7 7 7 3 7 5 7 4 5 9 6 5 Table 5 Characteristics of species chosen (those positively identified in faecal material) and those avoided (those not confirmed in faecal material) in the diets of the black-gloved wallaby and western grey kangaroo populations in Whiteman Park. NS is not significant. Black-gloved wallaby Western grey kangaroo Characteristic Chosen Avoided t or X2 Sig. Diff. Chosen Avoided t or X2 Sig. Diff. Field vegetation Field Cover (%) 0.43 ± 0.18 0.24 ± 0.06 1.33 NS 0.42 ± 0.19 0.25 ± 0.06 1.07 NS Plant Frequency (%) 12.81 ± 3.94 11.53 ± 1.88 0.29 NS 14.25 ± 4.48 11.24 ± 0.70 0.70 NS Chemical, moisture and total ash contents Calcium (%) 0.66 ± 0.11 0.75 ± 0.06 -0.76 NS 0.61 ± 0.12 0.76 ± 0.06 -1.19 NS Chloride (%) 0.75 ± 0.03 0.71 ± 0.09 0.17 NS 0.99 ± 0.38 0.64 ± 0.08 1.42 NS Copper (ppm) 5.17 ± 0.48 6.18 ± 0.61 -0.96 NS 5.49 ± 0.49 6.03 ± 0.59 -0.49 NS Iron (ppm) 148.40 ± 36.20 117.00 ± 14.70 0.96 NS 174.23 ± 40.07 110.73 ± 14.08 1 .89 NS Magnesium (%) 0.20 ± 0.03 0.25 ± 0.02 -1.35 NS 0.19 ± 0.04 0.25 ± 0.02 -1.45 NS Manganese (ppm) 46.12 ± 6.89 71 ±94 -1.62 NS 61.55 ± 10.50 66.05 ± 8.75 -0.26 NS Nitrate (ppm) 33.18 ± 9.10 27.14 ± 1.34 1.02 NS 38.42 ± 10.27 25.82 ± 1.32 2.07 P<0.05 Nitrogen (%) 1.10 ± 0.09 1.17 ± 0.07 -0.54 NS 1.16 ± 0.12 1.15 ± 0.07 0.09 NS Phosphorus (%) 0.04 ± 0.00 0.05 ± 0.01 -0.88 NS 0.05 ± 0.01 0.05 ± 0.01 0.25 NS Potassium (%) 0.58 ± 0.09 0.55 ± 0.09 0.22 NS 0.63 ± 0.10 0.53 ± 0.06 0.82 NS Sodium (%) 0.31 ± 0.10 0.25 ±0.03 0.17 NS 0.34 ± 0.11 0.25 ± 0.03 1.15 NS Sulphur (%) 0.19 ± 0.02 0.24 ± 0.02 -1.23 NS 0.21 ± 0.02 0.24 ± 0.02 -0.72 NS Zinc (ppm) 30.73 ± 5.30 43.72 ±8.78 -0.87 NS 40.89 ± 6.06 40.04 ± 8.40 0.05 NS Moisture Content (%) 52.25 ± 2.60 55.83 ± 1.55 1.17 NS 51.86 ± 3.18 55.72 ± 1.46 1.18 NS Ash Content (%) 3.75 ± 0.73 4.60 ± 0.43 -0.98 NS 3.58 ± 0.87 4.60 ± 0.41 1.11 NS 60 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 5 (continued) Black-gloved wallaby Western grey kangaroo Characteristic Chosen Avoided tor X2 Sig. Diff. Chosen Avoided t or X2 Sig. Diff. Morphological features Apex Type — — 12.08 NS — — 12.57 NS Apex soft or hard — — 0.51 NS — 0.79 NS Glands present or absent — — 1.31 NS — 6.55 P<0.05 Consistency — — 5.22 NS — 6.60 NS Adaxial tomentosity — — 2.30 NS — 2.54 NS Abaxial tomentosity — — 2.64 NS — — 5.30 NS The morphological features of the plants selected and avoided by the two macropods were not generally differ¬ ent (Table 5). The type of leaf apex, whether the apex was soft or hard, the presence or absence of glands, leaf consistency, and adaxial and abaxial tomentosity, had no relationship to the types of plants consumed by black- gloved wallabies. Western grey kangaroos, however, avoided plants with leaf glands, notably the Myrtaceae. Other morphological characteristics had no relationship to the choices made by the western grey kangaroos. Discussion Central to the study of animal ecology is the use a herbivore makes of its environment, specifically the types of plants it consumes in the different habitats it occupies. It is not only worthwhile knowing what a herbivore eats, but also when and how much is consumed, the availabil¬ ity of the resource and the nutritional value of the plant to the animal (Free et al. 1970). It is also advantageous to understand the relationship between a herbivore and other species with which it co-exists, particularly the level of competition for shared resources. When shared resources are in limited supply, competition will eventually lead to the more successful competitor excluding the other. Management of animal populations in restricted range habitats, therefore, requires a wide array of information. In this study, the black-gloved wallabies of Whiteman Park consumed a wide range of species, generally in moderate or small amounts rather than a few species in large amounts. However, Carpobrotus edulis, Cynodon dactylon, and Nuytsia floribunda were found in abundant amounts in the faeces. The succulent Carpobrotus edulis and the lawn Cynodon dactylon were the most abundantly consumed species in the drier seasons, a time when free¬ standing water was not readily available in the study area. The hemi-parasitic Nuytsia floribunda, the only species noted as abundant in the faecal samples of winter samples, might also be a likely source of moisture due to its ability to absorb water from neighbouring plants. The black-gloved wallaby has been noted by Christensen (1983) to be able to survive without free water. The Whiteman Park population of black-gloved wallabies would be especially well served with these moisture-rich plants for their source of water. The western grey kangaroos of Whiteman Park showed a similar tendency as black-gloved wallabies to consume a wide variety of plant species in limited amounts. As for black-gloved wallabies, these larger marsupials consumed abundant quantities of the exotic, succulent Carpobrotus edulis and the lawn grass Cynodon dactylon among the 25 dietary plant species. As in pre¬ vious studies (Halford et al. 1984b; Shepherd et al. 1997), it appears that the western grey kangaroo is a versatile feeder, consuming a wide variety of plant species and plant types. Consumption of a wide variety of plant species is seen both as an adaptation to the prevention of adverse effects from plant poisons (Howe & Westley 1988), but also a necessity, as the plant species of Whiteman Park have very low cover percentages. A mixed species diet is therefore required to provide sufficient energy for sur¬ vival. Limited grazing and browsing on a range of plant species is also of significance to the management of Whiteman Park, as herbivore pressure on any individual plant species is not likely to result in local extinction. Hume's (1982) review of marsupial nutritional require¬ ments concludes that potoroid and small macropod species generally ingest plants of higher nutritive value than do larger (greater than 10 kg) macropods. A relationship be¬ tween body size and nutritive value of preferred diets has also been reported for eutherian herbivores (Demment & Van Soest 1985). Total energy requirement of animals tends to increase with increasing body mass, but at a progressively slow rate. Thus, although larger animals require more total energy (kj d1), small animal require more energy relative to body mass (kj kg'1 d1) (Freudenberger et al. 1989). It may be that subtle differ¬ ences in food perception are responsible for the observed differences in selection. A smaller animal may be capable of distinguishing food types on a finer scale than larger animals (Norbury & Sanson 1992). In this study, how¬ ever, the measured chemical constituents and the range of morphological features described for leaves between chosen and avoided species were not different. It is believed that different plants within a plant community have varying payabilities for grazing sheep (Krueger et al. 1974). Although taste is of primary importance in diet selection in sheep, smell, sight and touch also play a role, although they may be of minor importance. Unfortunately, such information is not known for the majority of herbi¬ vores, including macropods. Until it is known whether senses such as taste, smell, sight and touch significantly affect a macropod's food perception, and if so to what extent, problems will remain as to whether measured food availability is the same as perceived food availability. Halford et al. (1984b) showed that plants potentially high in nitrogen (i.e. legume species) were preferentially 61 Journal of the Royal Society of Western Australia, 80(2), October 1997 browsed by western grey kangaroos in a Eucalyptus wandoo- grass pasture landscape near Bakers Hill. The re¬ sults in the present study indicated that nitrate levels of selected species was higher than avoided species. Al¬ though an ability to select plant species of high nitrate condition or high protein level would be a valuable asset for the herbivore, we feel this finding should be viewed with caution for a number of reasons. Firstly, it is doubt¬ ful that western grey kangaroos would not obtain enough nitrogen in their normal diet and, therefore, seek out plants high in nitrates. Even if they did, kangaroos would need the appropriate microbial symbionts in their digestive systems to convert the nitrates to nitrites, so they could be assimilated. We are not aware if this species has such microbial symbionts. Secondly, it is also doubtful if these kangaroos can perceive parts per million concen¬ tration differences in leaf content of nitrates. Thirdly, due to the number of comparisons analysed, it could be ex¬ pected that one comparison might be falsely indicated as different. Further detailed comparisons would be neces¬ sary to confirm the ability of this marsupial herbivore to select high nitrate containing plants. Western grey kangaroos appear to avoid plant species with oil glands. Black-gloved wallabies, on the other hand, consumed a number of species with oil glands, but only in small or moderate amounts. The evolutionary sig¬ nificance of oil glands in the Myrtaceae and Rutaceae families is uncertain, but it is possible that they play a role in the prevention of herbivory. However, the detec¬ tion of diet selection by herbivores is a very complicated task. Many factors must be considered, such as positive nutritional elements, negative compounds, such as lignin, tannins, pectins and toxin, water requirements and mor¬ phological characteristics of leaves, stems and branches. Considering each of these factors separately can give clues to diet selection, but an ideal method of analysis would simultaneously take all factors into consideration for each plant species. For example, a plant species may have high positive nutrients and high water content, but the present of pungent spines or toxins may render the plant inedible for certain herbivore species. Searching for trends in the light of specific characters may, therefore, be an oversimplified and inappropriate method of deter¬ mining diet selection. Although the use of more sophisti¬ cated multivariate statistical techniques might lend further knowledge on diet selection, the characters are not inde¬ pendent of one another, and therefore, this statistical technique is invalid on mathematical grounds. This study of dietary preference in black-gloved wallabies and western grey kangaroos in Whiteman Park has confirmed their polyphagous nature. The diversity of plant species in Whiteman Park provides a wide range of dietary choices for both herbivores. Competitive exclusion is not considered to be cause for concern as the vegeta¬ tion of Whiteman Park provides a large number of ac¬ ceptable food resources for both animals. The provision of significant numbers of high moisture content plants would favour the continued survival of both species, but especially the black-gloved wallabies that require plant- derived moisture rather than free-water sources. Acknowledgments: We are grateful for cooperation throughout the study to A Brian, H Gratte and R Wagland of the Whiteman Park Staff and K A Shepherd of the Department of Botany, University of Western Australia. Funds for the research were provided through the Department of Botany Teaching and Research Block Grants. References Anon. 1989 Ecological Studies - Whiteman Park. Mattiske and Associates Report. The Western Australian State Planning Commission. Perth, Western Australia. Anon 1992 Gnangara Mound Vegetation Stress Survey. Results of Investigation. The Water Authority of Western Australian, Leederville, Western Australia. Arnold G W 1991 Whiteman Park Fauna Survey. CSIRO Division of Wildlife and Ecology, Perth. Bell D T 1994 Plant community structure in southwestern Australia and aspects of herbivory, seed dispersal and pollination. In: Plant-Animal Interactions in Mediterranean Ecosystems (eds M Arianoutsou & R H Groves). Kluwer Academic Publishers, Dordrecht, 63-70. Bettenay E, McArthur W M & Hingston F J 1960 The soil associa¬ tions of part of the Swan Coastal Plain, Western Australia. Soil and Land Use Series, 35. CSIRO, Division of Soils, Perth. Christensen P 1983 Western brush wallaby. In: The Australian Museum Complete Book of Australian Mammals (ed R Strahan). Angus and Robertson, Sydney, 235. Demment M W & Van Soest P J 1985 A nutritional explanation for body size patterns of ruminant and non-ruminant herbi¬ vores. American Naturalist 125:641-672. Free J C, Hansen R M & Sims P L 1970 Estimating dry weight in food plants in faeces of herbivores. Journal of Range Manage¬ ment 23:300-302. Freudenberger D O, Wallis I R & Hume I D 1989 Digestive adaptations of kangaroos, wallabies and rat-kangaroos. In: Kangaroos, Wallabies and Rat-Kangaroos (ed G Grigg, P Jarmann & I Hume). Surrey Beatty & Sons, Sydney, 179-187. Halford D A, Bell D T & Loneragan W A 1984a Epidermal char¬ acteristics of some Western Australian wandoo-woodland species for studies of herbivore diets. Journal of the Royal Society of Western Australia 66:111-118. Halford D A, Bell D T & Loneragan W A 1984b Diet of the western grey kangaroo (Macropus fuliginosus Desm.) in a mixed pasture-woodland habitat of Western Australia. Journal of the Royal Society of Western Australia 66:119-128. Howe F H & Westley L C 1988 Ecological Relationships of Plants and Animals. Oxford University Press, Oxford. Hume I D 1982 Digestive Physiology and Nutrition of Marsupials. Cambridge University Press, Cambridge. Jain K K 1976 Hydrogen peroxide and acetic acid for preparing epidermal peels from conifer leaves. Stain Technology 51:202-204. Krueger W C, Laycock W A & Price D A 1974 Relationships of taste, smell, sight and touch to forage selection. Journal of Range Management 27:258-262. Norbury G L & Sanson G D 1992 Problems with measuring diet selection of terrestrial mammalian herbivores. Australian Journal of Ecology 17:1-7. Poole W E 1983 Western grey kangaroo. In: The Australian Museum Complete Book of Australian Mammals (ed R Strahan). Angus and Robertson, Sydney, 248-249. Priddel D 1986 The diurnal and seasonal patterns of grazing of the red kangaroo. Macropus rufus, and the western grey kanga¬ roo, M. fuliginosus. Australian Wildlife Research 13:113-120. Shepherd K A, Wardell-Johnson G W, Loneragan W A & Bell D T 1997 Dietary of herbivorous marsupials in a Eucalyptus marginata forest and their impact on the understorey vegetation. Journal of the Royal Society of Western Australia 80:47-54. Storr G M 1961 Microscopic analysis of faeces, a technique for ascertaining the diet of herbivorous mammals. Australian Journal of Biological Sciences 14:157-164. 62 Journal of the Royal Society of Western Australia, 80:63-72, 1997 Waychinicup Estuary, Western Australia: the influence of freshwater inputs on the benthic flora and fauna J Phillips & P Lavery Department of Environmental Management, Edith Cowan University, Joondalup WA 6027: email p.lavery@cowan.edu.au Received October 1996 , accepted March 1997 Abstract Waychinicup Estuary (south-western Australia) is dominated by tidal and swell exchange with the ocean and receives seasonal, but low, freshwater inflows. The estuary was surveyed in September 1995 to determine the significance of freshwater inflows to its ecology and to provide a detailed inventory of benthic flora and fauna because, like other estuaries, Waychinicup may come under pressure for diversion of inflows to meet demand for potable water. Gradients detected in benthic flora and fauna, sediment grain size and organic content, and water quality, were related to distance from the river mouth. Sediments closest to the river mouth had a higher organic content and a higher percentage of <63 pm particles. These areas had a low water clarity (SDD = 0.6m). Further from the river, the sediments became less organic and coarser, and water clarity increased (SDD >7 m). We hypothesise that biotic gradients in the estuary are determined by freshwater inflow, principally through its contribution of fine and organic sediments, and less so through its effect on salinity. These sediments are prone to resuspension, reducing water clarity, limiting plant growth and providing a niche for euryhaline, low light-adapted macroalgae and a polychaete-dominated infauna in the upper estuary. Change in sediment type, increase in water clarity and increasing salinity produces a shift to a diverse macroalgal and seagrass community, and an amphipod-dominated benthic fauna in the lower estuary. Despite the dominance of oceanic exchange in the estuary, freshwater inflow is a key determi¬ nant of the composition and distribution of biota. Any significant reduction in freshwater inflow is likely to cause a change in water quality and biota; the system would have lower organic inputs, become clearer, and the plant and animal assemblages currently found in the upper reaches of the estuary may well be lost. Introduction The south coast of Western Australia has a diverse array of estuaries and coastal lagoons, ranging from those permanently open to the ocean through to some which are permanently closed. Cumulatively, as well as indvidually, these systems have enormous ecological, sci¬ entific and educational value. Like many ecosystems, however, these estuaries and coastal lagoons face potential threats. Among the most serious of these is the increasing demand for potable water supplies that accompanies rec¬ reational activities and tourism developments adjacent to estuaries. One means of meeting this demand is damming of the small rivers which feed the estuaries, and most at risk are those estuaries with relatively pristine catchments and reliable flows of freshwater. Estuaries are characterised by highly variable ecological conditions due to the mixing of marine and fresh water. The stress this places on estuarine biota is a major deter¬ minant of the biotic community structure and leads to gradients in both floral species diversity (Doty & Newhouse 1954; Munda 1978; Josselyn & West 1985; Montague & Ley 1993) and faunal species diversity (Jones et al. 1986; Montagna & Kalke 1992) related to salinity gradients. Changes to the freshwater inflow regime may alter the complex biophysical relationships and are © Royal Society of Western Australia 1997 likely to have significant implications for estuarine biotic distribution. While Waychinicup Estuary is not at all typical of other estuaries in the region, it does typify those estuaries most at risk of disturbance due to the future demand for freshwater. It is located close to significant demands for freshwater (the major urban centre of Albany is only 44 km to the west while nearby Cheyne Beach and the Two Peoples Bay Nature Reserve are increasingly used by tourists who require freshwater). Coupled with this, Waychinicup Estuary has a relatively reliable streamflow of potable water from a largely undisturbed catchment. The role of freshwater inflow into Waychinicup Estuary was examined in relation to the benthic flora and fauna, and water quality parameters. The estuary itself is only about 1300 m in length and over half of this length is less than 2 metres deep. The system is well flushed by tidal and swell action and riverine input is small in compari¬ son to the degree of oceanic exchange. Thus, it might be easy to dismiss the significance of relatively small inputs of freshwater as a determinant of biotic composition. However, we show that freshwater inputs are significant in this regard, but not through the simple effect of salinity variation. A further objective of the study was to provide a baseline inventory of benthic biota in the estuary. The highly marine nature of this estuary also suggests that 63 Journal of the Royal Society of Western Australia, 80(2), October 1997 the benthic flora within the estuary may serve as a useful indicator of benthic flora in adjacent marine areas. Despite increasing threats to the estuaries of the south-west, there is a noticeable absence of comprehensive baseline data, apart from the preliminary inventories by Hodgkin & Clark (1988a,b, 1989a, b, 1990a,b). Methods Study area Waychinicup estuary is located on the southern coast of Western Australia (34° 54’ S 118° 19' E; Fig 1), approxi¬ mately 44 km east of Albany. The area has warm summers and cool winters, and receives approximately 750 mm rainfall per annum. The estuary is unique in the region. It is the only estuary with steep rock shores along most of its 1300 m length. Rocky headlands maintain a wide opening to the ocean, allowing relatively undampened influences of oceanic swells and tides and maintaining permanent exchange with the ocean. The Waychinicup River, the estuary's only riverine source, has a catchment of 145 km2 of which 41% has been cleared (Hodgkin & Clark 1990b). The lower valley of the Waychinicup River and the Waychinicup Estuary are both within the Waychinicup National Park. Data were collected on two visits to the estuary, on 2 September and 18 September 1995. Biotic and physico¬ chemical variables were examined on each occasion. River flow Total monthly stream records for the period 1970-1995 were obtained for the Waychinicup River Gauging Weir (# 602031) from the Water and Rivers Commission of Western Australia. These data were analysed to provide total monthly, winter and annual stream flow volumes and long-term average winter and annual flow volumes. Macroalgae and seagrass Nine transects were established at intervals along the length of the estuary as well as individual quadrat samples at the head of the estuary where the narrowness of the estuary did not allow transects (Fig 1A). The cover and composition of macroalgae and seagrass were recorded in quadrats (0.25 m2) at 10 m intervals along the transect. Where transects occurred on areas of rocky granite and steep sides, quadrats were also taken at depth intervals. Data on seagrass distribution was supplemented by spot dives. All samples were identified to the lowest taxo¬ nomic level possible using keys of Huisman & Walker (1990) and Womersley (1984, 1987, 1994). Patterns in macroflora assemblages were examined using the multivariate statistical analysis package PATN (Belbin 1993). All analyses were based on a presence/ absence data matrix recorded for each quadrat using non-hierarchical allocation (Belbin 1987). Allocation was conducted following a Bray & Curtis association measure. This was selected as a dissimilarity measure since it is the most accepted measure used for ecological data (Faith et al. 1987). A radius of 0.6 was selected to define groupings in the allocation. Benthic invertebrates and infauna The presence and abundance of benthic invertebrate fauna was determined along three transects (Fig 1A). Figure 1. Sampling site locations for (A) benthic flora and macroinvertebrates and (B) sediments and water quality profiles in Waychinicup Estuary. Three replicate core samples (120 mm diameter and 10 cm deep) were randomly taken from within the first, central and last 0.25 m of the transect. Core samples were sieved (1.0 mm) on location and preserved with buffered seawater formalin solution. Material retained on the sieve was flooded in a tray to sort for invertebrates. One observer spent 15 minutes picking out any visible invertebrates from each sample. Then, a second observer spent five minutes on the same sample to check for any overlooked organisms. This standardised processing was intended to reduce observer bias and to impose a common effort on each sample. Invertebrates were identified to the lowest taxonomic level possible using Brusca & Brusca (1990), Wells & Bryce (1984, 1988), Hutchings (1984) and Shepherd & Thomas (1982, 1984). PATN was used to explore patterns in infauna assem¬ blages, based on a species abundance data matrix recorded for each replicate (core). Classification of the replicates following the Bray & Curtis association was conducted by flexible UPGMA, using a beta value of zero. Species groupings were determined by a two-step association measure prior to a flexible UPGMA. Quadrats were also ordinated in two dimensions following Bray & Curtis association, using non-metric multidimensional scaling. Sediment sampling Sediment grain size and organic content were sampled at three sites on a longitudinal transect in the centre of 64 Journal of the Royal Society of Western Australia, 80(2), October 1997 the estuary (Fig IB). At each site, three replicate sedi¬ ment cores were collected to a depth of 2 cm using 50 mm perspex corers. Sediments were weighed then dried for 82 hours at 75 °C. Organic content was determined by combusting samples for two hours at 550 °C. Samples were then re-wet with deionised water and allowed to stand for two hours before wet-sieving through a 63 pm sieve. Water and sediment passing through the sieve were transferred into 10 mm centrifuge tubes and centrifuged for 15 minutes at 1800 rpm. The supernatant was removed and the remain¬ ing pellet of sediment was placed into a crucible and dried at 75 °C to determine the dry weight of the <63 pm fraction. A subsample of each <63 pm sediment sample was combusted for one hour at 950 °C to determine the calcium carbonate (CaCCX) content. Complete burn-off of CaC03 was verified by calculating the reduction in mass of CaC03 standards combusted with the samples. The CaC03 mass was used as an indication of fine particulate origin, with low content being interpreted as indicating a greater likelihood of terrigenous origin. The mean proportion of organic matter and <63 pm fraction recorded for each site was compared by analysis of variance (ANOVA). Separate one-way ANOVAs were conducted for each, using arcsine-transformed data, fol¬ lowed by Fisher's PLSD pairwise comparisons (Zar 1984). Water quality Horizontal and vertical salinity profiles were recorded along the length and width of the estuary (Fig IB), using a Yeo-kal Hamon Salinity-Temperature Bridge (MKI1 Model 602). Surface and bottom readings for temperature and dissolved oxygen (DO) were taken using a Yeo-kal Dissolved Oxygen/Temperature Model 630 meter. Light penetration was recorded using a Secchi disc. Surface and bottom water samples were collected during each visit to the estuary, at three sites (Fig lb). Bottom water samples were taken using a vanDorn bottle. Fil¬ tered (GFC 0.45 pm) samples were analysed for nitrate and nitrite, ammonia and orthophosphate using a Hach DR 2000 spectrophotometer, as per the cadmium reduc¬ tion, salicylate, and ascorbic acid methods respectively (Anon 1989). Results Stream flow Monthly, winter and annual streamflow for Waychinicup River are shown in Fig 2. Monthly flow was highly variable with the majority of flow occurring in the winter (June-September) period. Average annual flow (1970-1995) was 9.15 106 m3, while the average winter flow over the same period was 6.02 106 m3. The recorded winter flow in 1995 was 1.88 106 m3, or only 31 % of the long-term average winter flow volume and the third lowest flow on record. However, interannual variation was high and the long-term average was greatly enhanced by the high flows in 1978 and 1979. Eleven years (44 % of re¬ corded winters) had flows of less than 3 106 m3 indicating that low flow years are not rare. The following results are probably indicative of the lower flow years. 1970 1975 1980 1985 1990 1995 E ! £ (c) Winter Row (June-Sept.) ..iiiJilllL 1970 1975 1980 1985 1990 1995 Figure 2. River flows for the Waychinicup River, 1970-1995. Dashed lines indicate the 1970-1995 average. Physico-chemical variation All of the measured variables showed detectable changes clearly associated with increasing distance from the river mouth. There were both vertical and horizontal variations in salinity through the estuary with surface salinities con¬ sistently lower than bottom salinities (Fig 3). In September, stratification was greatest at the river mouth and least at 350 m from the river mouth where water depth was only 1.25 m. Surface salinity varied from 21.6 ppt at the head of the estuary to 33.9 ppt at the seaward mouth (Fig 3). Bottom salinity readings showed a smaller range of variation, from 27.4 ppt at the head of the estuary to 35.6 ppt in the lower reaches. Surface temperatures ranged from from 14.7 °C at near the river mouth to 16 °C at 810 m and 1070 m from the river mouth respectively (Fig 3). Similarly, bottom temperature readings increased from 14.9 °C at site 1 to 16.2 °C at site 13. Temperature stratification was evident throughout the estuary, though only marginally so at 400 m from the river mouth. Oxygen concentrations were consistently high throughout the estuary (Table 1; corresponding saturation values ranged between 94 - 102 %) with only minor spatial differences. Variations in the concentrations of dissolved inorganic nitrogen (ammonia, nitrate+nitrite) and phosphorus (orthophosphate) were similar on both sampling dates, and so only data for 18/9/95 are shown in Table 1. Only orthophosphate showed a clear spatial variation, with very high concentrations near the river mouth and low 65 Journal of the Royal Society of Western Australia, 80(2), October 1997 concentrations elsewhere. The concentration of inorganic nitrogen varied negligibly through the estuary. There was noticeable stratification of orthophosphate in the upper reaches of the estuary (site 1) with the upper, less saline layer having almost four times the concentrations of the bottom waters. Corresponding to the variation in ortho¬ phosphate concentrations was a gradient in N:P ratios of the water column. The upper and middle estuary had ratios less than 16:1 and so could be classified as poten¬ tially N limited waters. It must be emphasised that these values are for a single sampling occasion at a time of low flow. Sediments and light attenuation As with the water column parameters, the sediment characteristics showed a clear gradient corresponding to distance from the river mouth. Mean organic matter con¬ tent decreased with increasing distance from the river mouth (Table 2). The variation between sites was signifi¬ cant (ANOVA, P < 0.05; data arc-sine transformed). Site 1 had a significantly higher percentage organic matter than site 2 (P = 0.02, Fisher's PLSD) and site 3 (P < 0.01), but sites 2 and 3 were not significantly different. The greatest variation was observed at site 2, which sup¬ ported a seagrass meadow. Salinity Table 2 Mean (± standard deviation, n = 3) of sediment characteristics, Secchi disc depth and water depth in Waychinicup Estuary. Sites 1, 2 and 3 were 125 m, 550 m and 1070 m from the river mouth respectively. Sediment data are the transformed values derived from percentage by weight data. Parameter Site 1 Site 2 Site 3 Organic matter (%) 13.4 ± 0.8 9.3 ± 2.5 7.8 ± 0.6 Grain size <63 pm (%) 18.0 ± 0.7 14.8 ± 2.9 12.1 ± 0.4 Secchi disc depth (m) 0.65 1.4 7.5 Water depth (m) 1.6 2.0 9.2 Temperature Figure 3. Water column salinity (top) and temperature (bottom) for surface and bottom water in Waychincup Estuary, Septem¬ ber 1995. Table 1 Nutrient concentrations in Waychinicup Estuary on 18 Septem¬ ber 1995. N:P ratio is for dissolved inorganic ions. NUTRIENT Site 1 Site 2 Site 3 P04-P (mg L-1) Surface 0.11 0.00 0.01 Bottom 0.03 0.02 0.02 N03 -N (mg L1) Surface 0.07 0.08 0.07 Bottom 0.03 0.03 0.03 NH3 -N (mg L'1) Surface 0.00 0.00 0.01 Bottom 0.00 0.00 0.00 N:P Ratio Surface 0.6 16 8 Bottom 1.0 1.5 1.5 02 (mg L'1) Surface 8.2 7.0 8.2 Bottom 8.7 7.3 8.3 The above trends were also reflected in fine particle fraction (Table 2). Site 1, closest to the river mouth, had the highest fine particle fraction. The variation between sites was statistically significant (ANOVA, P < 0.05). Fisher's PLSD indicated significant differences between sites 1 and 3 (P < 0.01). The greatest variation was again observed at site 2, which supported a benthic seagrass community. Secchi disc depths increased with increasing distance from the mouth of the river (Table 2) thus varying posi¬ tively with temperature and sediment organic content and fine particulate fraction. Secchi disc depth in the upper estuary was 0.65 m or 40% of the water column depth. In contrast SDD was over 7.50 m or 80% of the water depth at the lower site. Benthic flora There were clear spatial patterns in the benthic plant assemblages through the estuary which again coincided with increasing distance from the river mouth and the gradient in physico-chemical parameters. Forty species of macroalgae and five species of seagrasses were recorded in the estuary (Table 3). Four¬ teen groups of benthic flora were identified, falling into two broad categories; those on the estuary floor or those on the vertical rock walls. The distribution of 'estuary floor' groups showed a strong gradient of species change in the upper reaches of the estuary (Fig 4). Fine green and red algae were limited to the upper 80 m of the estuary, whilst a mixed algal assemblage dominated by Cystoseira trinodis was restricted to an area of hard sub¬ stratum between 150 m and 300 m from the river mouth. Elsewhere, the substratum was muddy to sandy and sup¬ ported seagrasses (Fig 4). The seagrasses Posidonia australis and Heterozostera tasmanica were widespread although their distribution was generally limited to areas further 66 Journal of the Royal Society of Western Australia, 80(2), October 1997 Chaetomorpha aurea & filamentous red algae Filamentous red algae Mixed algae dominated by Cystoseira trinodis Heterozostera tasmanica Posidonia australis Halophila australis Amphibolis antarctica Mixed P. australis and H. tasmanica Posidonia sinuosa Mixed algal assemblage with vertical zonation (see inset figure) Mixed algae dominated by Cystophora brownii and C. retorta Figure 4. Distribution of benthic flora assemblages in Waychinicup Estuary, September 1995. than 200 m from the river mouth. Posidonia sinuosa, a seagrass that is usually found in higher energy environ¬ ments, was found only near the mouth of the estuary. Three other species, Halophila australis, Heterozostera tasmanica and Amphibolis antarctica occurred in small patches throughout the estuary. The vertical rock faces in the lower reaches of the estuary supported a mixed algal assemblage dominated by brown and red algae. Allocation analysis indicated vertical zonation (Fig 4). Benthic invertebrate and infauna assemblages The two-dimensional ordination from pooled replicate data showed patterns in sites relating to location within the estuary (Fig 5). There were three clear groups, with all sites for each of the the upper, middle and lower estuary transects clustering together. The dendrogram from clas¬ sification of all replicate data from the nine sites also reflected this pattern with the upper estuary sites (1, 2 & 3; 150 m from the river mouth) showing a high dissimilar¬ ity to other sites. A cut-off value of 0.96 (a very high degree of dissimilarity) produced five groups of sites with the three upper estuary sites clearly separating out as one group (Fig 5). There was a shift from small crustacean dominated sites at the mouth of the estuary to sites dominated by polychaete worms in the upper reaches of the estuary (Table 4). The pattern of species shift shown by the clas¬ sification and ordination analyses (Fig 5) corresponded well to the dominant sediment type at each site (Table 5). 67 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 3 Benthic macrophyte species recorded in Waychinicup Estuary (September 1995) and their approximate distances from the river mouth. * indicates presence. metres from river mouth 5 30 60 80 100 120 160 210 270 360 510 610 650 810 1070 Chaetomorpha aurea * * * filamentous red sp 1 * * * Heterozostera tasmanica * * * * * * Cystoseira trinodis * * * * * * * filamentous red sp 2 * * * * * Sargassum spinuligerum * * * * * * Posidotiia australis * * * * * * * * * Colpomenia sinuosa * * * Laurencia sp 1 * * * * Phloiocaulon spectabile * * Cystophora brownii * * * * Cystophora retorta * * * * encrusting red coralline sp 1 * * * encrusting red coralline sp 2 * * * * Hormosira banksii * * Posidonia sinuosa * Amphibolis antarctica ♦ Dictyopteris sp * Chondria sp X- Laurencia sp 2 * Metagoniolithon radiatum * * Halophila australis * * Ecklonia radiata * * Lobophora variegata * * Sargassum distichum * * Haliptilon roseurn * * Sargassum sp 1 * Dictyosphaeria sericea * Cystophora sp * Caulerpa distichophylla * Amphiroa anceps * Halimeda cuneata * Phacelocarpus adopus * Metamastophora sp * Zonaria tumeriana * filamentous red sp 3 * Dictyota dichotoma * Hypnea sp * Sargassum sp 2 * Sargassum tristichum * Cystophora monilifera * Jania micrarthrodia * Rhipiliopsis peltata * Lenormandia spectabilis * Stenogramme leptophylla * Discussion The unique physical and hydrologic features of Waychinicup Estuary, which have been described above, are clearly reflected in its benthic flora. With over forty five species, Waychinicup Estuary has an exceptionally rich benthic flora. By comparison, the vast majority of other south coast estuaries are dominated by three species of benthic macrophyte, Ruppia megacarpa, Polyphysa penicitlus and Lamprothamnion papillosum with its associated epiphytes ( e.g . Hodgkin & Clarke 1989a,b, 1990). The benthic flora of Waychinicup Estuary is predomi¬ nantly a marine intrusion flora. The seagrasses in the system are typical of coastal marine areas not estuaries, and the macroalgal assemblages in the lower reaches are dominated by species typical of southern Australian ma¬ rine flora (Womersley 1984, 1987, 1994). In this respect, the algal composition of the lower reaches recorded in this winter survey is likely to be a better representation of open coastal marine flora of the south coast than it is of other estuaries. 68 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 4 Benthic macroinvertebrates recorded in Waychinicup Estuary (September 1995) and their approximate distances mouth. (See Fig 1A for site locations), p = Polychaetae, mx = Maxillopoda, m = Malacostrata, g = Gastropoda, s = o = Oligochaetae, b = Bivalva. 1 is < 250 nr2, 2 is < 500 nr2, 3 is < 800 nr2, and 4 is > 800 nr2. from the river = Sipunculada, Region: Site: 1 Lower 2 3 4 Middle 5 6 7 Upper 8 9 Lyssianassidae sp (m) 2 Spio pacifica (p) 4 4 1 1 1 maldanid sp 1 (p) 1 1 Odontosyllis sp (p) 1 1 1 1 1 Birubius sp (m) 1 1 1 Paphies sp (b) 1 1 1 1 1 Tipimegus sp (m) 1 1 1 Anthuridae sp (m) 1 mysid sp 1 (m) 1 Aricidea fauveli (p) 1 1 1 1 Phallodrilus zvellsi (o) 1 1 1 amphipod sp (m) 1 Austropheonoides sp (m) 1 Oedicerotidae sp (m) 1 Paradexamine sp (m) 1 ostracod sp 1 (mx) 1 Rhinothelepus setosus (p) 1 1 Guemea endota (m) 2 1 Golfingia sp (s) 1 2 1 1 Leonnates sp (p) 1 1 oligochaete sp (o) 1 Bittium sp 1 (g) 3 Bittium sp 2 (g) 1 Spisula trigonella (b) 1 1 1 Paratanais ignatus (m) 1 1 1 Apseudidae sp (m) 1 1 1 1 Caullierella sp (p) 2 1 1 Capitella cap it a ta (p) 1 2 4 maldanid sp 2 (p) 1 1 Notomastus estuarius (p) 1 Naineris grubei (p) 1 1 1 1 Leitoscoloplos bifurcatus (p) 1 1 Katyelsia rhytiphora (b) 1 Tethygeneia nalgo (m) 1 mysid sp 2 (m) 1 Total Species 19 23 12 Notwithstanding the above, our results clearly indicate that other factors are important in shaping the distribution patterns in the upper reaches of the estuary. Salinity may be an important determinant of macroalgal distribution in Waychinicup Estuary, but in association with turbidity. Only two algal assemblages penetrated to the very head of the estuary. This trend of reduced diversity in the upper reaches of estuaries is often ascribed to salinity effects (McLusky 1989; Morrisey 1995) though in this case it appears to be a combination of both salinity and light attenuation. The fine green alga in the upper reaches (Chaetomorpha) is typically adapted to variable salinity regimes and better adapted to poor light conditions than larger brown and red algae (Lavery 1989). Other macroalgal groups in the upper and mid-estuary reaches showed a strong gradient of species changes, which may be due to factors other than light penetration. Not only did benthic plant groups change along a gradient of in¬ creasing salinity, but vertical groups also changed with increasing salinity. Although vertical species shift would be partly depth-related those groups that occurred at the water line must be able to tolerate, in addition to wave action, a relatively wide range of salinity given the fresh¬ water lens that was evident at the time of this study. All of our results suggest that freshwater inflow is a significant determinant of biological gradients in Waychinicup Estuary. This is despite the relatively small volume of inflow, its seasonality, and the degree of ocean exchange which is large. Thus while it might be argued that this system is in fact a simple river mouth with granite headlands, the parameters measured in this study clearly suggest it is, at least seasonally, a stratified estuary. In drawing this conlcusion, it must be stressed that the estuary was surveyed over a limited period in one of the lowest flow years recorded. While the conditions in the estuary therefore represent a 'snapshot', they were probably a conservative estimate of the role freshwater inputs play in the system. It is probable that the role of freshwater in 69 Journal of the Royal Society of Western Australia, 80(2), October 1997 Table 5 Qualitative description of sediment characteristics in Waychinicup Estuary. Location Site Sediment type lower reaches 1-3 coarse sand /shell fragments middle reaches 4 medium to coarse sediments; some organic matter 5 fine to medium sediments; some organic matter 6 medium to coarse sediments; accumulation of dead shells upper reaches 7 fine mud and sands with fine organics and seagrass detritus 8 fine muds /silt 9 seagrass detritus over medium coarse sediment and shells driving biological gradients will be greater in higher flow years. There was a gradient in most variables that related to distance from the source of riverine inflow, indicating that during these flow conditions freshwater inflows are affecting several variables, not just salinity. In particular, freshwater inflow appears to influence sediment distribution with subsequent effects on water clarity and sediment characteristics which directly affect benthic floral and faunal distributions. Three possible sources of fine sediment within Waychinicup Estuary are the inflowing river, the coast, and from the margins of the estuarine basin itself (Carter, 1988). The margins were considered to contribute the smallest proportion of fine sediment, however, due to the large area of granite rock bordering the estuary. Fine sediments of marine origin in the form of calcium car¬ bonate (from organisms such as foraminiferans and the weathering of shell material) were eliminated from samples taken from all three sediment sampling sites, leaving only clays and silicas. Since organic matter had been removed, it can be concluded that the remaining fine sediments were predominantly introduced with freshwater inflow, and this appears to be supported by the gradient of decreasing proportion of fine particles with increasing distance from the river source. The gradient in sediment characteristics in Waychinicup Estuary provides a range of microhabitats for benthic invertebrates. Richness and abundance of benthic invertebrate species appeared to be related to dif¬ ferences in microhabitats and in particular sediment characteristics. Furthermore, invertebrate assemblages varied not only between transects but also within transects, suggesting that sediment characteristic gradi¬ ents also existed across the estuary, most likely the result of seagrass distribution modifying the longintudinal gradi¬ ent of sediment type. The salinity stratification was correlated with higher levels of nutrients in surface waters in the upper reaches of the estuary. High levels of phosphate were expected, given that the upper catchment of Waychinicup River is NMDS invertebrate ordination; 0.3100 0.4460 0.5820 0.7180 0.8540 0.9900 I l l i I i Group 1 Group 2 Group 3 Group 4 Group 5 " si repl si rep3 si rep2 s2 repl s2 rep2 s2 rep3 s3 rep2 s3 repl - s3 rep3 - s4 repl s4 rep3 s4 rep2 s5 rep3 s6 rep2 s7 rep2 s5 repl L s6 repl - s7 repl s8 repl s8 rep2 s9 rep2 s7 rep3 s8 rep3 - s9 rep3 Cs6 rep3 s9 repl - s5 rep2 1 n i ' i i i 0.3100 0.4460 0.5820 0.7180 0.8540 0.9900 Figure 5. Ordination and associated dendrogram of sites and replicates based on the presence/absence of benthic macroinvertebrate fauna. See Fig 2 for site locations (si rep 1 = site 1 replicate number 1 etc). cleared for agriculture, and this was indeed the case at the time of this study. The dissolved phosphate concen¬ trations were comparable to those in highly eutrophic estuaries of south-western Australia (Lukatelich & McComb 1986). However, at the time of sampling a high degree of tidal exchange was evident as well as short term oscillations in the direction of flow due to incoming swells through the estuary mouth. Together with the degree of stratification, these exchange will significantly reduce nutrient accumulation, and the associated effects, within the estuary. Management implications It is clear that freshwater inflow into Waychinicup Estuary has a significant influence on the nature of the estuary. The results of the winter survey clearly emphasise the uniqueness of this estuary, and its ecological 70 Journal of the Royal Society of Western Australia, 80:29-39, 1997 and conservation value. Within this very small inlet/ estuary there is a clear gradation from an 'estuarine' to open marine flora which is not recorded for any other system along the south coast. Any diversion of freshwater inflows to this system would therefore impact its ecol¬ ogy. It can be easy to overlook the importance of a small, seasonal freshwater input to an inlet which is so obviously dominated by tidal and oceanic swell exchange. Yet, any significant reduction in freshwater flow would reduce particulate loads to the estuary and, subsequently, sediment characteristics and the distribution and composition of benthic biota. While seagr asses might extend further up the estuary, the coarser sediments and lower turbidity would produce a more homogenous flora and possibly exclude the benthic invertebrate assemblages associated with fine mud and silt. Finally, reduction in freshwater inflow may reduce the degree of stratification and enhance vertical mixing in the estuary. In conclusion, there are expected to be changes to the nature of the estuary in the event that freshwater is dammed or diverted, but how important such changes would be depends on how the estuary is valued. Given the estuary's location within Waychinicup National Park, it is assumed that it would be a management objective to preserve it for its intrinsic values, and so any alteration in the nature of the estuary can only be regarded as un¬ desirable. Acknowledgements: This study was funded by the Water Authority of Western Australia; we thank P Helsby of the Albany branch for his assis¬ tance. Special thanks go to G Kendrick, J Nielson and M Platell for their assistance with species identifications. Thanks also go to the two anony¬ mous referees who provided useful comments on the manuscript. References Anon 1989 Water Analysis Handbook. Hach Company, Loveland, Colorado, USA. Belbin L 1987 The use of non-hierarchical allocation methods for clustering large sets of data. Australian Computer Journal 19:32-41. Belbin L 1993 PATN Technical Reference. CSIRO Division of Wildlife and Ecology, Canberra. Brusca R C & Brusca G J 1990 Invertebrates. Sinauer Associates, Massachusetts. Carter R W G 1988 Coastal Environments. Academic Press, London. Doty M S & Newhouse J 1954 The distribution of marine algae into estuarine waters. American Journal of Botany 41 ROS¬ SIS. Faith D P, Minchin P R &Belbin L 1987 Compositional dissimi¬ larity as a robust measure of ecological distance. Vegetatio 69:57-68. Hodgkin E P & Clark R 1988a An inventory of information on the estuaries and coastal lagoons of South Western Australia. Nomalup and Walpole inlets and the estuaries of the Deep & Frankland Rivers. Environmental Protection Authority, WA. Estuarine Studies Series No 2. Hodgkin E P & Clark R 1988b An inventory of information on the estuaries and coastal lagoons of South Western Australia. Beaufort and Gordon Inlet. Estuaries of the Shire of Jerramungup. Environmental Protection Authority, WA. Estuarine Studies Series No 4. Hodgkin E P & Clark R 1989a An inventory of information on the estuaries and coastal lagoons of South Western Australia. Estuaries of the Shire of Esperence. Environmental Protec¬ tion Authority, WA. Estuarine Studies Series No 5. Hodgkin E P & Clark R 1989b An inventory of information on the estuaries and coastal lagoons of South Western Australia. Estuaries of the Shire of Manjimup. Environmental Protec¬ tion Authority, WA. Estuarine Studies Series No 6. Hodgkin E P & Clark R 1990a An inventory of information on the estuaries and coastal lagoons of South Western Australia. Estuaries of the Shire of Ravensthorpe and the Fitzgerald National Park. Environmental Protection Authority, WA. Estuarine Studies Series No 7. Hodgkin E P & Clark R 1990b An inventory of information on the estuaries and coastal lagoons of South Western Australia. Estuaries of the Shire of Albany. Environmental Protection Authority, WA. Estuarine Studies Series No 8. Huisman J M & Walker D I 1990 A catalogue of the marine plants of Rottnest Island, Western Australia, with notes on their distribution and biogeography. Kingia 14:349^159. Hutchings P 1984 An Illustrated Guide to the Estuarine Poly- chaete Worms of New South Wales. Coastal and Wetlands Society, Sydney. Jones A R, Watson-Russell C J & Murray A 1986 Spatial patterns in the macrobenthic communities of the Hawkesbury Estuary, New South Wales. Australian Journal of Marine and Fresh¬ water Research 37:521-543. Josselyn M N & West J A 1985 The distribution and temporal dynamics of the estuarine macroalgal community of San Francisco Bay. Hydrobiologia 129:139-152. Lavery P 1989 Factors controlling the abundance and diversity of macroalgal species in eutrophic estuaries. PhD Thesis. Dept of Botany, The University of Western Australia, Perth. Lukatelich & McComb 1986 Nutrient levels and the develop¬ ment of diatom and blue-green algal blooms in a shallow Australian estuary. Journal of Plankton Research 8:597-618. McLusky D S 1974 Ecology of Estuaries. Heinemann Educa¬ tional Books, London. Montagna P A & Kalke R D 1992 The effect of freshwater inflow on meiofaunal and macrofaunal populations in the Guadalupe and Nueces Estuaries, Texas. Estuaries 15:307- 326. Montague C L & Ley J A 1993 A possible effect of salinity fluc¬ tuation on abundance of benthic vegetation and associated fauna in northeastern Florida Bay. Estuaries 16:703-717. Morrisey D 1995 Estuaries. In: Coastal Marine Ecology of Temperate Australia (eds AJ Underwood & MG Chapman). University of New South Wales Press, Sydney, 152-170. Munda I M 1978 Salinity dependent distribution of benthic algae in estuarine areas of Icelandic fjords. Botanica Marina 21:451- 468. Shepard S A & Thomas I M 1982 Marine invertebrates of Southern Australia. Part I. South Australian Government Printer, Adelaide Shepard S A & Thomas I M 1982 Marine invertebrates of South¬ ern Australia. Part II. South Australian Government Printer, Adelaide Wells F & Bryce C W 1984 A Guide to the Common Molluscs of South-Western Australian Estuaries. Western Australian Museum, Perth. Wells F & Bryce C W 1988 Seashells of Western Australia. West¬ ern Australian Museum, Perth. Womersley H B S 1984 The Marine Benthic Flora of Southern Australia: Part I. South Australian Government Printer, Adelaide. Womersley H B S 1987 The Marine Benthic Flora of Southern Australia: Part II. South Australian Government Printer, Adelaide. Womersley H B S 1994 The Marine Benthic Flora of Southern Australia: Part III. Australian Biological Resources Study, Canberra. Zar J H 1984 Biostatistical Analysis. 2nd edition. Prentice-Hall International Editions, Englewood Cliffs, NJ. 71 Journal of the Royal Society of Western Australia, 80:73-77, 1997 1 Contributions of N H Speck to the biogeography of Proteaceae in Western Australia N Gibson1, G J Keighery1 & B J Keighery2 1 Science and Information Division, Department of Conservation and Land Management, PO Box 51 Wanneroo WA 6065; email neilg@calm.wa.gov.au and gregk@calm.wa.gov.au 2 Department of Environmental Protection, PO Box K822, Perth WA 6842; ema i / bronwen _keighery@envi ro n . wa .gov. a u Received January 1997; accepted May 1997 Abstract When considering the study of the biogeography of the Western Australian flora, botanists such as Diels, Gardner and Beard spring to mind, but what has not been generally recognised outside of Western Australia was the substantial contribution of Dr Nathaniel Speck. Speck was the first to quantify patterns of species richness for the Proteaceae in south-western Australia. In addition he provided detailed community mapping of the Swan Coastal Plain in particular, and of the south west in general. Speck was also the first to propose that the Mt Lesueur region (one of the two major centres of biodiversity in Western Australia) should be recognised as a separate botanical district. The biogeographic patterns of the Proteaceae reported by Speck are compared here to a recent synthesis based on over 21000 collections held in the Western Australian Herbarium. The current analysis was largely consistent with the Speck's 1958 work, except that the gradients in species richness away from the two major centres of species richness are less steep than previously described. This strong correlation between the two studies is despite a taxonomy that recognises more than twice the number of taxa that he dealt with and availability of many more collections across the south west. A brief outline of Speck's botanical and ecological work and a bibliography of his published and unpublished work are included. Introduction The diversity and richness of the Western Australian flora have long been recognised (Diels 1906; Gardner 1944; Hopper 1979), as has the biogeographical pattern¬ ing of the flora (Diels 1906; Speck 1958; Beard 1980, 1990). While the names of Diels, Gardner and Beard are readily recognised, the contribution of Nathaniel Henry Speck has often been overlooked outside of Western Australia. Speck mapped the vegetation pattern on the Swan Coastal Plain in the early 1950s and he later extended this across the whole of the southwest (Speck 1958). Based on this work he proposed that the Lesueur region should be recognised as a separate botanical district. Beard's (1980, 1990) phytogeographic mapping did not accept this division, but recent work on the biogeography of the genus Banksia has strongly argued for the resurrection of this district (Lamont & Connell 1996). The detailed biogeographic patterning of many of the major genera and families is now well documented across the south-west of Western Australia (Speck 1958; Hopper & Maslin 1978; Hopper 1979; Taylor & Hopper 1988; George 1991; Keighery 1996; Lamont & Connell 1996). Speck's (1958) analysis of the biogeography of the Proteaceae was the first study to quantify these patterns. It is now almost 40 years since this ground-breaking work was published, and here we compare the results of his 1958 study with our present understanding of bio- © Royal Society of Western Australia 1997 geography of the Proteaceae based on collections held in Western Australian Herbarium. Methods Speck (1958) mapped 426 taxa of the Proteaceae across the southwest on a 50000 yard (28 mile) grid, and distribu¬ tion patterns were analysed on both a genus and family basis. A comparable recent analysis was compiled from distribution data for 882 taxa (in 17 genera and 779 species) based on collections held in the Western Australian Her¬ barium. Over 24000 collections of Proteaceae are held in the herbarium and in excess of 21000 have geographic coordinates that could be used in this analysis. Distribution patterns were determined on a one degree latitude and longitude grid basis across the State. This grid was both larger and more extensive than that used by Speck (1958). Analyses were undertaken for all taxa and for Grevillea, Lambertia and Adenanthos. These were compared with Speck's (1958) earlier analysis. Results and Discussion Comparison of biogeographic patterns Speck's (1958) analysis of the 426 taxa of Proteaceae showed two centres of species richness, one on the northern sandplain centred on the Mt Lesueur area and a second on the south coast stretching from the Stirling Range to the Fitzgerald River area (Fig 1). 73 Journal of the Royal Society of Western Australia, 80(2), October 1997 Figure 1. Isoflor maps from Speck (1958) based on a 50000 yard grid; A) Lambertia, B) Adenanthos, C) Grevillea and D) Proteaceae. Some of the small genera such as Adenanthos and Lambertia exhibited only one main centre of species rich¬ ness on the south coast while the larger genera such as Grevillea exhibited a bimodal pattern of species richness, as did the family as a whole (Fig 1). At both the genus and family level, a feature of the pattern that Speck (1958) showed was the very steep nature of the isoflor gradients away from the centres of species richness. The more recent analysis of 21000 records held in the Western Australian Herbarium (Fig 2) confirms the general pattern described by Speck (1958). The main difference between the two analyses is the estimate of number of taxa at the centres of species richness (180-194 taxa per grid cell in the current analysis compared with 110-132 taxa per grid cell in Speck's analysis) and the more gradual nature of the species richness gradients now seen. These differences are accounted for by a taxonomy that now recognises more than twice the number of entities recognised in 1958, and the huge increase in collections and access across the south-west that have become avail¬ able over the last 40 years. The state-wide analysis also shows a small but significant concentration in the Kimberley region, an area not covered by Speck's earlier analysis (Fig 2). Both analyses show evidence of sample bias. In Speck's (1958) analysis there is a large bulge along Great Eastern Highway that runs from Perth to Kalgoorlie, indi¬ cating the poor access north and south off this transport corridor in 1958. The bias in the recent analysis is more subtle, but the small peaks in the arid zone reflect either range systems or major conservation reserves. As a result, it is not possible to unequivocally attribute these small peaks to particular land forms. Speck (1958) concluded that his isoflor maps sup¬ ported the theory that the two major centres of high species richness are the centres of origin of the Proteaceae in the southwestern flora. His prediction that the genera of the Myrtaceae and Epacridaceae would reveal essentially similar patterns have been borne out (George 1991; Keighery 1996). Speck (1958) further argued that the high degree of endemism and species richness indicate that these centres of origin were of greater age than if they 74 Journal of the Royal Society of Western Australia, 80(2), October 1997 Lambertia Adenanthos Grevillea ■9-11 ■7- 9 □5- 7 03- 5 □ l- 3 Proteaceae ■so -60 ■40 -50 ■30 -40 □20 -30 □ 10 -20 □ 1 10 Figure 2. Isoflor maps from a recent analysis of 21000 collections held in the Western Australian Herbarium using one degree grid; A) Lambertia, B) Adenanthos, C) Grevillea and D) Proteaceae. 75 Journal of the Royal Society of Western Australia, 80(2), October 1997 were only refugia and centres of redistribution following an ameliorating climate from the Last Glacial. Hopper's (1979) review of species richness across the southwest suggested that the transitional rainfall zone (in which the two centres of species richness are found) have had an extremely long history of changing climate and the species diversity of this region is a function of the both the stability of the landscape and the repeated pattern of sweeping climatic change, echoing Speck's conclusions of some 20 years previously. In addition to his analysis of distribution patterns in the Proteaceae, Speck also undertook structural mapping of the vegetation communities across the southwest. He recognised 62 community types in 26 vegetation systems. Using these 26 vegetation systems he proposed modifica¬ tions to the phytogeographic districts of Diels (1906). These modifications included the eastward movement of the Coolgardie and Austin boundary and the inclusion of a separate Lesueur district. Speck split his southern node of species richness between the Stirling district (which includes the Stirling Range) and the Eyre district (which includes the Fitzgerald River area). Beard's (1980, 1990) later mapping did not recognise the Lesueur district and incorporated all of the southern node of species rich¬ ness (stretching from the Stirling Range to Fitzgerald River) into his Eyre district. In a recent analysis of the pattern of biogeography of the genus Banksia, Lamont & Connell (1996) argue for the recognition of the Lesueur district, a south Stirling dis¬ trict and a west Eyre district corresponding to centres of Banksia species richness and similarity. Our current analysis was undertaken at a coarser scale than that of Lamont & Connell (1996), and only considered species richness; nevertheless, the northern node of species rich¬ ness seen in our data (Fig 2) is centred on their Lesueur node (Fig 6b in Lamont & Connell 1996). At the one degree grid scale used in our analysis, no subdivision of the southern node of species richness is evident (Fig 2). While phytogeographic boundaries are not based on species rich¬ ness criteria our data lends some support to a reappraisal of the status of the Lesueur phytogeographic district. A synopsis of Nathaniel Speck's botanical career Nathaniel Henry Speck was born in South Australia on 6 December 1906 and died in Canberra in 1970. Speck was one of those rare people who built two successful careers during his life time. The first as a science teacher then secondly as a botanist and ecologist. He began his second career with a BA degree majoring in botany which he started in 1942 at the age of 35. He graduated from the University of Western Australia in 1948. Between 1949 and 1952 he worked on his masters dissertation in which he described and mapped the major plant commu¬ nities of the Swan Coastal Plain around Perth. At this time he was also responsible for the establishment of the Herbarium at the Botany Department of the University of Western Australia, which played a major role in the development of the illustrated key to the Western Aus¬ tralian flora produced by Blackall and Grieve (Grieve 1953). Speck was awarded his MSc in 1952. He immedi¬ ately enrolled as a PhD candidate and in 1953 was awarded a Senior Research Fellowship. His PhD topic was very ambitious; he attempted to describe and map the vegetation communities of south-western Australia from Shark Bay to Esperance, and as an aside he under¬ took a biogeographical analysis of the family Proteaceae across this region. During this time he continued to assist Professor Grieve with preparation of the illustrated key to Western Australian flora, in redrawing and redrafting some of Blackall's earlier illustrations, and by taking many colour photographs for illustration of those volumes (Grieve 1953). In July 1953, Speck applied for a position as a Technical Officer with CSIRO's Land Research and Regional Survey Section. It appears that CSIRO was so impressed with his application that they readvertised the position at a Research Officer level and Speck was offered this position, which he subsequently accepted in November 1953 at the age of 47. He and his family moved to Canberra where he com¬ menced work as a Research Officer Grade 3 - Ecologist on 5 April 1954. Speck immediately fitted into the survey section and undertook field work in the Gilbert-Leichardt area, around Wiluna, Western Australia and in the west Kimberley. He was made permanent in November 1954, and reclassified to Senior Research Officer in 1957. Reports of his ability and performance in CSIRO over this period are uniformly glowing. Speck was awarded his PhD in 1959 from the Univer¬ sity of Western Australia. His thesis was in two parts. Volume 1 described and mapped 62 plant associations in 26 vegetation systems across the southwest of Western Australia. The second volume mapped the distribution of 426 species of Proteaceae across the same area on a 50000 yard grid, and for the first time quantified the patterns of diversity of this family. It is unfortunate that Speck never formally published any of this work and this probably cost him the national and international recognition that he certainly deserved. Speck was by that time 53 years old. His work commitments in a senior position in CSIRO protracted the completion of his thesis which he initially expected to complete it in two or three years before his appointment to CSIRO. It in fact took him over six years to complete. In the last few years of his professional life, Speck undertook consulting work as a survey ecologist for the United Nations Food and Agriculture Organisation and the UN Development Program in Argentina while on un¬ paid leave from CSIRO. He retired from CSIRO at the age of 60 on 2 February 1966. In his second career Speck made a major contribution to the fields of botany, ecology and biogeography in Aus¬ tralia. He made over 8500 collections of which over 1800 are lodged in the Wesrem Australian Herbarium. His published ecological work covered areas as diverse as the Swan Coastal Plain to the north Kimberley to Queensland (Appendix 1). He collected four type specimens ( Eremophila congest a ms, Thysanotus speckii, Conostylis crassinervia , Grevillea makinsonii) of which the Thysanotus bears his name. He had a life long interest in botany and ecology, collecting and drawing plants wherever he lived. His type collection of Grevillea makinsonii was made 76 Journal of the Royal Society of Western Australia, 80(2), October 1997 the year before he died. With his passing in 1970 at the age of 63, Australia lost one our most significant but unrecognised field ecologists and plant biogeographers. In his second career. Speck made a very significant contribution to the knowledge of plant biogeography in Western Australia, so much so that many of his insights are now considered common knowledge. Perhaps there is no greater accolade. Acknowledgments: We would like to thank librarians L Wright and B Waugh for their help in locating material related to Dr Speck, Dr Speck's family for permission to access his University file and M Lyons for draft¬ ing the figures. This paper is based on a presentation prepared for the Proteaceae Conference held in Melbourne in 1996. References Beard JS 1980 A new phytogeographic map for Western Australia. Western Australian Herbarium Research Notes 3:37-58. Beard ]S 1990 Plant life of Western Australia. Kangaroo Press, Kenthurst. Diels L 1906 Die Pflanzenwelt von West-Australien stidlich des Wendekreises. Die Vegetation der Erde. Vol 7. Englemann, Leipzig. Gardner CA 1944 The vegetation of Western Australia. Presi¬ dential Address. Journal of the Royal Society of Western Australia 28: xi-lxxxvii. George AS 1991 New taxa, combinations and typification in Verticordia (Myrtaceae: Chamelaucieae). Nuystia 7:231-393. Grieve BJ 1953 Editorial preface. In: How to know Western Aus¬ tralian wildflowers (Blackall WE 1959). University of Western Australia Press, Perth. Hopper SD & Maslin BR 1978 Phytography of Acacia in Western Australia. Australian Journal of Botany 26:63-78. Hopper SD 1979 Biogeographical aspects of speciation in the south-western Australian flora. Annual Review of Ecology and Systematics 10:399-422. Keighery GJ 1996 Phytogeography, biology and conservation of Western Australian Epacridaceae. Annals of Botany 77:347- 355. Lamont BB & Connell SW 1996 Biogeography of Banksia in southwestern Australia. Journal of Biogeography 23:295-309. Speck NH 1958 The vegetation of the Darling - Irwin botanical districts and an investigation of the family Proteaceae in south Western Australia. 2 vols. PhD Thesis. University of Western Australia, Perth. Taylor A & Hopper SD 1988 The banksia atlas. Australian Flora and Fauna Series No 8. Australian Government Service, Canberra. Appendix 1. Bibliography of Speck's research work (in chronological order) Speck NH 1952 Plant ecology of the metropolitan sector of the Swan Coastal Plain. MSc Thesis, University of Western Aus¬ tralia, Perth. Speck NH 1958 The vegetation of the Darling - Irwin botanical districts and an investigation of the family Proteaceae in south Western Australia. 2 vols. PhD Thesis, University of Western Australia, Perth. Speck NH 1960 Vegetation of the north Kimberley area, West¬ ern Australia, CSIRO Land Research Series 4:41-63. Mabbutt JA, Speck NH, Wright RL, Litchfield WH, Sofoulis J & Wilcox DG 1963 Land systems of the Wiluna-Meekatharra area. Land Research Series 7:24-72. Speck NH 1963 Vegetation of the Wiluna-Meekatharra area. Land Research Series 7:143-161. Wilcox DG & Speck NH 1963 Pastures and pasture lands of the Wiluna-Meekatharra area. Land Research Series 7:143-161. Speck NH 1964 Introduction and summary description of the West Kimberley area. Land Research Series 9:9-23. Speck NH & Lazarides M 1964 Vegetation and pasture of the West Kimberley area. Land Research Series 9:140-174. Speck NH, Fitzgerald K & Perry RA 1964 Pasture lands of the West Kimberley area. Land Research Series 9:175-191. Speck NH, Wright RL & Rutherford A 1964 Land systems of the West Kimberley area. Land Research Series 9:24-75. Speck NH 1965 Introduction. Land Research Series 13:7-10. Speck NH 1965 Pasture lands of the Tipperary area. Land Research Series 13:99-103. Speck NH 1965 Summary description of the Tipperary area. Land Research Series 13:11-17. Speck NH 1965 Vegetation and pastures of the Tipperary area. Land Research Series 13:81-98. Speck NH, Wright R L & van de Graaff RHM 1965 Land systems of the Tipperary area. Land Research Series 13:18-38. Speck NH 1968 Vegetation of the Dawson-Fitzroy area. Land Research Series 21:157-173. Speck NH & Baird AM 1984 Vegetation of Yule Brook Reserve near Perth, Western Australia. Journal of the Royal Society of Western Australia 66:147-162. 77 Journal of the Royal Society of Western Australia CONTENTS VOLUME 77 PART 1 (Published March 1994) Page Recent Advances in Science in Western Australia 1 Royal Society of Western Australia Medal Recipients 3 1993 Medal Recipient: Professor J R De Laeter 4 A question of time: Royal Society Medallist's Lecture for 1993 J R De Laeter 5 The impact of prolonged flooding on the vegetation of Coomalbidgup Swamp, Western Australia RH Froend & PG van der Moezel 15 Rottnest Island artifacts and palaeosols in the context of Greater Swan Region prehistory CE Dortch & PA Hesp 23 PART 2 (Published June 1994) Recent Advances in Science in Western Australia 33 Convergent evolution in the dentitions of grazing macropodine marsupials and the grass-eating cercopithecine primate Theropithecus gelada N Jablonski 37 Re-examination of the Murchison Downs meteorite: A fragment of the Dalgaranga mesosiderite? AWR Bevan & BJ Griffin 45 Invertebrate community structure related to physico-chemical parameters of permanent lakes of the south coast of Western Australia DHD Edward, P Gazey & PM Davies 51 PART 3 (Published September 1994) Biosystematics of Australian mygalomorph spiders: Description of a new species of Aname and its aerial tube (Araneae: Nemesiidae) B York Main 65 Wet heathlands of the southern Swan Coastal Plain, Western Australia: A phytosociological study RS Smith & PG Ladd 71 Seed dispersal of Hibbertia hypericoides (Dilleniaceae) by ants A Schatral, SG Kailis & JED Fox 81 Holdings in the Library of The Royal Society of Western Australia 87 PART 4 (Published December 1994) PLANT DISEASES IN ECOSYSTEMS Forward WA Cowling & RT Wills 97 Symposium summary SH James 99 Session 1: Biology K Old 101 Session 2: Impact on ecology S Hopper 103 Session 3: Impact on industry L Mattiske 105 Session 4: The future SH James 107 Ecosystem pathogens: A view from the centre (east) P Bridgewater & B Edgar 109 The major plant pathogens occurring in native ecosystems of south-western Australia BL Shearer 113 Role of environment in dieback of jarrah: Effects of waterlogging on jarrah and Phytophthora cinnamomi, and infection of jarrah by Phytophthora cinnamomi EM Davidson 123 Ecological impact of plant disease on plant communities RT Wills & GJ Keighery 127 Smut and root rots on native rushes (Restionaceae) and sedges (Cyperaceae) KA Websdane, IM Sieler, K Sivasithamparam & KW Dixon 133 Impact of plant diseases on faunal communities BA Wilson, G Newell, WS Laidlaw & G Friend 139 Disease and forest production in Western Australia with particular reference to the effects of Phytophthora cinnamomi DS Crombie & FJ Bunny 145 The impact of plant disease on mining IJ Colquhon & AE Petersen 151 Threats to flora-based industries in Western Australia from plant disease RT Wills & CJ Robinson 159 Management of access K Gillen & A Napier 163 Control options of plant pathogens in native plant communities in south-western Australia GE StJ Hardy, PA O'Brien & BL Shearer 169 Future ecosystems - use of genetic resistance JA McComb, M Stukely & IJ Bennett 179 Future ecosystems - ecological balance (ecological impact of disease causing fungi in Western Australia) GJ Keighery, DJ Coates & N Gibson 181 The future - effects of plant diseases on society JT Young 185 79 Journal of the Royal Society of Western Australia CONTENTS VOLUME 78 PART 1 (Published March 1995) Page Recent Advances in Science in Western Australia 1 The role of diet in determining water, energy and salt intake in the thorny devil Moloch horridus (Lacertilia: Agamidae) PC Withers & CR Dickman 3 New records and further description of Macrothrix hardingi Petkovski (Cladocera) from granite pools in Western Australia NN Smirnov & IAE Bayly 13 Preliminary observations on termite diversity in native Banksia woodland and exotic pine Pinus pinaster plantations M Abensperg-Traun & DH Perry 15 Membership of The Royal Society of Western Australia 1994-1995 and Index of Interests 19 PART 2 (Published June 1995) Recent Advances in Science in Western Australia 29 The value of macroinvertebrate assemblages for determining priorities in wetland rehabilitation: A case study from Lake Toolibin, Western Australia RG Doupe & P Horwitz 33 An Upper Cretaceous chert nodule, apparently marine ballast, from Princess Royal Harbour, Western Australia JE Glover, RJ Davey & CE Dortch 39 Freshwater biogenic tufa dams in Madang Province, Papua New Guinea WF Humphreys, SM Awramik & MHP Jebb 43 PART 3 (Published September 1995) Recent Advances in Science in Western Australia 55 An early Triassic fossil flora from Culvida Soak, Canning Basin, Western Australia GJ Retallack 57 New Pleistocene and Holocene stratigraphic units in the Yalgorup Plain area, southern Swan Coastal Plain V Semeniuk 67 Seagrass communities in Exmouth Gulf, Western Australia: A preliminary survey LJ McCook, DW Klumpp & AD McKinnon 81 PART 4 (Published December 1995) 1995 Medal Recipient: Professor A R Main 89 The Royal Society of Western Australia Medal Recipients 90 The study of nature - A seamless tapestry: Royal Society Medallist's Lecture for 1995 AR Main 91 Diurnal stratification of Lake Jandabup, a coloured wetland on the Swan Coastal Plain, Western Australia DS Ryder & P Horwitz 99 Cocoon formation by the treefrog Litoria alboguttata (Amphibia: Hylidae): A 'waterproof' taxonomic tool? PC Withers & SJ Richards 103 Foraging patterns and behaviours, body postures and movement speed for goannas, Varanus gouldii (Reptilia: Varanidae), in a semi-urban environment GG Thompson 107 Constitution and Rules and Regulations 115 81 Journal of the Royal Society of Western Australia CONTENTS VOLUME 79 PART 1 (Published March 1996) DE LAETER SYMPOSIUM ON ISOTOPE SCIENCE Page Forward KJR Rosman iii de Laeter Symposium Sir M Oliphant iv Science to de Laeter: What now comes later HS Peiser 1 High accuracy mass spectrophotometry for isotopic abundances. EL Garner 5 Ultra-high accuracy isotopic measurements: Avogardro's constant is up! P de Bievre 11 Atomic weights: From a constant of nature to natural variations NE Holden 21 Ion optical design for mass spectrometers SWJ Clement 27 Clean laboratories: Past, present and future JR Moody 29 Meteorites recovered from Australia AWR Bevan 33 Isotopic anomalies in extraterrestrial grains TR Ireland 43 Solar and solar system abundances of the elements M Ebihara 51 Origin of the terrestrial planets and the moon SR Taylor 59 Cosmogenic noble gases in silicate inclusions of iron meteorites: Effects of bulk composition on elemental production rates O Jentsch & L Schultz 67 Aspects of low energy nuclear fission JW Boldeman 73 Isotopic studies of the Oklo fossil reactors and the feasibility of geological nuclear waste disposal RD Loss 81 Isotopic signatures: An important tool in today's world DJ Rokop, DW Efurd, T M Benjamin, JH Cappis, JW Chamberlin, H Poths & F Roensch 85 Stable heavy isotopes in human health BL Gulson 91 Lead isotopes and pollution history KJR Rosman & W Chisholm 97 Surface ionization sources and applications JE Delmore, AD Appelhans, J E Olson, T Huett, GS Groenewold, JC Ingram & DA Dahl 103 SHRIMP: Origins, impact and continuing evolution W Compston 109 Zircons: What we need to know RT Pidgeon 119 A review of Pb-isotope constraints on the genesis of lode-gold deposits in the Yilgarn Craton, Western Australia NJ McNaughton & DI Groves 123 Isotopic constraints on the age and early differentiation of the Earth MT McCulloch 131 A tale of two cratons: Speculations on the origin of continents AF Trendall 141 Neutrinos: Ghosts of creation JR de Laeter 143 PART 2 (Published June 1996) Recent Advances in Science in WA 149 Installation of Professor Michael Jones as the new President of the Royal Society of Western Australia. His Excellency Major General Michael Jeffery, AC MC, Governor of Western Australia 153 The impact of vegetated buffer zones on water and nutrient flow into Lake Clifton, Western Australia. PM Davies & JAK Lane 155 A new species of Lerista (Lacertilia: Scincidae) from Western Australia: Lerista eupoda LA Smith 161 Biogeography of the herpetofauna of the Archipelago of the Recherche, Western Australia LA Smith & RE Johnstone 165 Population and plant growth studies of six species of Eremophila (Myoporaceae) from central Western Australia GS Richmond & EL Ghisalberti 175 PART 3 (Published September 1996) Plant breeding for stable agriculture: Presidential Address 1994 WA Cowling 183 Egg laying by thorny devils ( Moloch horridus ) under natural conditions in the Great Victoria Desert G Pianka, ER Pianka & GG Thompson 195 Sea breeze activity and its effects on coastal processes near Perth, Western Australia G Masselink 199 A genetic perspective on the specific status of the Western Rock Lobster along the coast of Western Australia - Panulims cygnus George 1962 or P. longipes A Milne-Edwardes 1868? AP Thompson 207 Roost selection by the lesser long-eared bat, Nyctophilus geoffroyi, and greater long-eared bat, N . major, (Chiroptera: Vespertilionidae) in Banksia woodlands DJ Hosken 211 Waterbirds and aquatic invertebrates of swamps on the Victoria-Bonaparte mudflat, northern Western Australia SA Halse, RJ Shiel & GB Pearson 217 83 PART 4 (Published December 1996) SYMPOSIUM ON DESIGN OF RESERVES Preface P Horwitz 111 History of conservation reserves in the south-west of Western Australia GE Rundle 225 Assessing the conservation reservation system in the Jarrah Forest Bioregion NL McKenzie, SD Hopper, G Wardell-Johnson & N Gibson 241 A floristic survey of the Tingle Mosaic, south-western Australia: applications in land use planning and management G Wardell-Johnson & M Williams 249 Terrestrial invertebrates in south-west Australian forests: the role of relict species and habitats in reserve design B York Main 277 Aquatic fauna of the Warren Bioregion, south-west Western Australia: does reservation guarantee preservation? KM Trayler, JA Davis, P Horwitz & D Morgan 281 Ecosystem dynamics and management in relation to conservation in forest systems RJ Hobbs 293 Forests reservations: an overview AR Main 301 84 yV Journal of the Royal Society of Western Australia co-sponsored by: OUNDATiON and The University of Western Australia Granite Outcrops Symposium September 14-15, 1996 ISSN 0035-922X The Royal Society of Western Australia To promote and foster science in Western Australia Patron Her Majesty the Queen Vice-Patron His Excellency Major General Michael Jeffery AD MC Governor of Western Australia COUNCIL 1997-1998 President M G K Jones MA PhD Immediate Past President S Hopper BSc (Hons) PhD Senior Vice-President H Recher BSc PhD Junior Vice-President A George BA Hon Secretaries P Gardner BEng (Hons) GDipCSci DipEd V Hobbs BSc (Hons) PhD M Lund BSc (Hons) PhD Hon Treasurer G G Thompson MEd PhD Hon Editor P C Withers BSc (Hons) PhD Hon Journal Manager J E O'Shea BSc (Hons) PhD Hon Librarian M A Triffitt BA ALIA Members V Semeniuk BSc (Hons) PhD L N Thomas BSc MSc PGDipEIA H Stace BSc (Hons) MSc PhD M Harvey BSc (Hons) PhD N Gibson BSc (Hons) PhD S Griffin BSc (Hons) L Broadhurst BSc (Hons) The Royal Society of Western Australia was founded in 1914. The Society promotes exchange among scientists from all fields in Western Australia through the publication of a journal, monthly meetings where interesting talks are presented by local or visiting scientists, and occassional symposia or excursions on topics of current importance. Members and guests are encouraged to attend meetings on the third Monday of every month (March - December) at 8 pm Kings Park Board offices. Kings Park, West Perth, WA 6005. Individual membership subscriptions for the 1997/1998 financial year are $40 for ordinary Members and $20 for Student and Associate Members. Library, company and institution subscriptions are $60 for the 1996 calendar year. For membership forms, contact the membership Secretary, % W A Museum, Francis Street, Perth, WA 6000. The journal of the Royal Society of Western Australia was first published in 1915. Its circulation exceeds 600 copies. Nearly 100 of these are distributed to institutions or societies elsewhere in Australia. A further 200 copies circulate to more than 40 countries. The Society also has over 350 personal members, most of whom are scientists working in Western Australia. The Journal is indexed and abstracted internationally. Cover design : Mangles' kangaroo paw (Anigozanthos manglesii ) and the numbat (Myrmecobius fasciatus) are the floral and faunal emblems of Western Australia. Also depicted is a collection of living stromatolites which are of particular significance in Western Australian geology. The three subjects symbolize the diversity of sciences ambraced by the Royal Society of Western Australia. (Artwork: Dr Jan Taylor). Journal of the Royal Society of Western Australia, 80(3), March 1997 Granite Outcrops Symposium Registered Participants Bates, T: Leederville Baxter, A: Narrogin Bayly, I: Melbourne Beard, J: Applecross Bicknell, D: Narrogin Biedinger, N: Bonn, Germany Bindon, P: Perth Blackwell, M: Nedlands Bourne, J: Adelaide Bradshaw, SD: Perth Bussell, J: Perth Cahill, J: Perth Campbell, E: Adelaide Christensen, L:Hilton Clack, T: York Corbyn, D: Willagee Couper, D: Nedlands Cranston, P: Canberra Crossley, N: Narrogin Dell, E: Bickley Dixon, K: Perth Dodd, N: Kalannie Dodd, N: Kalannie Doronilla, A: Perth Edward, D: Perth Fisher, J: City Beach Fleay, J: York Forbes, S: Perth Fox, J: Perth Froend, R: Perth George, A: Perth Harrison, A: South Africa Harvey, M: Perth Hopper, S: Perth Howell, C: York Hussey, P: Perth Huston, R: Gidgegannup Jackson, B: Wannamal Jackson, T: Wannamal James, S: Perth Jones, M: Perth Judd, D: Washington, USA Kornoff, H: Germany Laing, I: Perth Lloyd, T: Dumbleyung Main, AR: Perth Mares, M: Oklahoma, USA Maslin, B: Perth McCrum, E: Sawyers Valley McQuoid, N: Perth John Myers RSWA President, Mike Jones Ian Bayly Mike Mares Don Couper Bert Main iii journal of the Royal Society of Western Australia, 80(3), March 1997 Minett, T: York Morgan, R: Perth Mouritz, R: Hyden Mouritz, V: Hyden Murray, S: Kulin Leigh Murray, S: Kulin ge Myers, J: Perth Needham, D: Keysbrook Numeijer, M: York Ohlemriller, R: Bonn, Germany Oliver, K: Hamersley Pigott, P: Perth Porembski, S: Bonn, Germany Alex George Bruce Maslin Robson, A: Perth Sage, L: Armadale Schmitz, A: Chidlow Thompson, G: Perth Elizabeth Wayne, A: Perth Campbell Weston, A: St James Withers, P: Kalamunda Wyatt, R: Georgia, USA Arthur York Main, B: Perth Harrison Cicely Howell Barbara York Main Stephen Hopper Robert Wyatt Stefan Porembski Ralf Ohlemiiller Nadja Biedinger Journal of the Royal Society of Western Australia, 80: 87-100, 1997 Geology of granite J S Myers Geological Survey of Western Australia, 100 Plain Street, East Perth WA 6004 email: j.myers@dme.wa.gov.au Abstract The genesis of granite is intimately related to the dynamic structure of the Earth. Granite is the main component of continents; it is one of the oldest known rocks; and the geological history of granite provides the main evidence about the growth and evolution of continents through time. Granite formed in a number of different situations. Some granite was generated in zones of rifted continental or oceanic crust, but most granite was generated in zones of collision between continents and oceanic crust, and where continents were amalgamated. Granite formed by two different processes: by fractional crystallization of basaltic magma; and by melting older continental crust. Between these end members, there is a spectrum of hybrid processes, including mixing of basaltic and granitic magmas, and contamination of basaltic magma by partial melts of different kinds of continental crust. Although superficially simple and similar, most granites reflect a complicated history of multistage, hybrid processes. This complexity has led to a diversity of interpretations, and the origin of granite has been one of the most hotly debated topics in geology. The following review outlines the nature and diversity of granite; where, how, and when granite was generated; how it was intruded through the crust; the structures that it formed; and the history of debates over its origin. What is Granite? Granite is one of the most abundant, and most widely known, rocks on Earth. The continents are dominated by granite; it forms the most ancient cores of long eroded continents, as well as lofty peaks of the youngest mountain ranges. Granite is an igneous rock comprising crystals of quartz, feldspar, mica and/or hornblende or pyroxene. The crystals are generally large (a few mm); they can be seen directly on outcrops, and give granite its rough texture on weathered surfaces. If most crystals are of similar size, a granite is described as even grained. If a granite contains very large crystals of feldspar (0.5 - 3.0 cm long), set in an even grained matrix, it is described as porphyritic, and the very large crystals are called phenocrysts (Fig 1). The crystals in granite are large because granite crystallized slowly from molten rock (magma), over 2 km below the Earth's surface. Where the same magma escaped upwards, and was erupted through fissures or volcanoes, it crystallized rapidly, forming small crystals or glass, and produced volcanic rocks. The colour of most granites ranges from pink to cream, white and grey, and generally reflects the colour of the dominant feldspars. Many granites contain straight, parallel-sided, cream or pink veins of pegmatite (Fig 1). This is generally a very coarse grained variety of granite that crystallized from residual granitic magma in cracks that formed during, or soon after, the solidification of the host granite. Quartz and feldspar are the dominant minerals in granite, and together make up 90% of the rock. Quartz © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 itself generally ranges from 20-45% and feldspar up to 60% of the granite. The proportions of these minerals, and the kinds of feldspars, are used to divide granites into a number of types. Feldspars are alumino-silicates Figure 1. Typical outcrop of (porphyritic) granite in Western Australia. The rock comprises large, short tabular (white) feldspar crystals (phenocrysts) in an even grained matrix of smaller crystals of biotite (black), quartz (white), and feldspar (white). The granite is crossed by a number of white pegmatite veins, comprising quartz and feldspar crystals that grew in fissures as the fissures opened. The granite is about 2700 million years old, and is located about 120 km northwest of Meekatharra. 87 Journal of the Royal Society of Western Australia, 80(3), September 1997 that contain a range of calcium, potassium and sodium. They are divided into two groups on the basis of these three elements: potassium or alkali feldspar, and plagioclase that itself ranges from calcium-rich to sodium- rich end members. Granite sensu stricto is rich in potassium feldspar (up to 65% of feldspars) relative to plagioclase, and has a high content of quartz (20-60%). With increasing plagioclase/ potassium feldspar ratio, and increasing calcium/sodium ratio in plagioclase, granite grades into rocks called granodiorite (plagioclase 65-90% of feldspars) and tonalite (plagioclase 90-100% of feldspars). With decreasing quartz content (quartz less than 20%), granite grades into (potassium feldspar - rich) syenite, monzonite, and (plagioclase-rich) quartz monzonite. For simplicity in this review, all these rocks of the granite family are referred to broadly as granite. They are all generally massive, coarse grained rocks, and form similar kinds of outcrops. Granite Generation in a Dynamic Earth The crust of the Earth is divided into two types, continental and oceanic. The continents are generally 30-40 km thick and mainly consist of granite, whereas the crust beneath the oceans is generally 5-10 km thick, and mainly consists of basalt. Granite largely consists of silica (65-75%) and has a lower density (average 2.7) than more iron- and magnesium-rich basalt (average Continent\Continent Rifting of Oceanic Crust\Continent collision Oceanic Crust collision density 3.0) with only 50% silica. Therefore the continents are more bouyant, more elevated, and thicker than oceanic crust. The main structure of the Earth is shown in Figure 2. Both the oceanic and continental crusts of the Earth are relatively thin, and their thickness is exaggerated in this illustration. They float on a region of ductile, semi-molten rock called the Mantle, and below that lies a ductile Outer Core and a more rigid Inner Core of iron and nickel. Heat generated by radioactive decay leads to convection within the Mantle. Rift zones Where convection currents rise and move apart, the thin crust is split, and basaltic magma is erupted through the resulting fissures. Where this occurs beneath the oceans along mid-ocean ridges, more oceanic crust is generated. Most of this remains submerged, except locally where the submarine mountain chains rise above sea level, or where the oceanic crust is abnormally thick, such as in Iceland (Fig 3). In some cases, the magma collects in pools deep in the crust and starts to crystallize. The first crystals to form, dense iron and magnesium - rich minerals (olivine, pyroxene, and spinel), may sink in the magma and accumulate at the bottom of these magma chambers. Thus the composition of the residual magma gradually becomes relatively richer in silica and alumina. This process is called fractional crystallization and leads to the generation of a variety of igneous rocks. Only a small amount of granite is generated in this situation where oceanic crust is pulled apart, but where continents are pulled apart some of the older continental crust may be partly melted and rise as granitic magma. A good example of this can be seen on the east coast of Greenland which was rifted from northwest Europe about 53 million years ago as the Atlantic Ocean started to form. Basaltic magma was erupted through steep fissures called dykes in the splitting continental crust (Fig 4). The magma was erupted as lava flows over the subsiding edges of the rifting continent and formed extensive piles of basalt several kilometres thick. These basalt flows are well preserved in East Greenland and in northwest Britain, including the famous localities of the Giant's Causeway in Ireland, and Fingal's Cave on Staffa in Scotland. Some basaltic magma that remained in pools within the crust crystallized slowly as gabbro (Fig 6), the coarse grained equivalent of basalt. Many of these gabbros are compositionally layered and show fractional crystallization. In some places fractional crystallization of basaltic magma, and partial melting of the continental crust, generated granitic magma (Fig 5). Collision zones Most granite was not formed in rift situations, but where Mantle convection currents converged. There the crust was pushed together and thickened, and the deeper parts melted. The two main situations are shown on Figure 2; 1. continent/oceanic crust collision in which granite was generated in a number of stages, and intruded passively into extending crust, and 88 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 3. Basalt lava (grey) forming extensive flows and small volcanic cones that were erupted between 1725 and 1729 AD at Krafla in northeastern Iceland. Iceland is located on the mid-Atlantic ridge that marks rifting between North America and Europe. The small valley extending between the volcanic cone in the left foreground and the cone on the far right, lies along the axis of the main rift, and formed during an episode of rifting at about 1725 AD. The darker, new lava, in the vicinity of the fumaroles in the centre and left centre of the photograph, was erupted in December 1975 at the start of a new phase of rifting and volcanic activity that continues to the present. View to the north in July 1977. Figure 4. Black dykes of basalt intruded into fissures in much older ( ca 2800 Ma) granitic gneiss, during initial rifting between North America and Europe. Headland between Tasilaq and 0stre Tasissaq, west of Kap Gustav Holm, East Greenland. The mountain is 1000 m high. 89 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 5. Granite generated during continental rifting between North America and Europe. The mountain exposes the upper, dome¬ shaped portion of a circular granite body (white), intruded into black diorite 35 million years ago. Thin sheets of granite occur in the diorite parallel to the roof of the main granite. Auluiartik island, Kialineq, East Greenland. The mountains are 500 m high. Figure 6. Igneous layers rich in plagioclase feldspar (pale) or pyroxene (dark) in uniform gabbro, formed during the crystallization of basaltic magma. The magma was intruded into an opening cavity at the base of a thick pile of basalt lava flows (forming the upper parts of the mountains), during initial rifting between North America and Europe about 53 million years ago. Skaergaard Intrusion, East Greenland. 90 Journal of the Royal Society of Western Australia, 80(3), September 1997 2. continent/continent collision in which granite was generated directly by melting of continental crust, and was intruded as sheets that were deformed during and after crystallization. Continent/Oceanic Crust Collision The main features of continent/oceanic crust collision are shown in Figure 7. Oceanic crust is pushed down into the Mantle below the continental crust, forming a structure called a subduction zone. At depth, within this zone, the oceanic crust melts, and also introduces saline fluids which induce melting in the overlying Mantle. The melts are basaltic in composition. They collect, rise, and spread out in the lower crust where they may crystallize as gabbro, or form granitic magma by fractional crystallization and contamination with Mantle 5-10 km Continental Crust 20-30 km ^7 Brittle Ductile Melting Figure 7. Schematic crustal section showing the generation and intrusion of magmas above a slab of oceanic crust subducted below a continental margin. melting continental crust. The granitic magmas collect and rise to form large bodies that may fractionate further, forming a whole spectrum of different granitic magmas that, with increasing contents of silica and alkalis, may crystallize as tonalite, granodiorite and granite. Granite intrusion processes In the lower, ductile, part of the continental crust, granitic magma may rise as diapirs that push aside the adjacent rocks (Fig 8). In higher, semi-ductile, parts of the crust, where the magma can no longer rise diapirically, it may spread out to form bodies with flat floors and dome¬ shaped roofs called laccoliths (Fig 8). In the upper, brittle, part of the crust, granitic magma is intruded along fractures (Figs 8 and 9). In some cases the fracturing leads to the collapse of large blocks of crust, with dimensions of several kilometres, that founder into underlying pools of magma. The magma rises up the vertical fractures and fills the space being created by the sinking block. This is the main process by which large volumes of granitic magma are intruded to high levels in the crust, and is called cauldron subsidence (Figs 8 Brittle CAULDRON SUBSIDENCE DIAPIR Ductile f t Rise of Magma Coalescence of melts Melting and/or crystal fractionation Figure 8. Schematic sections showing modes of granite intrusion at different crustal levels. 91 Journal of the Royal Society of Western Australia, 80(3), September 1997 and 9). The process of fracturing of the older rocks around granite intrusions is called stoping, and leads to the incorporation of angular fragments of wall and roof rocks into granitic magma chambers (Fig 10). These rock fragments are trapped in the granite as it solidified and are known as xenoliths or inclusions. Inclusions can also arise by a completely different process. In many instances, dykes of basaltic magma were intruded from deeper magma chambers up through and into granite plutons and crystallizing bodies of granitic magma. Where the granite was already solidified, the basaltic dykes crystallized in the straight fractures into which they were intruded. But where the granite was still a mixture of liquid and crystals, the dykes, after intrusion as linear bodies along shock induced fractures, broke up into linear trains of globular fragments called pillows within the still unconsolidated granite (Fig 11). Figure 9. Left: Vertical sections showing progressive stages a - e in the intrusion of granite by cauldron subsidence in the Coastal Batholith of Peru, a: fractures define the framework of cauldron subsidence; b: turbulent mixtures of gas and magma (black) move up the fractures; c: the central block of older rocks (white) subsides and granitic magma (crosses) is intruded into the opening space; d: the granite body is displaced downwards by a second episode of cauldron subsidence; e: flat-lying sheets of pegmatite (dots) are intruded into the roof region of the second granite body during further slight subsidence. Right: Horizontal sections of the granite plutons at the five levels marked on section c. Figure 10. Angular fragments of older rocks (stippled) incorporated by stoping into granite (white) 66 million years ago. Puscao granite. Coastal Batholith of Peru, 25 km ESE of Huarmey. Granite structures and cordilleran batholiths Magmas generated along subduction zones rise to form narrow, linear belts, generally 50 - 100 km wide and thousands of kilometres long. These belts occur within, and parallel to, the margins of continents. The Andes (Fig 12) provides one of the simplest examples. Here, in a head-on collision, the South American continent is being pushed over Pacific oceanic crust. This situation has existed for over 100 million years, and Figure 12 shows the extent of granite and associated volcanic rocks generated during this time by the processes shown on Figure 7. The granites and volcanic rocks form a major part of the Andes, and in Peru their structure can be clearly seen in mountainous desert with vertical relief of 5 km (Fig 13). The granite bodies have steep walls and flat roofs. The cross section, Figure 14, shows some of the complexity of repeated granite intrusion by cauldron subsidence into the same 50 km wide belt, over a period of 35 million years, between 100 and 65 million years ago. A complex of related granite bodies such as this is called a batholith, and batholiths that formed above subduction zones are known as cordilleran batholiths. 92 Journal of the Royal Society of Western Australia, 80(3), September 1997 Associated volcanic debris Here in the Andes, the granites rose into volcanic rocks derived from the same magmas. In some places they are associated with calderas, large circular depressions, up to 25 km in diameter, in which the older volcanic rocks sank into underlying bodies of magma. At the same time, large volumes of volcanic rocks were explosively erupted from circular fissures around the caldera rims. High clouds of ash, steam and gas, dropped rock and crystal fragments over vast areas, forming deposits that later solidified to form a volcanic rock called tuff. In some cases the ash columns collapsed, or the eruptions did not become airborne, and the erupted mixtures of hot gas, magma, crystals and rock fragments travelled as turbulent flows for huge distances (often over 50 km), very fast on thin cushions of gas, and left deposits of fused fragmentary volcanic rocks called ignimbrite. These deposits are widely preserved in southern Peru and northern Chile (Fig 15) where, because of extreme aridity, there has been less incision by erosion than in central and northern Peru, and southern Chile. Figure 11. Black pillow fragments of a basalt dyke intruded into a granite before the granite had solidified. Note the cauliform margins of the pillows indicating that both basalt and granite were ductile when the basalt dyke was fragmented. Lillee, Kialineq, East Greenland. Figure 12. Map showing major outcrops of granite and volcanic rocks of the Andes, formed along the continental margin of South America where Pacific oceanic crust is subducted below continental crust. Arrows indicate the main direction of movement of the two slabs of crust. Continent/Continent Collision Granite was also generated from melting of continental crust where continents collided and thickened as one wras pushed over the other (Fig 2). This situation is still active in the Himalayas and beneath the Tibetan Plateau where the Indian continent has already been pushed for 2000 km under the edge of the Eurasian continent during the past 50 million years. Many continental margins involved in continent/ continent collisions were preceded by episodes of continent/ocean collision in which granite was generated above subduction zones, as in the Andes. In the Himalayas, the 2500 km long Transhimalayan Batholith and part of the Karakoram Batholith (Fig 16) formed as Andean-type cordilleran batholiths above subduction zones, before the collision of India and Eurasia. In many cases, thin strips of older continental crust, ocean ridges, oceanic plateau, and basaltic island arcs such as those now forming in Java and Japan where oceanic crust is subducted below another slab of oceanic crust, were swept against a continental margin and trapped and deformed during later continent/continent collision. Therefore the structure and geological history of most continent/continent collision zones is much more complicated than those of continent/ocean collision. 93 Journal of the Royal Society of Western Australia, 80(3), September 1997 mm Figure 13. Extensive granite outcrops forming a belt 50 km wide and 2000 km long in the Andes in Peru. View southeast along the Coastal Batholith from an altitude of ca 2100 m, west of Chasquitambo. The batholith was mainly intruded between 100 and 65 million years ago into coeval volcanic rocks. The volcanic rocks (black), forming the flat roof of the batholith, can be seen in the distant (left) mountain peaks, and on a chain of peaks in the distant centre within a ring dyke of granite, 1 km thick and 25 km in diameter, emplaced by cauldron subsidence. Similar volcanic rocks occur beneath the fog (distant right), where they form the western vertical margin of the batholith adjacent to the Pacific Ocean. The black rocks in the middle foreground are large fragments of gabbro that crystallized from basaltic magma before the intrusion of the granite. sw NE Figure 14. Vertical section across a typical cordilleran batholith showing the structure formed by repeated episodes of igneous intrusion of gabbro, tonalite and granite by cauldron subsidence into related volcanic rocks. Section across the Coastal Batholith of Peru, located due west of the word continental crust on Figure 12. Note that horizontal and vertical scales are equal. 94 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 15. Valley (foreground) incised into extensive sheets of ignimbrite, erupted from granitic magma in northern Chile. The ignimbrites were erupted from fissures parallel to the western summits of the Andes (left) and flowed down towards the Pacific Ocean (off to the right), between 10 and 5 million years ago. Volcanic cones (distant left) formed from eruptions of basalt and andesite during the past 1 million years. View south from an altitude of 4300 m between Tatio and Lasana. Figure 16. Granite (distant snow-covered 7000 - 8000 m peaks) of the Karakorum Batholith that initially formed at c. 90 Ma during subduction of oceanic crust beneath the margin of Asia before the collision with India about 50 million years ago. View to the north from an altitude of ca 2500 m, east of Hunza, Pakistan. The Hunza valley (foreground) was incised during recent, ongoing uplift of the Himalayan range. 95 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 17. Map showing the assembly of three independent rafts of continental crust (cratons) 1300 million years ago, and the location of the Albany-Fraser Orogen, Arrows indicate the main directions of movement of the three continental rafts. Albany-Fraser Orogen Some of the products of continent /continent collision can be seen in the old, deeply eroded remnants of former mountain belts in Western Australia, such as the Albany-Fraser Orogen. About 1300 million years ago, three rafts of continental crust were joined to form a major part of what is now the Australian continent (Fig 17). The south coast of Western Australia exposes the roots of a mountain belt that formed at this time when a West Australian craton or continent, was joined to a combined South Australian - East Antarctic continent called the Mawson Craton. This kind of collision zone is called an orogen, and this particular zone is known as the Albany-Fraser Orogen. Granite was extensively generated during two distinct episodes of this continental collision, and forms the main rock outcrops of the region. The older granite, called Recherche granite, was intruded intermittently during the collision 1300 million years ago (Fig 18). At that time the margin of the Mawson continent was being deformed and thickened as slabs of continental crust were stacked over each other. Some of the granite was intensely deformed and recrystallized, deep in the crust. Pegmatite veins and inclusions were all streaked out into parallel layers and the granite was converted into a metamorphic rock called gneiss (Fig 19). There is widespread evidence that a smaller amount of granite also locally formed in situ by melting of older (1700 - 1600 Ma) granite of the Mawson continent, immediately after the initial episode of deformation and crustal thickening at 1300 Ma (Fig 20). This kind of metamorphic rock, comprising a mixture of older rocks with younger veins of granite, and diffuse patches of melting, is known as migmatite. The younger granite, intruded about 1180 million years ago, is called Esperance granite (Fig 21), and is prominent to the east and northeast of Esperance, along the south coast from William Bay through Albany to Mount Manypeaks, and forms the Porongorups. This granite, like the Recherche granite, may be largely derived by melting of older continental crust at depth, during another episode of oblique compression in which 96 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 18. 1300 Ma Recherche Granite with angular stoped and veined fragments of previously deformed, darker, 1700 - 1600 Ma granite containing remnants of basaltic dykes. Albany-Fraser Orogen: coast south of Lake Gore, west of Esperance. Figure 19. Banded granitic gneiss formed by intense deformation of granite and pegmatite veins (Recherche Granite) 1300 million years ago in the Albany-Fraser Orogen. Butty Head, west of Esperance. Figure 20. Irregular, diffuse veins of (pale) granite (beneath the hammer) formed 1300 million years ago by melting of older banded granitic gneiss (1650 Ma) in the Albany-Fraser Orogen. South of Lake Gidong, west of Esperance. 97 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 21. Outcrops of porphyritic Esperance granite in the AlbanyFraser Orogen, east of Esperance. This granite contains fragments of older Recherche granite, into which it was intruded 1180 million years ago. View to the southeast, at Thistle Cove, east of Cape Le Grand. GRANITE OUTCROPS IN WESTERN AUSTRALIA Geraldton Albany 11300-1000 Ma 2000-1600 Ma | « » | >2500 Ma Figure 22. Map showing the main granite outcrops in Western Australia, with the ages of granites given in million years (Ma). Note that the granite in the vicinity of Cape Leeuwin is much younger than all the other granites shown. the West Australian continent continued to slide northeastward relative to the more westerly-moving Mawson continent. Older orogens in Western Australia There are several remnants of even older orogens in Western Australia that mark former episodes of continental collision and granite generation. The most prominent orogens with extensive granite outcrops occur in the Kimberley region (Halls Creek and King Leopold orogens), and in the Gascoyne region (Capricorn Orogen) (Fig 22). The Capricorn Orogen marks the joining of the formerly independent Pilbara and Yilgarn continents about 2000 million years ago. Granite was generated in a cordilleran-type batholith before the collision (Fig 23), and again during and after the collision by melting of thickened continental crust. These older Pilbara and Yilgarn continents also largely consist of granite, although in the Pilbara much of this is buried by volcanic rocks and the Hamersley banded iron formations. The most extensive granite outcrops form part of the Yilgarn Craton or continent. Most of these granites were generated by melting of older continental crust between 2700 and 2600 million years ago (Fig 1), when this continent was assembled from a number of smaller rafts of continental crust. Deformed remnants of older continental crust are exposed in the Narryer gneiss complex that forms the northwest part of the Yilgarn Craton, northwest of Meekatharra. These gneisses are largely derived from granite, including some of the oldest known rocks on Earth that formed 3600 million years ago (Fig 24). 98 Journal of the Royal Society of Western Australia, 80(3), September 1997 2000 Ma PILBARA CRATON YILGARN CRATON Island arc volcanics and batholiths Back-arc volcanics PILBARA CRATON Fold-and thrust-belt \ w Deformed foreland basin and older passive YILGARN CRATON Obduction of arc volcanics JSM12 16 10 95 Figure 23. Schematic, vertical cross-sections showing the generation (at 2000 Ma) and deformation (at 1850 Ma) of granite (diagonal crosses) in a cordilleran batholith involved in continental collision between the Pilbara and Yilgarn cratons. Granite in Historical Perspective During the late 18th century, when geology was in its infancy, it was widely believed that granite was the oldest of rocks and was a precipitate from a world-wide ocean of molten rock. Supporters of this idea, advocated most strongly by Abraham Werner from the Mining Academy of Freiburg in Saxony, were known as Neptunists. This concept was first challenged in 1795 by James Hutton of Edinburgh who described veins of granite cutting across sedimentary rocks in Scotland, in his book "Theory of the Earth". He saw that the sedimentary rocks were baked by the granite and concluded that the granite must have been intruded as a hot liquid from subterranean regions of molten rock. His supporters were known as Plutonists, but were a small minority until joined by the influential Charles Lyell in the 1830s. During the 1830s, just as Hutton's concept of the igneous origin of granite was becoming widely accepted, granite was found in many places to be closely associated with rocks that had been transformed under conditions of high temperature and pressure, deep in the Earth, into rocks that Charles Lyell called metamorphic. Various kinds of igneous and sedimentary rocks appeared to have been converted into granite by a process known as granitisation. This theory of granite formation became predominant during the latter part of the 19th century, and remained so until the early 1960s. Its supporters were generally known as transformists. This theory was widely supported in France by Michel-Levy, Lacroix, Termier and others; in Scandinavia by Keilhau and Sederholm; in Switzerland by Wegmann; and in Britain by Read. Meanwhile, during the first half of the 20th century, a minority led by the experimental and Figure 24. Granite (below the hammer) intruded between 2700 and 2600 Ma into banded gneiss (top right), derived from much older, 3600 Ma, granite by deformation and recrystallization. Narryer gneiss complex, 170 km northwest of Meekatharra. 99 Journal of the Royal Society of Western Australia, 80(3), September 1997 geochemical work of Bowen and Nockolds demonstrated the importance of fractional crystallization in producing a diversity of igneous rocks from basalt to granite. The last 30 years has seen a general decline in the popularity of granitisation, and more popular support for an igneous origin of granite. The advent of plate tectonic theory during the late 1960s provided a framework in which a diversity of different interpretations of different kinds of granite could be reconciled. Some granites formed by melting, or partial melting, of older heterogeneous continental crust, ranging from gabbro and granite to sedimentary rocks. Some granites formed by fractional crystallization from basaltic magmas. Many granites were derived by a combination of both processes: by fractional crystallization of basaltic magma, contaminated by heterogeneous continental crust; and by mixing of basaltic and granitic magmas. The bulk composition and texture of many granites are superficially similar, but this often masks a complicated history of multistage, hybrid processes. Detailed geochemistry, especially of trace elements and isotopes, and detailed geochronology, especially analysis of individual zircons, are becoming increasingly important in deciphering the source rocks, the conditions of melt generation, and the evolution of granitic magmas. Debates on the origin of granite continue. Although abundant and superficially simple, the interpretation of granite still offers exciting challenges for future generations of geologists. References Much of what is said here about granite can be found in current geological textbooks. In addition, a recent book by Pitcher (1993) on granite, provides a detailed account of current knowledge of granite and its historical perspective. Another overview of granite geology is given by Atherton (1993). The proceedings of a 1995 conference on granite (Brown et al. 1996) present a spectrum ot recent work on the geology of granite. Other recent key references to Western Australian granites mentioned above are included in this selected bibliography. Anon 1990 Geology and Mineral Resources of Western Australia. Geological Survey of Western Australia, Perth. Memoir 3. Atherton M P 1993 Granite magmatism. Journal of the Geological Society of London 150:1009-1023. Brown M, Candela P A, Peck D L, Stephens W E, Walker R J & Zen E-an 1996 Third Hutton Symposium - The origin of granites and related rocks. Transactions of the Royal Society of Edinburgh, Earth Sciences, Volume 87. [Also published as Geological Society of America Special Paper 315, 370 pp] Griffin T J & Tvler I M in press 1996 Geology of the King Leopold Orogen. Geological Survey of Western Australia, Perth. Bulletin 143. Hickman A H 1983 Geology of the Pilbara Block and its environs. Geological Survey of Western Australia, Perth. Bulletin 127. Myers J S 1995a Albany, Western Australia. 1:1 000 000 Geological Series Map and Explanatory Notes. Geological Survey of Western Australia, Perth. Myers J S 1995b Esperance, Western Australia. 1:1 000 000 Geological Series Map and Explanatory Notes. Geological Survey of Western Australia, Perth. Pitcher W S 1993 The Nature and Origin of Granite. Chapman & Hall, Glasgow. Williams S J 1986 Geology of the Gascoyne Province. Geological Survey of Western Australia, Perth. Report 15. 100 Journal of the Royal Society of Western Australia, 80:101-112, 1997 Granite landforms E M Campbell Department of Geology and Geophysics, University of Adelaide, North Terrace, Adelaide SA 5005: email campbell@geology.adelaide.edu.au Abstract Fresh granite is low in porosity and permeability, but is highly pervious by virtue of a connected series of orthogonal and sheet fractures. Granite is susceptible to weathering by moisture, leading to the formation of a regolith. The course and rate of weathering are influenced by the structure of the rock, including fractures; mineral composition; texture, especially size of the crystals; and the physical, chemical and biotic nature of the invasive water. Plains occupy extensive areas of granite outcrops, but positive relief forms - bornhardts, nubbins and castle koppies - have attracted most attention. Minor forms developed on granite are due primarily to weathering, e.g. boulders, flared slopes, rock basins, tafoni. Many of them are initiated in the subsurface. Some, like A-tents, are tectonic. Some forms are convergent as they evolve in different ways. Several are not yet satisfactorily explained. Introduction The aim of this review is to describe and explain the assemblage of landforms, both major and minor, characteristically developed on granite. None of these landforms is peculiar to granite: each of the forms is developed in a variety of different materials and they all have developed under a wide range of climates (Campbell & Twidale 1995a). Nevertheless, the assemblage of landforms is typical of granite, so much so that it is possible, in many circumstances, to predict from afar the nature of the underlying rock. Weathering patterns are of primary importance in the development of many granite forms, though others are due principally to tectonic forces (see below). The interaction of granite with the atmosphere and the hydrosphere leads to weathering of the constituent minerals and the formation of a regolith. Weathering, the disintegration or alteration of rocks at and near the Earth's surface, can conveniently be divided into those processes which involve chemical reactions and the formation of new minerals (chemical weathering) and those that involve only physical breakdown of the rock (physical weathering). Most commonly, however, these several processes operate together. The type and rate of weathering of rocks depend on rock characteristics, the physical, chemical and biotic characteristics of the weathering fluids, the nature of reactions at the mineral surface and environmental factors (Colman & Dethier 1986). The base of the regolith, the contact of the regolith with the unweathered bedrock, is called the weathering front (Mabbutt 1961). The weathering front may be irregular, partly due to structural weaknesses of various types and partly as a result of the concentration of moisture. The rock characteristics which influence the rate of weathering are fracture density, mineral composition © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 and texture. Being composed of interlocking crystals of quartz, feldspar and mica, fresh granite is of low porosity and permeability. Granite is pervious because water penetrates the rock along partings, especially orthogonal joints and sheet fractures. Fractures occur at all scales from microcracks within minerals to partings that can be kilometres in length. Highly fractured rocks are more susceptible to weathering than those in which the fractures are absent, widely spaced or tightly closed. In some instances, even though no fracture is discernible, the crystals may be in strain, leading to a higher susceptibility to weathering (Russell 1935). The stability of the common silicate minerals is closely related to their order of crystallisation from a silicate melt (Goldich 1938). Minerals that crystallise at the highest temperatures, especially biotite and plagioclase feldspar, are unstable at the Earth's surface and are susceptible to alteration, and granites with a higher than average proportion of these more readily weathered minerals tend to form negative features whereas those with a greater abundance of quartz and potassium minerals, which crystallise at lower temperatures and are more stable, tend to be more resistant (e.g. Brook 1978). This variation in suceptibility to weathering is demonstrated at the crystal scale by the occurrence of pitted surfaces where granite is in contact with water. The biotite and plagioclase feldspars are chemically altered, leaving the quartz and potassium feldspars in relief. As the surface is rapidly reduced by flaking, such pitted surfaces are an indicator of recent exposure from beneath the regolith (Twidale & Bourne 1976a). Variations in crystal size do not produce consistent effects (Twidale 1982). Finer-grained granites tend to be more susceptible to weathering due to the greater surface area to volume ratio of each crystal. However, coarser crystals, especially those that are in strain and hence traversed by an array of microcracks, may be more strongly weathered (Russell 1935; Pope 1995; Hill 1996). Organisms, especially microorganisms such as bacteria, algae and lichens, significantly affect the rate of weathering of granite. Algae readily colonise rock 101 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 1. Ucontitchie Hill, Eyre Peninsula, South Australia, is an inselberg which rises abruptly from the surrounding plain. It is also a bomhardt due to the domical profile, a reflection of the sheet fractures essentially parallel to the surface (CR Twidale). surfaces. Their hyphae penetrate between and within the mineral grains by a combination of physical activity and biochemical reactions (Yatsu 1988). Lichens are characteristic of granite exposures and although in humid environments they protect the surface, in arid environments they contribute to the physical and biochemical weathering of the rock (Fry 1927; Viles 1988). The retention of moisture by plants and soils, especially in fissures and basins, and the humic acids resulting from plant and animal decay also increase weathering. The thickness of the regolith is a function of the relative rates of weathering and erosion, or the wearing away of the land surface. In most areas erosion is principally by running water. In tectonically stable regions, or those situated far from the sea, which is the ultimate base level of erosion, or those in humid tropical environments where weathering proceeds rapidly, or those in arid or semiarid environments where erosion of the regolith may be restricted, great thicknesses of regolith (100 m or more) may develop. - V" Y widely spaced closely spaced fractures orthogonal fractures Figure 2. Schematic diagram illustrating the two-stage development of bornhardts. The subsurface exploitation of fractures by weathering is followed by the erosion of the regolith and the exposure of the weathering front. A similar two-stage or etch explanation of boulders is also illustrated (after Twidale 1982). 102 Journal of the Royal Society of Western Australia, 80(3), September 1997 Major Landforms Plains are well representative of granite outcrops; they take various forms. Peneplains are gently rolling surfaces, like those on northwest Eyre Peninsula, South Australia, and the southern Yilgarn of Western Australia. Pediments, or rock platforms, are smooth and gently sloping, essentially bare rock surfaces located on the margins of uplands. Mantled pediments carry an in situ regolithic veneer. Plains develop in different ways. Some are epigenetic, being formed by subaerial agencies, particularly running water. Etch forms, such as in the plain around Meekatharra, Western Australia, develop as a result of weathering and the formation of a regular weathering front, followed by stripping of the regolith down to the weathering front. Some plains have been buried and later exhumed (Twidale 1982), like that resurrected from beneath a cover of Early Cretaceous strata, found in the north of Western Australia near Port Hedland. However, it is the positive relief features that have attracted the attention of most geomorphologists. Twidale (1982, 1993) divided the major forms, on the Figure 3. A: Flared slope on Ucontitchie Hill, Eyre Peninsula, South Australia. This prominent overhang is developed on a spur of massive rock, the shoulder of which is 5-6 m above the rock platform. The shoulder marks the level of the land surface prior to the stripping of the weathered material and exposure of the weathering front. B: Yarwondutta Rock, Eyre Peninsula, South Australia, showing flared side walls (S) and steps marking stages in exposure of the inselberg from beneath the regolith. The boulder with tafone developed underneath is 2 m high. (CR Twidale). 103 Journal of the Royal Society of Western Australia, 80(3), September 1997 basis of shape, into inselbergs, bornhardts, nubbins and castle koppies. Inselbergs are isolated steep-sided island mountains rising abruptly from the adjoining plains (Fig 1). The forms also occur in groups or massifs. The shape of the individual hills varies. Bornhardts are the basic form. They are rounded, domical forms in massive bedrock, bounded by orthogonal fractures giving rise to a circular to square or rectangular plan shape. Sheet fractures, which parallel the surface, are prominent, resulting in a domical profile. Several explanations have been proposed for bornhardts. Some, e.g. the Pic Parana in south-eastern Brazil, are upfaulted blocks (Lamego 1938), but most are not fault-defined. Some granite bornhardts, e.g. those near Hiltaba in the Gawler Ranges, South Australia, are small plutons that have been exposed as a result of the preferential erosion of the weaker host rocks into which they were injected (Hurault 1963; Campbell & Twidale 1991), though such an explanation cannot apply to those bornhardts shaped in sedimentary rocks. In some instances, the composition of the bornhardts is more resistant to weathering than the rock of the adjacent areas (Hurault 1963; Brook 1978). However, most bornhardts are composed of material of essentially the same composition as that underlying the surrounding plain. Two explanations proposed to explain bornhardts have been widely accepted. First, bornhardts are considered by many to be the last surviving remnants following long distance scarp retreat (King 1942, 1962), If scarp retreat is a valid explanation, bornhardts ought to be restricted to major drainage divides but this is not so; backwearing of marginal scarps is restricted to a few score metres at most (Twidale & Bourne 1975a) rather than the scores of kilometres demanded by this hypothesis; and, if scarp retreat is involved, no bornhardt should survive through more than one cycle of erosion or, in general terms, for more than about 33 million years. Many are of greater antiquity and the upper parts of some on Eyre Peninsula, South Australia, may be of Mesozoic age, at least 100 million years (Twidale & Bourne 1975a). Also, the scarp retreat hypothesis is unable to explain several aspects of the field evidence (Twidale 1982). Second, the field evidence points to bornhardts being structural forms developed on compartments of resistant rock with few open fractures whereas the surrounding rock is highly fractured and therefore subject to Figure 4. A: Boulder-strewn surface of a granite nubbin, Naraku, northwest Queensland. Note the termite mounds in the foreground. B: Castle koppie, formed of angular joint blocks, near Harare, central Zimbabwe. (CR Twidale). 104 Journal of the Royal Society of Western Australia, 80(3), September 1997 weathering (Fig 2). According to this theory, bornhardts develop in two stages; differential weathering beneath the surface followed by removal of the regolith to expose the weathering front. They are thus etch forms. The unfractured rocks are resistant to weathering and remain as high points in the landscape, whereas the well- fractured zones are easily weathered and are hence more susceptible to erosion. Once upstanding, bornhardts tend to shed water, which is concentrated in the lower zones where weathering proceeds more rapidly. In support of this hypothesis, numerous examples of convex-upward masses of unweathered granite have been exposed in excavations. Also, in many instances, for example at Ucontitchie Hill in South Australia, the fractures beneath the plain, and evidenced by dam construction, are closer than those widely-spaced fractures on the hill. In addition, the preservation of the outer layers of the bornhardt may be influenced by slight concentrations of iron oxide or silica associated with the weathering front; though lichens and mosses may also concentrate these elements after exposure (Twidale 1982). The relative rates of weathering and erosion of the bedrock may lead to variations of slope of the exposed bornhardts. For example, flared slopes, as at Wave Rock, southwest Western Australia, and Ucontitchie Hill, Eyre Peninsula, South Australia (Fig 3A), and stepped topography (Fig 3B; Twidale 1982) are expressions of subsurface weathering and stages in exposure, as a consequence of which the bornhardts increase in relative relief as the surrounding plain is lowered. Bornhardts are the basic major positive relief form from which nubbins and castle koppies are derived (Fig 4). The block- or boulder-strewn nubbins are due to the partial breakdown beneath the surface of sheet structures. Small, steep-sided castle koppies are explained as domes modified by pronounced marginal weathering along vertical fractures and in the subsurface. Alternatively, the presence of strongly-developed near-vertical fractures and pronounced frost action may result in the development of pointed or needle-like forms such as in the Organ Mountains in New Mexico (Seager 1981). An etch origin similar to that for bornhardts but on a smaller scale is suggested for granite corestones and boulders, which are perhaps the most typical global granite form (Fig 4A). Spherical to ellipsoid corestones of intrinsically fresh rock are set in a matrix of weathered material. As the land surface is lowered the friable weathered material is removed leaving the boulders in situ (Hassenfratz 1791). This is the origin of such forms as the Devil's Marbles in central Australia (Twidale 1980). Pillars, such as Murphy Haystacks, Eyre Peninsula (Fig 5), are intermediate forms consisting of attached blocks or towers from which the surrounding regolith has been removed (Twidale & Campbell 1984). Minor Forms Minor forms developed on granite outcrops have been classified by Campbell & Twidale (1995b) into those due mainly to weathering and those that are tectonic in origin. Here this classification is generally maintained, although subdivisions are modified. Many of the forms have evolved in different ways i.e. they are convergent (Table 1). In this brief review it is impossible to describe and explain each of these landforms. Rather a selected few will be discussed in order to illustrate the range of factors that are considered to be important in their development. Rock basins Rock basins, also known as gnammas, are circular, elliptical or irregular depressions in solid bedrock (Fig 6). Many are initiated at the weathering front which is demonstrated by the presence of shallow saucer-shaped depressions on recently exposed platforms. They commonly form along fractures and especially at fracture intersections. They are etch forms and after Table I A classification of minor granite landforms. Those forms in italics are probably convergent. 1. WEATHERING FORMS a. Initiated at the weathering front Pitting Flakes ami spalls Rock basins G u t ters a n d g r oaves Polygonal cracks Flared slopes Scarp foot depressions Deep indents Caves Clefts Pseudobedding Blocks Boulders b. Due to partial exposure Rock levees Rock doughnuts Fonts Pedestal rocks Plinths c. Initiated at the surface Rock basins Gutters and grooves Tafoni Flakes and spalls Blocks Boulders d. Due to crystal strain Clefts Gutters and grooves e. Due to gravitational pressure Rock basins Tafoni f. Due to intrusive veins Clefts Walls 2. CONSTRUCTIONAL FORMS Speleothems Boxwork pattern of ridges 3. TECTONIC FORMS A-tents Blisters Triangular wedges Fault scarps Orthogonal cracks 105 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 5. Pillars of granite with flared side walls, Murphy Haystacks, Eyre Peninsula, South Australia (CR Twidale). exposure develop varied morphologies according to the structure of the granite, the slope of the exposure and the depth of erosion. Flat-floored pans are the most common form, occurring on flattish crests in laminated rock (Fig 7). The laminations allow lateral weathering to outpace vertical weathering, in some cases resulting in overhanging sides. Hemispherical pits occur also on flattish crests but are developed where the granite is homogeneous, especially at depth beneath the superficial laminated zone. Armchair-shaped hollows are modified pans and pits and are restricted to steeper slopes Cylindrical hollows develop where deep, concentrated weathering and erosion have extended a pit through a sheet allowing throughflow and abrasion of the form (Twidale & Corbin 1963; Twidale & Bourne 1978). On the other hand, some rock basins are formed on exposed surfaces. For instance, a basin formed on the crest of a menhir, or granite monument, at St Uzek, in Brittany, must have developed after exposure when the crest was placed in a roughly horizontal position about 5000 years ago (Lageat et al. 1994). Similarly, basins have formed on recently deglaciated surfaces as, for instance, in northern Portugal and southern Galicia (Vidal Romani 1989). Gutters and grooves Some gutters and grooves (horizontal Rillen and vertical flutings respectively) are initiated at the weathering front, for, like basins, they are developed on recently exposed bedrock surfaces, as for example at Dumonte Rock on Eyre Peninsula (Fig 8; Twidale & Bourne 1975b). Just as the saucer-shaped precursors of basins are shallow, so the gutters on newly exposed surfaces are shallow and narrow. Epigene gutters can be traced into the subsurface along the weathering front where they may converge and become shallower and wider, presumably as flow filtering between soil particles becomes diffuse, until they fade at a depth of several metres. Equally, however, and spectacularly at the menhir of St Uzek, grooves have formed on the exposed flanks of the erected slab (Lageat et al. 1994). Gutters also are found on freshly deglaciated surfaces (Vidal Romani 1989). At Cash Hill, on Eyre Peninsula, gutters have developed on much of the exposed surface but do not extend beneath the regolith nor on the broad bench from which the regolith has apparently been recently stripped, suggesting that some gutters have formed subaerially. After exposure, incipient gutters are enlarged by running water. Abrasion is evidenced by the development of potholes. In some cases the gutters have become flask-shaped in cross section as a result of the undercutting of side walls by streams and in part a reflection of their development in the laminated surface zone (Fig 9). Some gutters have exploited and follow fractures, but that slope is the prime determinant of the path followed by streams, and also gutters, is demonstrated by the many places where the gutters leave fractures to follow the steepest local slope. There are also instances, e.g. on Wudinna Hill, Eyre Peninsula, where elements of the drainage have migrated to new positions suggesting that fractures and slope are not the only factors influential in their development. Gutters, in many instances, are developed along fractures but the fractures are not apparently present along the entire length of the landform and at some sites, e.g. Little Wudinna Hill, on Eyre Peninsula, gutters are paralleled by clefts in which no fractures are discernible; possibly the stresses responsible for fracturing affected the adjacent zones but were insufficient to cause rupture there (cf Russell 1935). Gutters which descend from the summit of Yarwondutta Rock and The Dinosaur, both on Eyre Peninsula, South Australia, on reaching the overhang of the flared slope, bifurcate into two grooves on either side of a central rib (Fig 10) In an immediate sense this is apparently due to the protection afforded by a thin veneer of desiccated algal slime, but why the present channel floors are not similarly protected has not yet been explained. 106 Journal of the Royal Society of Western Australia, 80(3), September 1997 B Figure 6. A: Rock basin or gnamma, a hemispherical basin or pit developed on homogeneous granite, Pildappa Rock, Eyre Peninsula, South Australia. B: A 1 m deep pan developed in laminated granite and with overhanging sides, Yarwondutta Rock, Eyre Peninsula, South Australia (CR Twidale). Flared slopes Flared slopes are found in a variety of lithological, climatic and topographic settings. They are particularly well developed in the granite of south-western and southern Australia. Flares are characteristically found around the base of hills but they are also found on higher slopes (Fig 3A,B) and in clefts. Several distinct flares may be present. Many flares are inclined rather than horizontal. On Eyre Peninsula, they are highest and best developed on the southern sides of the hills. Most significantly, flares shaped in bedrock have been exposed by excavation of the in situ regolith. Flared slopes evolve in two stages, the first involving subsurface weathering during which the hillslope is undermined. The tectonic 107 Journal of the Royal Society of Western Australia, 80(3), September 1997 homogeneous rock Pit Frequency of water/ bedrock contact in pit USMi indurated surface laminated rock + + + + + + + + + + + + + + + Pan homogeneous rock + + + T + + + + + + + + + + Cylindrical basin homogeneous rock sheet joint Figure 7. Schematic development of pits, pans, cylindrical gnammas and armchair shaped rock basins (After Twidale 1982). stability of southern Australia has allowed time for intense scarp foot weathering to develop. The near surface zone dries out in summer, the dry season in southern Australia, but moisture and hence weathering persist at depth. The second stage involves the stripping of the regolith to expose the smooth, concave weathering front (Twidale & Campbell 1993). Tafoni Tafoni are hollows formed at the base of boulders and sheets and protected by an outer rim or visor (Fig 11). Some are associated with basal fretting or with flared slopes, suggesting that they too are initiated at the weathering front. Tafoni evolve as inverted saucers and enlarge upward into the rock mass as a result of salt crystallisation, hence the occurrence of tafoni in arid and semiarid lands, in cold and arid Antarctica and also in some coastal areas. The preservation of the outer visor is an integral part of tafone development and is possibly due to concentrations of iron or silica. Tafoni are inevitably self destructive, for their very growth helps destroy the host mass in which they are located (Twidale 1982). They may also result from gravitational forces imposed by a large boulder resting on an outcrop and causing crystal strain at the point or points of contact, eventually leading to rock basins on the outcrop and tafoni on the underside of the boulder (Vidal Romani 1989, 1990). 108 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 8. Dumonte Rock, near Wudinna, Eyre Peninsula, South Australia. The regolith has recently been cleared to make a reservoir, and the former weathering front has been exposed. The gutters continue beneath the former soil level, X-X (CR Twidale). Figure 9. Undercut side walls of gutters on Wudinna Hill, Eyre Peninsula, South Australia (CR Twidale). Rock levees and rock doughnuts Some minor forms are best explained in terms of the partial exposure of the bedrock surface. Rock levees are residual rims bordering shallow channels or gutters scored in bedrock. Rock doughnuts are annular rims encircling basins (Fig 12; Blank 1951a,b; Twidale & Bourne 1977). Protection by organisms and a coating of opaline silica (Whitlow & Shakesby 1988) have been suggested to account for their preservation. They can also be explained by the contrasted behaviour of wet and dry granite (Barton 1916; Twidale 1988). Little weathering takes place in the seasonally dry regolith-free zones adjacent to the gutter or basin though weathering proceeds through all seasons beneath the regolith preserved over most of the surface. Speleothems Although the products of granite weathering form important components of sediments, constructional forms are rare in the granite context. Silica released by weathering is, in places, reprecipitated along fractures which are consequently sealed. After weathering of the 109 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 10. Inverted grooves on the flared basal slope of Yarwondutta Rock, Eyre Peninsula, South Australia. The gutters which drain from basins on the hill bifurcate at the top of the flared slope into two grooves. The central rib of each is covered with algal slime (CR Twidale). Figure 11. Tafoni developed beneath and within sheets 3 m thick at Ucontitchie Hill, Eyre Peninsula, South Australia (CR Twidale). 110 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 12. Rock doughnut, a raised rim encircling a small pan on Waddikee Rock, Eyre Peninsula, South Australia (CR Twidale). surrounding granite, the silica forms a boxwork pattern of miniature ridges. In some areas silica from water seepages in open fractures is reprecipitated as small (up to 5 mm high) speleothems, most commonly in the form of flowstone, stalagmites or stalactites (Caldcleugh 1829; Vidal Romani & Vilaplana 1984). The stems of the speleothems are composed of opal-A but, curiously, each has a tip of gypsum. Figure 13. An arched slab of granite 15 cm thick with a crestal fracture and termed an A-tent, on Freeman Hill, western Eyre Peninsula, South Australia (CR Twidale). Ill Journal of the Royal Society of Western Australia, 80(3), September 1997 A-tents Tectonic processes are responsible for a small but notable suite of landforms in granite. A-tents, or pop- ups, (Fig 13) involve a permanent expansion, are consistently oriented in a given area and, in some instances as at Quarry Hill (Eyre Peninsula), have been induced by detonation of explosives. They have been attributed to insolation and to erosional unloading but their consistent orientation suggests they are best explained as associated with the release of compressive stress, in natural conditions probably in response to earth tremors (Twidale & Sved 1978). Conclusion Many of the landforms developed on granite, both major and minor, are related to the characteristics of the rock, the composition of the penetrating water, and the relative amounts of weathering and erosion. Depositional forms are rare. Tectonic factors are significant in a suite of forms due to the release of compressive stress. References Barton DC 1916 Notes on the disintegration of granite in Egypt. Journal of Geology 24:382-393. Blank HR 1951a "Rock doughnuts", a product of granite weathering. American Journal of Science 249:822-829. Blank HR 1951b Exfoliation and granite weathering on granite domes in central Texas. Texas Journal of Science 3:376- 390. Brook GA 1978 A new approach to the study of inselberg landscapes. Zeitschrift fur Geomorphologie Supplementband 31:138-160. Caldcleugh A 1829 On the geology of Rio de Janeiro. Transactions of the Geological Society of London 2:69-72. Campbell EM & Twidale CR 1991 The evolution of bornhardts in silicic volcanic rocks in the Gawler Ranges. Australian Journal of Earth Sciences 38:79-93. Campbell EM & Twidale CR 1995a Lithologic and climatic convergence in granite morphology. Caderno Laboratorio Xeoloxico de Laxe (Coruna) 20:381M03. Campbell EM & Twidale CR 1995b The various origins of minor granite landforms. Caderno Laboratorio Xeoloxico de Laxe (Coruna) 20:281-306. Colman SM & Dethier DP 1986 Rates of chemical weathering of rocks and minerals. Academic Press, Orlando. Fry E 1927 Mechanical action of crustaceous lichens on substrate of shale, schist, gneiss, limestone and obsidian. Annales of Botany 41:437-460. Goldich SS 1938 A study of rock weathering. Journal of Geology 46:17-58. Hassenfratz J-H 1791 Sur ^arrangement de plusieurs gros blocs de differentes pierres que Lon observe dans les montagnes. Annales de Chimie 11:95-107. Hill SM 1996 The differential weathering of granitic rocks in Victoria, Australia. AGSO Journal of Australian Geology and Geophysics 16:271-276. Hurault J 1963 Recherches sur les inselbergs granitiques nus en Guvane Fran^aise. Revue de Geomorphologie Dynamique 14:49-61. King LC 1942 South African Scenery: A Textbook of Geomorphology. Oliver & Boyd, Edinburgh. King LC 1962 Morphology of the Earth. Oliver & Boyd, Edinburgh. Lageat Y Sellier D & Twidale CR 1994 Megaliths et memorisation des granites en Bretagne littorale, France du nord-ouest. Geographie Physique et Quaternaire 48:07-113. Lamego AR 1938 Escarpas do Rio de Janeiro. Departamento Nacional da Produgao Mineral (Brazil) Servigo Geologico e Mineralogies Boletim 93, Mabbutt JA 1961 "Basal surface" or "weathering front". Proceedings of the Geologists' Association (London) 72:357-358. Mustoe GE 1982 The origin of honeycomb weathering. Geological Society of America Bulletin 93:108-115. Pope GA 1995 Internal weathering in quartz grains. Physical Geography 16:315-338. Russell GA 1935 Crystal growth and solution under local stress. American Mineralogist 20:733-737. Seager WR 1981 Geology of Organ Mountains and southern San Andres Mountains, New Mexico. New Mexico Bureau of Mines and Mineral Resources, Socorro. Memoir 36. Twidale CR 1962 Steepened margins of inselbergs from north¬ western Eyre Peninsula, South Australia. Zeitschrift fiir Geomorphologie 6:51-69. Twidale CR 1980 The Devil's Marbles, central Australia. Transactions of the Royal Society of South Australia 104:41-49. Twidale CR 1982 Granite Landforms. Elsevier, Amsterdam. Twidale CR 1988 Granite landscapes. In: The Geomorphology of Southern Africa (eds BP Moon & GF Dardis). Southern Book Publishers, Johannesburg, 198-230. Twidale CR 1993 The research frontier and beyond: granitic terrains. Geomorphology 7:187-223. Twidale CR & Bourne JA 1975a Episodic exposure of inselbergs. Geological Society of America Bulletin 86:1473-1481. Twidale CR & Bourne JA 1975b The subsurface initiation of some minor granite landforms. Journal of the Geological Society of Australia 22:477-484. Twidale CR & Bourne J A 1976a Origin and significance of pitting on granitic rocks. Zeitschrift fiir Geomorphologie 20:405-416. Twidale CR & Bourne JA 1976b The shaping and interpretation of large residual granite boulders. Journal of the Geological Society of Australia 23:371-381. Twidale CR & Bourne JA 1977 Rock doughnuts. Revue de Geomorphologie Dynamique 26:15-28. Twidale CR & Bourne JA 1978 A note on cylindrical gnammas or weather pits. Revue de Geomorphologie Dynamique 26:135-137. Twidale CR & Campbell EM 1984 Murphy Haystacks, Eyre Peninsula, South Australia. Transactions of the Royal Society of South Australia 108:175-183. Twidale CR & Campbell EM 1993 Australian Landforms: Structure, Process and Time. Gleneagles Publishing, Adelaide. Tw'idale CR & Corbin EM 1963 Gnammas. Revue de Geomorphologie Dynamique 14:1-20. Twidale CR & Sved G 1978 Minor granite landforms associated with the release of compressive stress. Australian Geographical Studies 16:161-174. Vidal Romani JR 1989 Geomorfologia granitica en Galicia (NW Espana. Cuadernos Laboratorio Xeoloxico de Laxe (Coruna) 13:89-163. Vidal Romani JR 1990 Formas menores en rocas graniticas: un registro de su historia deformativa. Cuadernos Laboratorio Xeoloxico de Laxe (Coruna) 15:317-328. Vidal Romani JR & Vilaplana JM 1984 Datos preliminares para el estudio espeleotemas en Avidades graniticas. Cuadernos Laboratorio Xeoloxico de Laxe (Coruna) 7:305-324. Viles H 1988 Biogeomorphology. Blackwell, Oxford. Whitlow R & Shakesby RA 1988 Bornhardt micro¬ geomorphology: form and origin of micro-valleys and rimmed gutters, Domboshava, Zimbabwe. Zeitschrift fiir Geomorphologie 32:179-194. Yatsu E 1988 The Nature of Weathering. Sozosha, Tokyo. 112 Journal of the Royal Society of Western Australia, 80:113-122, 1997 Granite outcrops: A collective ecosystem B York Main Department of Zoology, University of Western Australia, Nedlands Western Australia 6907 email: bymain@cyllene.uwa.edu.au Abstract Ecological systems of granite outcrops are posed as having arisen through one of two processes or through a combination of both processes; through colonisation of exposed and weathered rock surfaces and/or through retention of components of relict biotic assemblages surrounding such exposures. In the context of such evolved ecological systems, the inter-relationships of outcrop configuration, the geological and climatic history and associated changes from mesophytic to sclerophyll or xeric vegetation of surrounding landscapes is discussed with reference to selected outcrops. Introduction Ecosystem, like biodiversity, is an "in" word. But what does it mean? As an abbreviation of ecological system, it denotes some sort of interaction of life forms and dependence within a physical, non-biological framework on which the living forms in turn have some effect. As a system, an inherent cohesion is intimated; external limits are inferred. All this in turn suggests that an ecosystem has boundaries. A dictionary supports this concept in the definition " The plants and animals of a particular habitat , together with the environment influenced by their presence" (Onions 1978). Ecosystems, as nominated in common biological parlance, can be large or small, relatively open or near-closed. An ecosystem can range from the complexity of a tropical rainforest to an individual tussock grass on Beauchene Island in the Falkland Islands archipelago (Smith & Prince 1984). How then to define a granite outcrop ecosystem? It must comprise some sort of cohesive biotic community supported by a physiographic matrix, which embraces both time and space components and be clearly demarcated from any adjacent, contiguous or surrounding ecosystems. Granite outcrops were referred to early as monadnocks or "island hills" (Jutson 1934) who recognized them as exposed residuals surrounded by more readily weathered components of the underlying bedrock. They are currently regarded popularly as "islands" (Anon, undated). In the context of "islands" many botanical studies have been undertaken in southwestern Western Australia (Ornduff 1987). Current understanding and interpretation of the peculiarites of the biota of the outcrops has been partially wrapped in island biogeography theory. The obvious refugial role of the rocks has also been noted in relation to isolates of former more widespread and continuous distributions of plants such as jarrah (Abbott 1984) and also of various animals. Recognition of both the refugial and insular nature of granite outcrops has underpinned much of the drive and justifeation for conservation and © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 inclusion of particular rocks in the nature reserve system of Western Australia. The "island" concept of outcrops has been generally stimulated by observation of their present biotic distinctiveness or biotic disjunction with their immediate surrounds, particularly in relation to vegetation and ephemeral aquatic fauna. While documentation of vegetation, vertebrates and aquatic invertebrates has already been undertaken on selected granite outcrops (see elsewhere in this issue), the terrestrial and/or lithic invertebrates are less well known. Few, if any, comprehensive studies have attempted to document the interactive relationships of the total biota of a granite outcrop although Main (1967) presented a simplistic naturalist's view of the ecology of a particular rock, Yorkrakine Rock, in the wheatbelt. McMillan & Pieroni (undated) have given a "primer" account of the life forms associated with granite outcrops. To understand the representativeness and cohesion of the biota on any one outcrop and to be able to arrive at generalisations regarding the composition of the biota over the array of outcrops in any region or on a continental scale, I believe one needs to adopt an historical perspective. While documenting (a) the characteristics of the outcrops as they appear now and (b) the present biota, one needs at the same time to ask; • How have these "ecosystems" come about ? • What were the antecedent landscapes like? • How has the biota responded to the sequential changes of the landscapes through to the present time? I shall now attempt to show that by adopting this historical approach we can come to some understanding of • the physiography of the outcrops, • the distribution and restriction of dominant elements of the biota, • "ecosystems" of selected rock configurations, and • over a geographic span suggest that there are biotic patterns peculiar to granite outcrops which comprise a "collective ecosystem". 113 Journal of the Royal Society of Western Australia, 80(3), September 1997 From the outset, my discussion will be limited largely to the granite outcrops of southern Western Australia, south of approximately 26 degrees latitude, and hence those which occur in a predominantly winter rainfall region. In the more northern and eastern areas particularly, they are also affected by summer rainfall in the form of thunderstorms. Outcrops near the south coast experience the benefit of coastal precipitation either as light drizzle or fog induced by onshore winds associated with high pressure systems. The relevance of the overall discussion to outcrops elsewhere, particularly Eyre Peninsula, is implicit. Distribution and Characteristics of Granite outcrops The present scenario Granite outcrops, as loosely defined, are scattered throughout southern Western Australia from north and west of the Nullarbor i.e. from the arid interior, through woodlands and forest country to the southwest coast. They appear as emergent protuberances of the basement rock over the Great Plateau or Archaean Shield (of the Yilgam craton) and similarly as promonotories on the dissected edges of the western escarpments and across the southern landforms of the state. Jutson (1934) referred to these "island hills" as remains of the "old plateau". Clarke (1994) dated the "stripping" of the regolith of much of the Yilgarn craton (which would account for "granite" exposures) as between Middle Jurassic and early Eocene and associated such stripping with uplift and incision of the palaeodrainage system. It is outside the scope of this paper to discuss the petrology or origin of the rocks which aspects are dealt with elsewhere in this issue. It is only relevant here to discuss the physiographic configurations of the rocks as they affect the biota and its distribution. Widely scattered as they are, the rocks fall within three major rainfall zones as defined by Hopper (1979) as the high, transitional and arid rainfall zones representing 800-1400, 300-800 and less than 300 mm per annum. Interestingly, rocks within these three zones can be more or less defined by their grey, brown or red colour respectively which reflects indirectly the response of living organisms e »g. types of predominant lichen growths, or conversely their absence, to the rainfall gradient. The composition of the ecosystems of particular rocks is determined partly by the present rainfall i.e. the zones as delimited above, by the geo-climatic history with the associated changes from a mesophytic to sclerophvll and xeric vegetation and the configuration and topography of the rocks. Configuration and topography of rocks Configuration of rocks ranges from high, sharply declivous domes to barely perceptible, flat exposures. Such configurations generally reflect the degree of erosion of the surrounding landscape which in turn has exposed the basement rock to varying degrees. A convenient characterisation of the predominant configurations adopted here (Fig 1) is as follows: • Single or multiple domes, monadnocks or tors, as "islands" on the dissected Darling Plateau and in a landscape of low relief e.g. the predominant outcrops of the Great Plateau ("inselbergs" of Jutson (1934); Fairbridge & Finkl (1978); Fink! & Churchward (1974) and in part "bornhardts" of Campbell (1997). • Flat, often disc-like pavements (Fig 1) which are also a frequent feature on the plateau. • Subsurface basement highs, here referred to as "fugitive outcrops" (Fig 1). • Asymmetric declivities or tors as exposures in mountain ranges e.g. the Porongurup Range and throughout the forested southwest of Western Australia (comparable to the domes of The Plateau). • Coastal declivities and near-coastal knobs protruding through Pleistocene limestone deposits. • Tumuli of boulders either around a central core or as rock "piles" (Fig 1; Plate 1) . The latter are sometimes adjacent to salt lake lineages where they may be erosional features of former domes which have weathered to blocks in association with the presence of salt. Such rock piles are comparable in part to "nubbins" and "castle koppies" of Cambell (1997). • Scattered boulders (especially on escarpments or ridges), or clusters of small tors, or "haystacks" (Twidale & Campbell 1984). It is the domes (monadnocks or tors) and the disc-like pavements which most closely fit the general concept of an "outcrop" i.e . as a sizable, isolated "intrusion" into the landscape (Ornduff 1987) and it is predominantly these configurations which will form the main focus of my discussion with reference to others as listed above for comparison. However, all types of configurations share some basic ecological elements while differing in structural and biotic detail. A combination of rock topography, which has been partly determined by past erosion and climatic patterns, and impacts of current weather conditions determines the level of complexity of the ecosystem of a particular outcrop. Topographic features and associated biota of outcrops The more subdued the profile of an outcrop the less complex the topography hence a "pavement" rock generally presents little sculpturing whereas a higher exposure exhibits a more varied topography (Fig 1). The main topographic features of an outcrop are associated with the sculptured surface and the surrounding "apron". The sculptured surface. The sculptured surface may be relatively "smooth" or pitted and dimpled (Jutson 1921). As well as the rock crust having partially detached pieces, including "pop-ups" or "A-tents" (sec Campbell 1997), the surface may also be littered with exfoliated plates, flakes or slabs. Both attached and loose rock plates provide habitat for invertebrates e.g. spiders, centipedes, scorpions, pseudoscorpions, opiliones, pupating insects, moths, beetles and also vertebrates particulary frogs and lizards such as geckoes and the agamid Ctenophorus ornatus. The wave-like slopes ("flared slopes" of Campbell 1997) of some higher domes have been variously interpreted as the result of subsurface weathering in the presence of water (see Campbell 1997) or partly by 114 Journal of the Royal Society of Western Australia, 80(3), September 1997 DOME PAVEMENT Figure 1. Examples of configurations and topography of granite outcrops. sandblasting by wind. However some initial sculpturing may have been by the Permian ice sheet which is known to have covered much of southwestern Western Australia (McWhae et al 1958). Earlier, Clarke (1919) had remarked on glacial deposits in inland southern Western Australia as have Fairbridge & Finkl (1978). However, Clarke (1994) doubts that any relicts of Permian landforms could be anything but minor and restricted today. The "waves" form dramatic drops and frequently direct rain onto either surrounding aprons or into flat bottomed gorges from which creeks may debouch. Both such sites provide moisture-holding habitats. Certain outcrops weather into block formations (Jutson 1934) and have large boulders both on their profiles and scattered around the base (see "nubbins" of Campbell 1997). These boulders sometimes erode into cave-like shells (tafoni) or narrow-based pedestal shapes 115 Journal of the Royal Society of Western Australia, 80(3), September 1997 (see Campbell 1997). Runoff water from the boulders maintains a moister habitat than the general rock surface and where there are clusters of boulders, intervening soil supports ephemeral and perennial vegetation. The boulder piles or clusters (Plate 1) form habitats for animals ranging from invertebrates to vertebrates such as bats and marsupials (euros and rock wallabies) and snakes and varanid lizards. "Fugitive" outcrops (Fig 1) of the basement rock often appear as open meadow-like spaces surrounded by higher shrubby vegetation, woodlands or forest. They are vegetated by the same fringe or meadow plants as occur on rocks in the same region. As summer-dry bogs they provide habitat for relict burrowing spiders e.g. the mygalomorph Teyl (Withers & Edward 1997, Fig 2) and various insects such as cicindellid beetles. Algae and lichens occur on most rock surfaces. Lichens are renowned for being able to withstand extremes of temperature. Pertinent to growth and survival on granite outcrops is the fact that dry thalli of some genera have been demonstrated to resist temperatures as high as 70-101 °C. However, the tolerance of moist thalli does not exceed 46 °C (Lange cited by Hale 1967). Some lichens are very long lived but they are also colonisers e.g. Parmelia occurs abundantly on the rocks. Lichens have the capacity to fix nitrogen and in the presence of rain to leach minerals from the rock surface. In so doing, they gradually break down the rock surface forming soil particles. Thus their role in making granite outcrops suitable for colonisation by mosses and finally other plants is highly important (Plate 2). While no studies appear to have been conducted in southern Western Australia on the possible association of organisms such as invertebrates living in the thalli, certain moth larvae feed on lichens on the outcrops (pers obs). In wet areas, considerable moss mats or swards occur where a little soil has accumulated. Larger rocks with higher profiles frequently have deep crevices which similarly support moss mats and rock ferns (the most common and widespread being Cheilanthes) and also the blind grass Stypandra and various shrubs e.g. Baeckea, Beaufortia, Kunzea or Anthocercus depending on the location. Kunzea, which is common on dryer rocks, characteristically also grows from deep narrow fissures and thus gives the appearance of hanging from the rock itself. In contrast Anthocercus which grows in higher rainfall areas grows in the moss mats or swards and consequently often suffers wind throw (Plate 3). Higher rocks may also have shelves with an accumulation of soil where again moss mats, ferns and shrubs flourish. Newbey (1995) remarked on deposits of "skeletal soil" up to 30 cm deep on lowlying areas of granite exposures. These deposits may also provide habitat for perennial mygalomorph spiders. The interface of rock and soil of crevices and shelves is frequently fringed with the liliaceous pincushion plant Borya in dryer areas, and in wetter areas the heather-like Andersonia. Borya , as one of the "resurrection" plants (Gaff 1981) is able to withstand considerable desiccation during which times it assumes a striking orange colour (Plate 2), but reverts quickly to a bright green following water uptake after rain (see Plate 5, and Hopper et al. 1997). A variety of orchids, composites and other ephemerals intermix near these fringes. Various invertebrates such as insects and spiders live in any soil accumulations e.g. of crevices and shelves, which support plant life. The gorges, in the higher rocks, support the same sort of vegetation as the shelves with occasionally the addition of the Christmas tree (Nuytsia floribunda), Allocasuarina huegeliana and Acacia rostrata . Large depressions ("dimples" of Jutson 1921; and in part "flat-floored pans" of Campbell 1997) occur on most outcrops which have at least some relatively flat areas. After rain these depressions fill with water to form ephemeral pools. The quillwort Isoetes frequently forms a dense lawn on the floor of such ponds. An array of invertebrates including the crustaceans Conchostraca (shield shrimp Triops ), Cladocera, Ostracoda, Anostraca (brine shrimp), insect larvae, water mites, flat worms (Platyhelminthes), rotifers and also frog tadpoles are active seasonally in the ponds. Rocks with higher topography encourage a cascading effect of runoff after rain. This runoff may erode the rock surface into tiers of pit-like depressions ("dimples"), of which the deeper ones are sometimes called "arm-chair basins", along drainage lines thus forming temporary waterfalls (Jutson 1921) which leave in their wake a series of longer lasting pools. A later serai stage of such stepped pools or water holes are soil-filled pockets which finally develop into "Babylonian gardens" (Plate 4). Gnammas (rock holes) may be present on both dome and pavement configurations and are frequently permanent water reservoirs. As well as habitat for invertebrates, the pools and gnammas are an important source of water for vertebrates, both permanent habitat associates and transitory fauna e.g. birds and also bats (R How, pers comm ; Dell & How 1984, p 69) and larger marsupials. The meadows, formed over infilled ephemeral ponds, are a common feature on outcrops (Plate 5). Depending on the serai stage these are crusted with lichens or vegetated by moss mats, Borya or Andersonia heaths, blind grass and restionaceous tussocks and small shrubs such as may also occur on shelves or in crevices of the outcrop. In that these meadows form seasonal bogs, usually in winter but also during summer thunderstorms, they provide important habitats for moisture dependent organisms particularly invertebrates including species of the mygalomorph spider Teyl. Fissures, crevices, shelf meadows and seral-pond meadows, boulder shadows where soil accumulates, lichen and moss around the lips of fixed rock plates, and even the soil spill around exfoliated slabs, all form to some degree relatively moist habitats when compared to the general rock surface. As such they form the principal habitats for smaller organisms, both invertebrates and vertebrates including frogs and lizards, while wallabies take advantage of larger boulder accumulations. The apron. The descending edges of outcrops frequently fan out into soil covered aprons of varying width depending on the declivity of the rock. Soil depth varies from 1.5 to 2.0 m (Newbey 1984 p33, 1988 pi 1 ) and throughout the seasons are alternately waterlogged and dry, thereby constraining both the vegetation and soil living invertebrates. Shallow soils with little vegetation apart from moss mats and rock fern, and meadows of Borya and ephemerals or in wetter areas 116 Journal of the Royal Society of Western Australia, 80(3), September 1997 Plate 1. A rock pile (tumulus) near Hyden, Western Australia, with surrounding Allocasuarina grove. Plate 2. A meadow of Borya at edge of rock with adjacent lichen crusts on rock, Plover Rock, Western Australia. Note the "drought" mode of the brightly coloured Borya. Plate 3. Wind-thrown Anthocercus plant on Torbay Hill, Western Australia. Plate 4. "Babylonian Gardens" growing in the pockets of soil formed in the tiers of infilled rock holes (armchair basins) on Torbay Hill, Western Australia. This stepped sequence of soil pockets would formerly have been a seasonal waterfall. Plate 5. A tiny Borya meadow formed in an infilled rock pool on a rock at Payne's Find, Western Australia. Such meadows provide habitat for the mygalomorph spider Teyl and other invertebrates. Plate 6. The stem-flowering ("cauliflory") of Hakea petiolaris, a condition possibly relictual from an earlier climatic period when the plants grew under a canopy. 117 Journal of the Royal Society of Western Australia, 80(3), September 1997 Andersonia, gradually blend into tussock and shrub thickets similar to the vegetation of the larger meadows and shelves on the rock proper or in the shallow gorges with the addition of sandal wood. In the semi-arid and arid region, the aprons may have an extension of indiginous grasses from park-like acacia groves. Allocasuarina shrubs and trees frequently fringe the outer extension of aprons or grow along creeks at the interface of rock edge and apron. Particularly on higher rocks which provide much runoff, the creeks at either the boundary of the apron or where they transect the apron are important seasonal habitats for invertebrates and frogs and in summer provide damp refuge areas. The apron meadows, like the meadows formed in infilled ponds and over "fugitive" outcrops, are sanctuary to relict burrowing spiders e.g. Teyl species which aestivate in sealed burrows, and insects whose larvae are dependent on a seasonally wet substrate. Microclimate of outcrops and effects on biota Outcrops, particularly in the semiarid and arid regions are subject to extremes of temperature and alternatively wet and dry seasonal conditions. Botanical studies and zoological works have both noted upper rock surface temperatures of over 50 °C (Marchant 1973; Bradshaw & Main, 1968). Bradshaw & Main (1968) also noted that air temperatures were at least ten degrees lower under the overhangs of attached rock plates and even under loose slabs. Animals such as the common ornate dragon lizard, Ctenophorus ornatus , take advantage of these retreats and also the shade cast by boulders. The latter also provide refuge from the heat for rock wallabies and euros. Presumably the large goannas also benefit from boulders and large semi- attached rock sheets. Rocks and rock slopes of coastal hills, the southern mountains and southwest peaks e.g. Porongorups, Granite Peak, Mt Lindsay, Many Peaks and peaks in the Darling Ranges such as Mt Cooke, all benefit from the altitudinal self-generated cloud and fog mantles which increase the moisture content of vegetation, litter and soil thus providing microhabitats for invertebrates. Conversely many rocks in regions of low rainfall while capturing isolated fog nevertheless do harvest moisture and thereby provide refuge for moisture dependent biota. It has been noted that even light rain generates run off from rocks. In winter 5 points (1.8 mm) of rain, in summer 7 to 9 points (2. 5-3. 2 mm) generates run off. More significantly it is recorded that "Kondinin Rock has been known to yield water .... from a light mist" (Fernie 1930). Such harvested water naturally channels into creeks at the rock/apron interface, around the base of boulders and into crevices and fissures and under overhangs and exfoliated slabs. It is also sufficient to wet dried beds of ponds and activate dormant insects and aestivating frogs. Thus, even minimal moisture catchment becomes of major importance for the survival of relict invertebrates. Origin and Distribution of the Biota Historical perspective and persistence The outcrops appear as present day intrusions in the surrounding landscape (Ornduff 1987) but they would not in earlier geoclimatic periods have stood out in such isolation. Although their lithic core would have set them apart topographically, a much wetter climate and mesophytic vegetation containing southern rainforest elements including Nothofagus and associated plants (Balme & Churchill 1959) would have maintained more continuity in the vegetation and fauna. Biotic isolation focused around granite outcrops would have begun in the Tertiary along with the sclerophylly of the vegetation. Outlying botanical components of the forest region of southwestern Western Australia associated with granite outcrops in the dryer transitional rainfall zone, such as the jarrah (Eucalyptus marginata) stand at Jilakin Rock (Abbott 1984; Churchill 1968) indicate the contraction of species to refuges in the face of changing climatic conditions. Another possible relict of an earlier bio- climatic scenario is Hakea petiolaris . This species exhibits the peculiar habit of stem-flowering or "cauliflory" (Plate 6), which White (1986) describes as characteristic of certain plants growing under a closed canopy- Thus it is possible that H. petiolaris is a relict of an earlier, wetter, forested landscape. With a continued trend towards a dryer, more seasonal climate, invertebrates dependent on a continuously moist habitat would have become concentrated in microhabitats such as already noted around granite rocks, while many (as with larger vertebrates) may have disappeared altogether. Some animals show definite adaptive responses through life history strategies to increased aridity and marked seasonal changes by retreating to permanently moist microhabitats at certain times of the year. Others, such as the aquatic invertebrates (Crustacea, insect larvae, platyhelminths, etc) are well suited through dormant life history stages to avoid severe summer drought. For example, many branchiopod crustaceans have resistant eggs and can remain viable in the dry soil of ponds for long periods. There appear to be few animals tied completely to rock habitats throughout their entire life cycle. The rock wallaby, although dependent for shelter on rocks, is more widespread and forages away from rocks. In some places euros similarly make use of rocks for shelter. Many other vertebrates such as kangaroos, echidnas, bats and a whole suite of birds make transitory use of granite outcrops especially as a water source and thus are probably dependent on them during summer drought but are not confined to them. Many frogs live and breed around rocks, but are not soley restricted to granite outcrops as a habitat. A notable vertebrate which is confined to granite outcrops is the widespread dragon lizard, Ctenophorus ornatus, which thus represents a collective of disjunct populations. Similarly a few plant species have a fragmented distribution with occurrences confined to granite rock habitats e.g. Isotoma (James 1982), Eucalyptus caesia (see Hopper et al. 1982), £. crucis (see Brooker & Kleinig 1990), Kunzea pulchella, and some orchids. Certain invertebrates likewise are tied to granite outcrops. The biogeographically-ancient midge Archaeochlus, whose larvae develop in seeps on granite rocks, is an example (Cranston et al. 1987). Nevertheless, over the range of the genus in southwestern Australia and extending to Central Australia (P Cranston, pers comm) three species are 118 Journal of the Royal Society of Western Australia, 80(3), September 1997 included. Most branchiopod crustaceans of the ephemeral rock pools occur on scattered outcrops. A well known example is the shield shrimp Triops which occurs not only on rocks but also along the edges of lakes and in temporary puddles in lowlying areas; there is no site specificity because of the wind-borne dispersal of the resistant eggs throughout the dry inland. The same probably holds for most aquatic invertebrates, except possibly for some water mites particulary in the forested south-west. Various insects such as lichen-eating moth larvae and possibly some of the moisture-loving grasshoppers are possibly restricted to granite outcrops. Amongst spiders, a Rebilus species (family Trochanteridae) occurs on many inland rocks where it lives under loose and attached slabs. Some other spiders with similarly flattened body form such as an unnamed lycosid, previously attributed to Pardosa (Main 1976, Plate 25) is confined to inland granite rocks while Hemicloea species occur under slabs on granites in southern forest areas but it is doubtful whether it or Rebilus are confined to rock habitats. Selenops, although a rock inhabiting genus, is not restricted to granite but is found also on sandstone formations. Other spiders frequently found on granites include wolf spiders such as Lycosa leackhartii, Miturga species and the redback spider Latrodectus hasselti. All are readily dispersed and not confined to granites. Site specificity In contrast to this pattern, an interesting mygalomorph spider genus, Teyl, which is widespread in south-west Western Australia (and occurs in restricted areas of Eyre Peninsula and western Victoria), appears to be one of the few genera of animals which has at least some species restricted to particular granite outcrops i.e. species that are site specific. A striking confirmatory example is an undescribed species from Paynes Find Rock in which the spider exhibits an aberrant carapace modification (an everted fovea). This peculiarity does not occur in species of neighbouring rocks and indicates development of a morphological feature in isolation. The spiders are extremely sedentary, do not readily disperse and are active for only a very short period of the year. The spiders live in burrows in the meadows (Plates 2 & 5) and apron soils of the rocks and seal the burrows when the soil begins to dry out. At least one species of pseudoscorpion Synsphyronus elegans, which lives under rock slabs, appears to be site-specific to Yorkrakine Rock (north of Tammin). While there are few data on either endemicity to granite outcrops and/or rock-site specificity of invertebrates there is apparently considerable botanical endemicity of particular rocks (see Hopper et al 1997). The lack of invertebrate data may be giving a false impression of low endemicity. Biogeographic patterns: relative ages Those invertebrate species which are confined to particular rocks probably indicate an older biogeographic phenomenon than those species which have not speciated in response to habitat isolation. I suggest this speciation began with the change to sclerophylly and seasonality in the early and mid-tertiary which would have induced a concentration of moisture loving organisms around the granite rocks, not because of any special affinity for rocks per se but because of their capacity to provide wet microhabitats in an otherwise dry terrain. This moisture holding capacity of areas around substantial granite outcrops also holds good for "fugitive rocks". Main (1996) noted the occurrence of Mesozoic genera of mygalomorphs such as Teyl in microhabitats over granitic "basement highs" on relictual parts of the old Tertiary plateau of the wheatbelt of southern Western Australia. The model she presented could explain the distribution of many archaic invertebrates associated with higher parts of the inland plateau which in turn embraces many granite outcrops. A richer fauna of both vertebrates and invertebrates must have existed during the mid-to-late Tertiary and into the Pleistocene, similarly focused around granite outcrops, but some of which has become extinct along with increased aridity. Fossil deposits of diprotodonts and other vertebrates at Balladonia (Glauert 1912) from a dam excavation adjacent to granite domes indicates just such a concentration (even if only seasonal) as we now see for certain extant organisms. Such species may also have been associated with other granite rocks which did not provide the same extensive boggy conditions suitable for fossilisation. At best, the granite outcrops currently preserve a partially relictual, but also a predominantly widespread although contracted and fragmented biota. Nevertheless, high botanical endemicity has been demonstrated for selected taxa which have been extensively studied. This suggests that the invertebrate fauna requires comparable in depth surveys, particularly of those forms with sedentary behaviour. Aeolian reinforcement The role of lichens and algae in both physically breaking down a "sterile" rock surface and also "capturing" atmospheric nutrients, primarily nitrogen, and thereby making a substrate or creating an environment for other organisms, has already been noted. An additional primary and continuous source of biological enhancement of granite outcrops is prevalent in what may be termed the aeolian medium. Swan (1992) defined an "aeolian zone" as "a region around and beyond the limits of the flowering plants". He suggested that the combined aeolian zones of the world constitute an "aeolian biome". An expanded definition of the "aeolian biome" could readily embrace granite outcrops both during the initial stages of their colonisation and later, even when a considerable associated biota has developed. The isolated granite outcrops of south-western Western Australia and/or other parts of southern Australia provide staging posts or scattered points in a biological lattice. Aerially-borne nutrients, and organic and inert debris as well as propagules of plants and animals (as spores, seeds and resistant eggs, and various other life history stages of invertebrates, including Collembola, aphids, thrips and large-wringed insects) all combine to provide a continuous aeolian reinforcement of spatially isolated granite outcrops. The seasonal behaviour and dormancy of some of the invertebrate fauna has already been mentioned. Spiders 119 Journal of the Royal Society of Western Australia, 80(3), September 1997 NEEDS OF BIOTA IN USE OF GRANITE ROCKS Permanent I ▼ ■ Lichens/Plants ■ Teyl SUBSTRATE SHELTER WATER Permanent 1 dragon lizards 1 pseudoscorpions 1 spiders Partial 1 ■ rock wallabies ■ echidnas ■ mud wasps Permanent | Partial Transitory 1 i ■ aquatic T ■ tadpoles 1 ■ birds invertebrates ■ insect ■ bats larvae ■ kangaroos Figure 2. Selected examples illustrating the needs of the biota in their use of granite outcrops summarised according to their residency status in relation to provision by rocks of a substrate, shelter and water. Amongst the transitory fauna dependent on rocks for water, the present environment includes feral animals from rabbits and foxes to camels and donkeys. such as the poor disperser Teyl are dormant for at least six months of the year secluded in sealed burrows, but other organisms deposit eggs or encysted dormant stages in the soil or dried out beds of ephemeral ponds. These latter organisms may be dispersed aerially, hence both reinforcing otherwise isolated populations and inhibitng allopatric speciation. Not only are propagules of plants and invertebrates aerially dispersed but also adult insects may form part of the aeolian fauna thereby accounting for the general lack of rock specificity. Irregular but recurrent weather events such as cyclonic winds, tornadoes, dust storms and local " cock¬ eyed bobs" (Hunt 1929) each with their propensity for lifting, distributing and dumping fine biotic material, all reinforce the aeolian element of granite rocks. Thus, finally, although from present knowledge the biota (especially of the fauna) associated with granite outcrops appears to have a distinctiveness, it represents predominantly a collective of "refugees" rather than an array of separate biotas tied to particular sites. Nevertheless I believe the relict mygalomorph genus Teyl may be representative of many other invertebrates of which we are not aware. This example indicates the need for thorough comprehensive surveys of invertebrates on selected outcrops over a wide geographic range. Geographic replacements In spite of the taxonomic commonality over a wide range of rocks, especially at the generic level and in many cases the lack of species endemicity, there are some noticeable geographic replacements by ecological "equivalents" of certain rock adapted organisms, both plants and animals, or organisms centred around outcrops. For example Anthocercus of the southern forests is replaced by Kunzea pulchella in the wheatbelt, which in turn is replaced by a Beaufortia in the eastern Goldfields. Amongst spiders over the same geographic span, Hemicloea is replaced by Rebilus and a lycosid. Trapdoor spiders of the Aganippini similarly replace one another in the apron meadows and fringing vegetation of disjunct rocks. Although superficially such replacements may be loosely regarded as "equivalents", a better term might be "behavioural counterparts" since the replacements are clearly associated with the rainfall zones and as such, physiological and interactive differences must be associated with the taxonomic shifts. Further study of this phenomenon as demonstrated by a selection of taxa over a geographic span of rocks could well prove worthwhile. Conclusions The ecology of the biota can be summed up in the concept of the needs of the respective taxa in relation to what the rocks offer towards their livelihood (see Fig 2). The use of granite outcrops in satisfying the needs of their resident biota are primarily related to the functions offered by (a) the substrate i.e. the rock surface as a "home base" or attachment plane, (b) the rocks as shelter, and (c) the rocks as a water source. The biota can be grouped into permanent, partial or transitory residents according to the degree of association throughout the organisms' life histories. 120 Journal of the Royal Society of Western Australia, 80(3), September 1997 No one rock maintains a "closed" ecosystem. Although some rocks with complex configurations and diverse biota with a high degree of interaction as measured by botanic richness, habitat availability for fauna, and food chains, may tend towards self¬ containment they are in fact still "open" in the same sense that a cave ecosystem is. Cave ecosystems are dependent on residents (such as bats) which live partly outside the cave to bring in some nutrients and also on the accidental deposition, such as by floods, of detrital matter. Granite rocks similarly are dependent on the partial and transitory residents to enrich the ecosystem. In addition the propagules and non-biotic detritus supplied by the aeolian medium continuously enhance the viability of a rock ecosystem. Thus even those elements which have permanent residency have some dependence beyond the margins or physical boundaries demarcating a particular rock. Figure 2 summarises the relationship of a granite outcrop biota to the outcrop and the surrounding environment. In essence it can be stated that granite outcrops support a characteristic assemblage of plants and animals comprised of isolated populations of relictual genera and/or species, and widespread examples from the surrounding environment. There are many examples of granite outcrops, as isolates, retaining components of a once more widely distributed biota, for example the stand of jarrah trees at Jilakin Rock (Abbott 1984). With increasing sclerophylly of the vegetation associated with a drying of the climate since the mid-late Tertiary, the granite outcrops because of their water harvesting/ retention capacity have become the foci around which moisture-dependent organisms have concentrated. Nevertheless they support both sedentary isolates and mobile, wide ranging organisms as well as wind-borne biota. It is notable that apart from the more sedentary organisms and particularly elements of an older evolutionary radiation, such as the Mesozoic mygalomorph spider genus Teyl , most animal genera have not speciated in association with particular outcrops in spite of population fragmentation. Few organisms are endemic to particular rocks. However, there is evidence that there is a high level of genetic difference between populations of particular species from different rocks. Endemicity i.e. rock-specificity amongst plants, appears to be more pronounced, on present evidence, than amongst the fauna. In contrast to the taxonomic conservatism of much of the outcrop biota there are some interesting examples of geographic replacement of behavioural counterparts represented by unrelated taxa particularly over the rainfall gradient. Also of interest, many organisms although now confined to granite outcrops either entirely or for a greater part of their life history do not exhibit adaptations peculiar to "rock- living" (the agamid lizard Ctenophorns ornatus and certain spiders e.g. a lycosid are exceptions) but rather are confined now to rocks because of their refugial role in maintaining cool, moist habitats on which such species are dependent as a legacy from their earlier historical environmental association. In summary, a granite outcrop at the individual level maintains a co-operative of refugees, but on a broad scale granite outcrops represent a collective ecosystem. References Abbott I 1984 Ecological features of an outlying stand of jarrah ( Eucalyptus marginata ) at Jilakin Rock, Western Australia. Journal of the Royal Society of Western Australia 66:107-110. Anon undated Granite Outcrops: Living Rocks. Department of Conservation and Land Management, Perth. Balme B E & Churchill D M 1959 Tertiary sediments at Coolgardie, Western Australia. Journal of the Royal Society of Western Australia. 42:37-43. Bradshaw S D & Main A R 1968 Behavioural attitudes and regulation of temperature in Amphibolurus lizards. Journal of Zoology 154:193-221. Brooker M I H & Kleinig D A 1990 A field guide to Eucalypts Vol 2. Southwestern and Southern Australia. Inkata Press, Melbourne & Sydney. Campbell E M 1997 Granitic landforms. Journal of the Royal Society of Western Australia 80:101-112. Churchill D M 1968 The distribution and prehistory of Eucalyptus diversicolor F. Muell., E. marginata Donn. ex Sm., and E. calophylla R. Br. in relation to rainfall. Australian Journal of Botany 16:125-151. Clarke E de C 1919 Notes on the occurrence of boulders, possibly glaciated, near Leonora and Laverton. Journal of the Royal Society of Western Australia 6:27-32. Clarke J D A 1994 Geomorphology of the Kambalda region. Western Australia. Australian Journal of Earth Sciences 41:229-239. Cranston P S, Edward D H D & Colless D H 1987 Archaeochlus Brundin: a midge out of time (Diptera: Chironomidae). Systematic Entomology 12:313-334. Dell J & How R A 1984 Vertbrate Fauna. In: Biological Survey of the Eastern Goldfields of Western Australia. 2. Widgiemooltha-Zanthus area. Records of the Western Australian Museum. Supp 18:57-69. Fairbridge R W & Finkl C W 1978 Geomorphic analysis of the rifted cratonic margins of Western Australia. Zeitschrift fur Geomorphologie 22:369-389. Fernie N 1930 Water supplies from rock catchments in the Western Australian wheat belt. Journal of the Institution of Engineers of Australia 2:198-208. Finkl C W & Churchward H M 1974 The etched surfaces of southwestern Australia. Journal of the Geological Society of Australia 20:295-307. Gaff D F 1981 The biology of resurrection plants. In: The Biology of Australian Plants (eds J S Pate & A J McComb). University of Western Australia Press, Perth, 114-146. Glauert L 1912 Fossil marsupial remains from Balladonia in the Eucla Division. Records of the Western Australian Museum 1: 47- 65. Hale M E 1967 The Biology of Lichens. Edward Arnold, London. Hopper S D 1979 Biogeographical aspects of speciation in the southwest Australian flora. Annual Review of Ecology and Systematics. 10:399-422. Hopper S D, Campbell N A & Moran G F 1982 Eucalyptus caesia, a rare mallee of granite rocks from south-western Australia. In: Species at risk: research in Australia, (eds R H Groves & W D L Ride). Australian Academy of Science, Canberra, 46-61. Hopper S D, Brown A P & Marchant N G 1997 Plants of Western Australian granite outcrops. Journal of the Royal Society of Western Australia 80:141-158. Hunt H A 1929 Record of notable meteorological events in Western Australia: windstorms, (hurricanes, cyclones, heavy gales, duststorms etc). In: Results of rainfall observations made in Western Australia. Commonweath of Australia Bureau of Meteorology, Melbourne, 162-187. James S H 1982 The relevance of genetic systems in Isotoma petraea to conservation practice. In: Species at Risk: Research 121 Journal of the Royal Society of Western Australia, 80(3), September 1997 in Australia (eds R H Groves and W D L Ride). Australian Academy of Science, Canberra, 63-72. Jutson J T 1921 Note on the "dimpling" of granite hills in sub- arid Western Australia. Proceedings of the Royal Society of Victoria 33:108-111. Jutson J T 1934 The physiography (geomorphology) of Western Australia. Geological Survey of Western Australia, Bulletin 95:1-365. Main B Y 1967 Between Wodjil and Tor. Jacaranda & Landfall, Brisbane & Perth. Main B Y 1976 Spiders. Collins, Sydney. Main B Y 1996 Microcosmic biogeography: trapdoor spiders in a time warp at Durokoppin. In: Gondwanan Heritage, Past Present and Future of the Western Australian Biota (eds S D Hopper, J A Chappill, M S Haney & A S George). Surrey Beatty & Sons, Chipping Norton, NSW. Marchant N G 1973 Granite Outcrops. In: Flowers and Plants of Western Australia (eds R Erickson, A S George, N G Marchant & M K Morcombe). A H & A W Reed, Sydney, 199-202. McMillan P & Pieroni M undated Exploring Granite Outcrops. Department of Conservation and Land Management, Perth. McWhae J R H, Playford P E, Lindner A W, Glenister B F & Balme B E 1958 The stratigraphy of Western Australia. Journal of the Geological Society of Australia 4:1-161. Newbey K R 1984 Physical environment. In: Biological Survey of the eastern Goldfields of Western Australia. 2. Widgiemooltha-Zanthus area. Records of the Western Australian Museum, Supp 18:29-40. Newbey K R 1988 Physical environment. In: The Biological Survey of the Eastern Goldfields of Western Australia. 4. Lake Johnston-Hyden. Records of the Western Australian Museum, Supp 30:7-16. Newbev K R 1995 Physical environment. In: The Biological Survey of the Eastern Goldfields of Western Australia. 11. Boorabbin-Southern Cross area (eds Keighery G J, McKenzie NL & Hall N J) . Records of the Western Australian Museum, Suppl 49:6-15. Onions C T 1978 The Shorter Oxford English Dictionary. Clarendon, Oxford. Ornduff R 1987 Islands on Islands: Plant Life on the Granite Outcrops of Western Australia. University of Hawaii Press, Honolulu. Smith R I L and Prince P A 1984 The natural history of Beauchene Island. Biological Journal of the Linnaean Society 24:233-283. Swan L W 1992 The aeolian biome ecosystems of the earth's extremes. Bioscience 42:262-270. Twidale C R & Campbell E M 1984 Murphy Haystacks, Eyre Peninsula, South Australia. Transactions of the Royal Society of South Australia 108:175-183. White M E 1986 The Greening of Gondwana. Reed, Frenchs Forest. Withers P C & Edward D H D 1997 Terrestrial fauna of granite outcrops in Western Australia. Journal of the Royal Society of Western Australia 80:159-166. 122 Journal of the Royal Society of Western Australia, 80:123-129, 1997 Reproductive ecology of granite outcrop plants from the south-eastern United States R Wyatt Institute of Ecology, University of Georgia, Athens, Georgia 30602 USA email: wyatt@dogwood.botany.uga.edu Abstract Most of the Piedmont Physiographic Province of the southeastern United States is covered with mixed mesophytic forest of oaks, hickories, and pines. Within this "sea", however, are "islands" of exposed granite and gneiss. A characteristic, and largely endemic, assemblage of plants has adapted to the environmental extremes that bare rock provides by strongly altering their morphology, physiology, and life history. With respect to their reproductive ecology, however, these plants appear very similar to their congeners and to the Piedmont flora as a whole. Except for ant- pollinated DiamorpJta smallii, most species show the expected range of pollen vectors, including wind, bees, flies, butterflies, moths and one species of hummingbird. Fruit and/or seed dispersal appears to be highly localised and effected primarily by wind and water. If anything, most species appear to possess adaptations against long-distance dispersal, which would carry propagules into the inhospitable matrix of oak-hickory-pine forest. Mating systems are variable, including examples of both self-compatible and self-incompatible taxa. Consistent with the expectation of low gene flow between populations on isolated outcrops, genetic data show strong differentiation and suggest the potential for genetic drift and/or natural selection to result in divergence. Some of the endemic species on granite outcrops have originated by allopolyploidy, whereas others appear to represent products of more gradual divergence in geographical isolation. There is reason to believe that some weedy species of early successful sites were originally restricted to granite outcrops and spread more recently to sites disturbed by human activities. Introduction and Background Most of the Piedmont Physiographic Province of the southeastern United States is covered with mixed mesophytic forest of oaks, hickories, and pines. Within this "sea", however, are "islands" of exposed granite and gneiss. Such areas constitute a habitat archipelago, supporting a largely endemic assemblage of plants and animals that are remarkably well-adapted to the environmental extremes that bare rock provides. These outcrops are known locally as "flat rocks" or "cedar rocks" and differ strikingly in vegetation from the surrounding Piedmont forest (McVaugh 1943). They occur from central Alabama to south-central Virginia and range from flat exposures of a few square metres to steeply sloping domes 200 m above base level that cover hundreds of hectares (Quarterman et al. 1993). Originally, biologists speculated that the granite outcrops were of relatively recent origin, their appearance having been initiated through burning by native Americans and hastened by the erosion that accompanied poor lumbering and farming practices of European settlers (e.g. Oosting & Anderson 1939). Geological evidence, however, indicates that these rock units are >350 106 years old and that the occurrence of exposed rock in the Piedmont probably dates back at least 150 mybp. This molten, igneous rock was intruded into pre-existing country rock and has become exposed in places where the erosional © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 cycle of the Piedmont has weathered away surrounding rock that proved less resistant (McVaugh 1943). Individual rock outcrops may, therefore, be much younger than might be suggested by the presumption of their existence 150 mybp. The endemic plants and animals have probably "island- hopped" from exposure to exposure as local erosional forces created new outcrops and covered older ones (Wyatt & Fowler 1977). Weathering of the granite itself produces distinctive patterns and creates special habitats to which plants have adapted (Burbanck & Platt 1964). Uneven weathering of the surface produces rock-rimmed depressions that retain water between sporadic rainstorms in the late winter to early spring. These weathering pits may later become filled with sand and organic debris. Invasion of these depressions by plants speeds the chemical weathering process, as rainwater combines with C02 from the plants to produce carbonic acid. On domed outcrops, exfoliation occurs regularly. Huge shells of rock fracture and slide over the top of underlying layers, exposing fresh granite and creating talus at the base of the dome. Environmental conditions on granite outcrops are harsh and differ sharply from conditions in the adjacent forest. Temperatures on the outcrops typically are much higher than in the forest because of high incident radiation, absorption of heat by the rock, and low evapotranspiration. Temperatures in excess of 50 °C are common at the rock surface during summer months. Moreover, the shallow, mineral soil overlying impervious rock and the low vegetation cover lead to extraordinarily high runoff. It is estimated that >95% of the annual 123 Journal of the Royal Society of Western Australia, 80(3), September 1997 precipitation in outcrop communities is lost as direct runoff versus only 10%, on average, for the Piedmont of Georgia (Duke & Crossley 1975). These characteristics of granite outcrops, despite their location within a region that receives >1200 mm of annual precipitation, make them islands of desert embedded in a sea of mesic deciduous forest. The decidedly desertic nature of the outcrops is reflected in the morphological, physiological, and life- history adaptations of the characteristic plants. A number of perennial species tolerate the extremely hot, dry conditions by storing water in succulent stems or leaves (e.g. Opuntia humifusa, Talinum teretifolium, Portulaca smallii). Others use CAM (Crassulacean Acid Metabolism), a photosynthetic pathway that enables them to use water very efficiently. By day, their stomata are closed as they store light energy in the form of organic acids. At night, their stomata open to take in C02 to be fixed into carbohydrates. Water loss by transpiration through open stomata is therefore minimised in such species as Diafnorpha smallii and Sedum pusillum • Perhaps the most widespread solution to the desert conditions, however, is drought avoidance. The overall abundance of different life-forms (sensu Raunkiaer 1934) shows a much lower proportion of trees and shrubs and a much higher proportion of annual herbs than are found in Piedmont forest (Phillips 1982; Walters & Wyatt 1982). In particular, the winter annual life history is extremely common among the characteristic plants of the granite outcrops. Seeds germinate after late September or early October rains to produce a rosette of frost-resistant leaves. When temperatures start to rise in late March or early April, the plants bolt and flower profusely- By early May, when the shallow soil depressions have become bone-dry, these winter annuals have matured a new crop of seeds, which require an after- ripening period of 4-5 months. Thus, they are able to survive the extremely hot, dry summer months as dormant populations of drought-resistant seeds. In addition to D. smallii and S. pusillum, _ there are many winter annuals, some of which, like Agrostis elliottiana, are the only annual members of groups that are typically perennial. Arenaria glabra, which is sometimes treated as merely a variety of the arctic-alpine perennial A. groenlandica ( e.g . Radford et al. 1964), differs chiefly in its annual life history, whose evolution presumably enabled the species to invade granite outcrops. Predictions We have, then, a situation on the granite outcrops of the southeastern US that should permit us to make certain predictions about the reproductive ecology of the characteristic plants. 1. The evolution of the characteristic, and largely endemic, outcrop flora will be paralleled by the evolution of the pollinating and dispersing fauna. 2. Because of their importance to colonising ability, self¬ compatibility and self-pollination will commonly evolve (Baker 1955). 3. As is often seen in island floras, there will be secondary losses of dispersibility and the evolution of dioecy (Carlquist 1974). 4. Population-level genetic differentiation among populations will be strong. 5. Speciation in these small, peripheral isolates will occur rapidly via allopolyploidy or other saltational mechanisms. 6. Because outcrops provide open, relatively low- competition habitats, many weedy species will be able to invade outcrop communities. Empirical Evidence: Patterns and Processes Despite the evolution of a number of animals that are endemic to granite outcrops (Quarterman et al. 1993), none of these serve as important vectors for pollen or seed dispersal. The activities of most, like the beetle Collops georgianus which feeds on pollen and seeds of Diamorpha smallii (Shure & Ragsdale 1977; King 1987), are largely destructive. Instead, the characteristic plants are typically pollinated by more wide-ranging, flying insects (e.g. Wyatt 1983, 1986). For the most part, the plants appear to have continued to maintain the usual pollination syndromes characteristic of their congeners. The grasses, sedges, and rushes (e.g. Agrostis elliottiana , Cyperus granitophihis and I uncus georgianus) are wind- pollinated. The aquatic Amphianthus pusillus is not water- pollinated, but rather insect-pollinated like its relatives in the Gratioleae of the Scrophulariaceae. Tradescantia hirsuticaulis is pollinated primarily by syrphid flies; Senecio tomentosus, by butterflies; and Sedum pusillum , by small native bees and flies. Few species are bird- pollinated, but this is true of the flora as a whole, probably reflecting the depauperate nectarivorous avian fauna, a single species, the ruby-throated hummingbird (Archilocus colubris). Some plant species associated with woods margins near outcrops (e.g. Bignonia capreolata and Aesculus pavia x A. sylvatica hybrids) are regularly visited by this species (dePamphilis & Wyatt 1989). Even when species with highly unusual pollinator relationships occur on the outcrops, they seem merely to have brought these coevolved mutualisms along with them (e.g. Yucca filamentosa and the moth Tegeticula yuccasella; Pellmyr et al. 1996). The one exception to this generalisation is Dia?norpha smallii, which has been reported to be effectively pollinated by ants (Wyatt 1981; Wyatt & Stoneburner 1981). The native ants Formica schaufussi and F. subsericea regularly visit the flowering plants to feed on the small quantities of nectar produced. In doing so, they pick up large numbers of pollen grains on their unusually hairy bodies, presumably transferring them between individuals of this incompatible species. Tests of pollen applied to ant bodies have shown that viability remains high even after several hours (linpubl. data), suggesting that these ants do not secrete substances that kill pollen, as many ant species do (Peakall & Beattie 1987). With respect to vectors for fruit and seed dispersal, the situation is much the same as for pollinators; the characteristic plants show much the same syndromes as their congeners and as the Piedmont flora as a whole. Most species are moved locally by water during convective storms and by wind during occasional wind storms. If anything, however, the species appear to show 124 Journal of the Royal Society of Western Australia, 80(3), September 1997 2 Photographs illustrating some of the characteristic granite outcrop habitats and plants described in the text. Plate 1. Stone Mountain, near Atlanta, Georgia, is a 237- hectare granite dome. Such steeply sloping exposures belie the name "flat rocks," which is commonly applied to granite outcrops in the south-eastern United States, Plate 2. A small, flat granite outcrop from eastern Alabama. Vegetation mats develop as soil accumulates in depressions weathered in exposed granite. These microcosms represent different stages of succession: Diamorpha smallii (Crassulacea), a red-bodied succulent, 4 dominates the shallow pool to the left; fruticose lichens and Andropogon virginicus (Gramineae) are abundant in the deeper depression to the right. Plate 3. Indicative of the hot, dry environment on granite outcrops is the stem succulent Qpuntia humifusa (Cactaceae). Here it flowers in mid-May on a granite outcrop in Alabama. Plate 4. Vegetation mats over granite develop concentric zones, with earlier colonists displaced to shallower soil at the edges. Here, progressing toward the center, we see Diamorpha smallii (Crassulaceae), Arenaria uniflora (Caryophyllaceae), Senecio tomentosus 6 (Compositae), and Andropogon virginicus (Gramineae). Plate 5. A large rock-rimmed pool on Heggie's Rock, near Augusta, Georgia, holds water between rainstorms during winter (December-March). This is the type locality for the endangered mat-forming quill wort, Isoetes tegetiformans. Plate 6. Two other endangered species that grow as aquatics in the rock-rimmed pools are Amphianthus pusillus (Scrophulariaceae) and Isoetes melanospora. Amphianthus is a monotypic genus that produces long stalks bearing two floating leaves, between which a single, white flower emerges. Plate 7. g Characteristic of the many winter annuals that occur on granite outcrops, Sedum pusillum (Crassulaceae) germinates in fall and over¬ winters as a frost-resistant rosette. It grows in moss mats of Hedwigia ciliata, typically in the shade of Juniperus virginiana (Cupressaceae) trees. Plate 8. Tradescantia hirsuticaulis (Commelinaceae) occurs in the thin woods fringing rock outcrops. Unlike most of the other endemics that have been studied to date, this species has relatively high levels of genetic variation. Plate 9. Another winter annual restricted to granite outcrops, Arenaria uniflora (Caryophyllaceae) flowers early in the jq spring. It has been hypothesised that competition for pollinators between the relatively large-flowered, out-crossing plants shown here and another species {A. glabra) with even larger flowers has served as the stimulus for the evolution of self-pollination in geographically marginal populations in Alabama and the Carolinas. Plate 10. The flowers of Talinum tnengesii (Portulacaceae) are open for only a few hours early in the afternoon. In addition to strong spatial separation of the anthers and pistil (herkogamy), out-crossing is promoted by a unique pollen germination delay mechanism. 125 Journal of the Royal Society of Western Australia, 80(3), September 1997 adaptations against too much dispersal. Long-distance transport, for the most part, would likely deliver a propagule to the inhospitable matrix of forest in which the islands of exposed rock occur. An example of this loss of dispersibility, which characterises many island floras (Carlquist 1974), is Diamorpha smallii, which differs from its close relatives in the genus Sedum by producing fruits that dehisce by a tear-shaped flap at the back of the follicle (Sherwin & Wilbur 1971). This results in the seeds being dropped directly below the parent plant, where they lie dormant during the hot, dry summer. I suspect that there may occasionally be directed long¬ distance dispersal of seeds between outcrops. Birds, such as killdeer, often visit the outcrops to drink from the pools and to forage for invertebrates in the seepage areas at the outcrop margins (pers. observations). In doing so, they may inadvertently pick up plant progagules in the mud on their feet. After travelling to another outcrop, the propagules may be washed off. It is possible that hawks and vultures, which are also commonly seen in the vicinity of the outcrops drinking from pools, hunting, and riding thermals, act similarly as dispersal vectors. In any event, however, gene flow via these dispersal vectors is likely to be rare. Other species, like the Georgia oak ( Quercus georgiana), pose additional difficulties, as their seeds are very large. Rarely, they may be transported between outcrops by blue jays, which are known to collect and cache seeds of oaks (Darley-Hill & Johnston 1981). With the exception of Isoetes, plants whose primary mode of dispersal is spores do indeed seem to be moved long distances by the wind. For example, Pellaea wrightiana is a common fern of rocky cliffs in the southwestern United States. Some years ago it was collected from a single granite outcrop in North Carolina, more than 1 600 km from the nearest populations in central Texas (Wagner 1965). Similarly, Astrolepis sinuata , a fern otherwise restricted to west Texas and northern Mexico, has been collected recently from a small granite outcrop in western Georgia (J R Allison, pers. comm.). The fern Pilularia americana and a number of lichens and mosses also show long-range disjunctions to outcrops in the southeast (Wyatt & Stoneburner 1982). The species of Isoetes that are endemic to granite outcrops in the southeastern United States are a notable exception to the pattern shown by homosporous pteridophytes. Genetic markers suggest that gene flow between outcrop populations of I. piedmont ana , /. melanospora, and l tegetiformans is very low (Van De Genachte & Wyatt, unpublished). This may be due to the fact that successful colonisation by these heterosporous plants requires simultaneous dispersal not only of the small microspores, but also of the much larger (and presumably, less dispersible) megaspores. Many of the characteristic granite outcrop species in the southwestern United States are characterised by high rates of outcrossing. Unlike the situation in many island floras (e.g. Hawaii; Carlquist 1974), however, there is not a disproportionately high number of dioecious species. The Piedmont flora as a whole is rather depauperate in dioecious species, which constitute only about 3.4% of the flora (Conn et at. 1980). The dioecious species that are common on outcrops are mostly woody taxa with wider ranges off the outcrops (e.g. Juniperus virginiana, Smilax smallii , Forestiera ligustrina). Dioecy is rare among annuals (Conn et al. 1980); thus, it is interesting to note that Rumex hastatulns is among the weedy species that may once have been restricted to granite outcrops (see below). Among the known outcrossers on granite outcrops are taxa demonstrated to be genetically self¬ incompatible (e.g. Diamorpha smallii, Wyatt 1981; Tradescantia hirsuticaulis, nnpubl. data) and those with well-developed herkogamy (e.g. Talinum mengesii, Murdy et al. 1970) or dichogamy (e.g. central populations of Arenaria uniflora , Wyatt 1984). Murdy & Carter (1987) discovered a unique pollen germination delay mechanism in Talinum mengesii. They reported that pollen grains deposited on the stigma during the few hours that each flower was open in the early afternoon failed to germinate immediately, but rather were delayed for up to 90 minutes. They speculated that this had the effect of opening up the range of mates siring seeds on each plant and of intensifying gametophytic competition. Even more numerous, however, are the many taxa, including some in the same genera listed above, that have evolved self-fertilisation. Self-compatibility has been documented in Sedum pusillum (Wyatt 1983), which also showed evidence of outbreeding depression, and Zephyranthes atamasco (Broyles & Wyatt 1991). Talinum teretifolium is self-compatible and highly self-pollinating relative to its diploid progenitor, T. mengesii (Murdy 1968; Murdy et al 1970; Murdy & Carter 1985). In this it contrasts sharply with the diploid /allotetrapolid species- pair Sedum pusillum/Diamorpha smallii, in which the former is self-compatible and the latter, self-incompatible. The marginal populations of Arenaria uniflora studied by Wyatt (1984) are self-compatible and strongly self- pollinating. This also appears to be true for Phacelia dubia var. georgiana (Levy 1988). I am inclined to discount early reports of a "balanced breeding system" in Amphianthus pusillus, in which plants produce immersed cleistogamous flowers at the base of the rosette and emersed chasmogamous flowers in the axils of paired bracts at the tips of elongated, floating branches (e.g. Pennell 1935). The "cleistogamous" flowers appear to be merely an earlier developmental stage, observed prior to elongation of the floating branch. True "hydroautogamy" appears to be rare (Philbrick 1991). Surprisingly few of the species endemic to granite outcrops have been analysed for genetic population structure. Using horizontal starch-gel electrophoresis, Wyatt et al. (1992) found low genetic variation in outcrop populations of Arenaria uniflora; the percentage of loci polymorphic per population averaged 17.9%; mean number of alleles per locus, 1.0; and expected heterozygosity, 0.048. The selfing populations at the margins of the range were often fixed for alleles that were polymorphic in the central, outcrossing populations. This led to a remarkably high value for GST, which expresses the proportion of the total gene diversity that exists as differences among populations. The GST of 0.572 indicates very strong genetic differentiation between outcrops. Similar results were obtained for Isoetes melanospora (25.4%, 1.32, and 0.069) and I. tegetiformans (12.5%, 1.14, and 0.020) by Van De Genachte & Wyatt (unpublished). Again, strong genetic differentiation was seen for these species (GST = 0.513 and 0.217, respectively). In contrast, Godt & Hamrick (1993) found high levels of genetic variation in 126 Journal of the Royal Society of Western Australia, 80(3), September 1997 Tradescantia hirsuticaiilis (53.5%, 1.72, 0.157). Nevertheless, the isolated populations were more strongly differentiated than expected (GST = 0.183). Only in T. hirsuticaiilis have gene flow estimates > 1 been estimated (Nm = 2.150; Godt & Hamrick 1993). Presumably, there is very little gene flow between outcrop populations of Aremria and Isoetcs , and genetic drift therefore can have strong effects on these isolated populations. Recently, Levy et ah (1996) have analysed genetic variation in Phacelia using restriction fragment length polymorphisms in chloroplast-DN A. They found surprisingly high levels of intraspecific polymorphism, especially in populations of the granite outcrop endemic Phacelia dubia var georgiana. Levy et ah (1996) discounted the likelihood of these polymorphisms being due to gene flow between outcrop populations. At least two of the most characteristic species endemic to granite outcrops have originated via allopolyploidy. Using isozymes as genetic markers, Murdy & Carter (1985) showed that Talinum teretifolium is an allotetrapolid that combines genomes from diploid T. mengesii and diploid T. parviflorum from the west- central United States and one locality in Alabama (formerly called T. appalachianum ; Carter & Murdy 1985). Despite conflicting evidence for some loci and a disquietingly high number of "null" alleles, they were able to reject clearly the alternative hypothesis that T. calycinum was the second progenitor. Although definitive proof remains elusive, it seems clear that Diamorpha smallii also is an allotetrapolid (n = 9), one of whose progenitors is diploid Sedum pusillum (n = 4). As Baldwin (1940) suggested, the other progenitor perhaps should be sought among those annual species of Sedum, such as S. nuttallianum of the west-central United States, with n = 5. Alternatively, D. smallii could be related to Parvisedum, a small genus of California and northern Mexico with n = 9 (Clausen 1975). Murdy (1968) also suggested that the outcrop endemic Cyperus granitophilus may have originated via autopolyploidy from the more widespread C. aristatus. For allopolyploid speciation, hybridisation is a necessary prerequisite. Interestingly, the outcrops seem to represent "hot spots" for hybridisation. To some extent, this may be due to their offering a wide range of habitats, some of which mimic those of other physiographic provinces. For example, the Coastal Plain Senecio tomentosus extends its range into the Piedmont on the outcrops, where it comes into contact with S. anonymus. Chapman & Jones (1971) documented hybridisation between these species, showing that the F1 hybrids have reduced pollen fertility. Other species notable for their tendency to hybridise in proximity to granite outcrops include Quercus (Q. georgiana x Q. nigra, unpiibl. data), Aesculus (A. pavia x A. sylvatica, de Pamphilis & Wyatt 1989, 1990), and Isoetes (I. tegetiformans x I. piedmontana, Van De Genachte & Wyatt unpublished; I. melanospora x /. piedmontana, Allison 1993). It would also appear likely that the geographical isolation of populations on granite outcrops could lead to allopatric speciation, especially following a long-distance dispersal event. This process may underlie the origin of the endemics Phacelia dubia var georgiana and P. maculata (Murdy 1966, 1968; Levy 1991a). Levy's (1991b, 1996) subsequent experimental studies have suggested how reproductive barriers can arise as taxa diverge and strongly hint that new incipient varieties are still being formed. Two other outcrop species that Murdy (1968) advanced as examples of ecogeographical differentiation are Portulaca smallii and Rhynchospora saxicola. This interpretation, however, may be backwards (see below), as the Coastal Plain habitats occupied by their presumed progenitors are probably younger than the Piedmont outcrops. Much interest has focused on two of the outcrop endemics that appear to be highly isolated in their families and to have no close relatives; Amphianthus pusillus and Viguiera porteri. McVaugh (1943) used the former to argue for the great antiquity of the outcrop flora and the latter to show an affinity with the Madro- Tertiary flora of the southwestern deserts, as all other species of Viguiera are restricted to that region. More recent evidence from phylogenetic reconstructions based on chloroplast-DNA gene sequences shows that Amphianthus is closely allied with the Gratioleae, including Gratiola, Bacopa and Mecardonia, among others (dePamphilis, unpubh data) Similarly, Schilling & Jansen's (1989) data concur with Schilling & Heiser's (1981) suggestion that V. porteri is not a Viguiera at all, but rather a Helianthus. The weedy species commonly observed on granite outcrops are a very limited subset of the total Piedmont weed flora. The group includes few, if any, exotics (McVaugh 1943). This observation led Wyatt & Fowler (1977) to propose that these species may, indeed, have once been endemics restricted to outcrop habitats. They possessed a number of features, howTever, that enabled them to invade disturbed habitats, such as old fields, following the arrival of humans. Marks (1983) has advanced this same argument. Thus, for example, rather than understanding the ecotypic adaptation of the old- field dominant grass Andropogon virginicus to outcrops, as Chapman & Jones (1975) did, it might be more appropriate to read the evolutionary process in the opposite direction. Broadening this view a bit, it might be useful to consider the outcrops as the possible original evolutionary springboard for such present-day widespread Coastal Plain plants as Rumex hastatulus, Linaria canadensis, and Crotonopsis elliptica. These pre¬ eminent weedy plants of the Coastal Plain may have perfected their ability to grow in open, disturbed areas of poor, sandy soil while existing primarily on granite outcrops. When similar habitats became available on the Atlantic and Gulf Coastal Plains, they may have undergone range expansion. Other species that appear to fit this pattern include Nothoscordum bivalve, Schoenolirion croceum, Trifolium carolinianum, Forestiera ligustrina, and Houstonia pusilla. Acknoivledgments: I thank the organisers of this symposium, S D Hopper, D Edward, and P Withers, for the invitation to participate and the opportunity to write on this subject. I also thank S D Hopper for stimulating my thinking about the unique habitats and organisms of granite outcrops. He and a long list of students and colleagues have kept my interest alive for many years; among others, these include ] Allison, M L Burbanck, S B Broyles, M E B Carter, C W de Pamphilis, B Ellwood, E A Evans, J W Glasser, M ] W Godt, J L Hamrick, P S King, F Levy, W H Murdy, D J Shure, J L Sorenson, A Stonebumer, and E E Van De Genachte. 127 Journal of the Royal Society of Western Australia, 80(3), September 1997 References Allison J R 1993 Technical draft recovery plan for three granite outcrop plant species. Federal Register. US Fish & Woldlife Service, Jackson MI, 1-41. Baldwin J T 1940 Cytophyletic analysis of certain annual and biennial Crassulaceae. Madrono 5:184-187. Baker H G 1955 Self compatibility and establishment after long¬ distance dispersal. Evolution 9:347-349. Broyles S B & Wyatt R 1991 The breeding system of Zephyranthes atatnasco (Amaryllidaceae). Bulletin of the Torrey Botanical Club 118:137-140. Burbanck M P & Platt R B 1964 Granite outcrop communities of the Piedmont Plateau in Georgia. Ecology 45:292-305. Carlquist S 1974. Island Biology. Columbia University Press, New York. Carter M E B & Murdy W H 1985 Systematics of Talinum parviflorutn Nutt, and the origin of T. teretifolium Pursh (Portulacaceae). Rhodora 87:121-130. Chapman G C & Jones S G 1971 Hybridization between Senecio smallii and S. tomentosus (Compositae) on the granite flatrocks of the southeastern United States. Brittonia 23:209- 216 Chapman R H & Jones S B 1975 Ecotypic differentiation in Andropogon virginicus (Gramineae). Bulletin of the Torrey Botanical Club '102:166-171. Clausen R T 1975 Sedum of North America North of the Mexican Plateau. Cornell University Press, Ithaca, New York. Conn J S, Wentworth T R & Blum U 1980 Patterns of dioecism in the flora of the Carolinas. American Midland Naturalist 103:310-315. Darley-Hill S & Johnston W C 1981 Acorn dispersal by the blue jay ( Cyanocitta cristata). Oecologia 50:231-232. dePamphilis C W & Wyatt R 1989 Hybridization and introgression in buckeyes (Aesculus: Hippocastanaceae): A review of the evidence and a hypothesis to explain long¬ distance gene flow. Systematic Botany 14:593-611 dePamphilis C W & Wyatt R 1990 Electrophoretic confirmation of interspecific hybridization in Aesculus (Hippocastanaceae) and the genetic structure of a broad hybrid zone. Evolution 44:1295-1317 Duke K M & Crossley D A 1975 Population energetics and ecology of the rock grasshopper, Trimerotropis saxatilis. Ecology 56:1106-1107. Godt M J W & Hamrick J L 1993 Genetic diversity and population structure in Tradescantia hirsuticaulis (Commelinaceae). American Journal of Botany 80:959-966. King P S 1987 Macrogeographic and microgeographic structure of a spatially subdivided beetle species in nature. Evolution 41:401-416 ' Levy F 1988 Effects of pollen source and time of pollination on seed production and seed weight in Pliacelia dubia and P. maculata (Hydrophyllaceae). American Midland Naturalist 119:193-198. Levy F 1991a Morphological differentiation in Pliacelia dubia and P, maculata. Rhodora 93:11-25. Levy F 1991b A genetic analysis of reproductive barriers in Pliacelia dubia. Heredity 67:331-345. Levy F, Antonovics J, Boynton J E & Gillham N W 1996 A population genetic analysis of chloroplast DNA in Pliacelia. Heredity 76:143-155. Marks P L 1983 On the origin of the field plants of the northeastern United States. American Naturalist 122:210-228. McVaugh R 1943 The vegetation of the granitic flatrocks of the southeastern United States. Ecological Monographs 13:119— 166. Murdy W H & Carter M E B 1985 Electrophoretic study of the allopolyploidal origin of Talinum teretifolium and the specific status of T. appalachianum (Portulacaceae). American Journal of Botany 72:1590-1597. Murdy W H & M E B Carter 1987 Regulation of the timing of pollen germination by the pistil in Talinum mengesii (Portulacaceae). American Journal of Botany 74:1888- 1892. Murdy W H 1966 The systematics of Pliacelia maculata and P. dubia var. georgiana, both endemic to granite outcrop communities. American Journal of Botany 53:1028-1036. Murdy W H 1968 Plant speciation associated with granite outcrop communities of the southeastern Piedmont. Rhodora 70:394-407. Murdy W H, Johnson T M & Wright V K 1970 Competitive replacement of Talinum mengesii by T. teretifolium in granite outcrop communities of Georgia. Botanical Gazette 131:186-192. Oosting H J & Anderson L E 1939 Plant succession on granite rock in eastern North Carolina. Botanical Gazette 100:750-768. Peakall R & Beattie A J 1987 Pseudocopulation of an orchid by male ants: A test of two hypotheses accounting for the rarity of ant pollination. Oecologia 73:522-524. Pellmyr O, Leebens-Mack J & Hutli C J 1996 Non-mutualistic yucca moths and their evolutionary consequences. Nature 380-155:156. Pennell F W 1935 The Scrophulariaceae of eastern temperate North America. Philadelphia Academy of Natural Sciences Monographs 1:1-650. Philbrick CT 1991 Hydrophily: Phylogenetic and evolutionary considerations. Rhodora 93:36-50. Phillips D L 1982 Life forms of granite outcrop plants. American Midland Naturalist 107:206-208. Quarterman E, Burbanck M P & Shure D J 1993 Rock outcrop communities: Limestone, sandstone and granite. In: Biodiversity of the southeastern United States/Upland terrestrial communities (eds W H Martin, S G Boyle & A C Echtemacht). John Wiley & Sons, New York, 35-86. Radford A E, Ahles H E & Bell C R 1964 Manual of the Vascular Flora of the Carolinas. University of North Carolina Press, Chapel Hill, North Carolina. Raunkiaer C 1934 The Life-forms of Plants and Statistical Plant Geography. Oxford University Press, London. Schilling E E & Heiser C B 1981 An infrageneric classification of Helianthus (Compositae). Taxon 30:393-403. Schilling E E & Jansen R K 1989 Restriction fragment analysis of chloroplast DNA and the systematics of Viguiera and related genera (Asteraceae: Heliantheae). American Journal of Botany 76:1769:1778. Sherwin PA & Wilbur R L 1971 The contributions of floral anatomy to the generic placement of Diamorpha smallii and Sedum pusillum. Journal of the Elisha Mitchell Scientific Society 87:103- 114. Shure D J & Ragsdale H L 1977 Patterns of primary succession on granite outcrop surfaces. Ecology 58:993-1006. Wagner W H 1965 Pellaea wrightiana in North Carolina and the question of its origin. Journal of the Elisha Mitchell Scientific Society 81:95-103. Walters T W & Wyatt R 1982 The vascular flora of granite outcrops in the Central Mineral Region of Texas. Bulletin of the Torrey Botanical Club 109:344-364. Wyatt R & Fowler N L 1977 The vascular flora and vegetation of the North Carolina granite outcrops. Bulletin of the Torrey Botanical Club 104:245-253. 128 Journal of the Royal Society of Western Australia, 80(3), September 1997 Wyatt R & Stoneburner A 1981 Patterns of ant-mediated pollen dispersal in Diamorpha smallii (Crassulaceae). Systematic Botany 6:1-7. Wyatt R & Stoneburner A 1982 Range extensions for some cryptogams from granite outcrops in Alabama. Bryologist 85:405-409. Wyatt R 1981 Ant-pollination of the granite outcrop endemic Diamorpha smallii (Crassulaceae). American Journal of Botany 68:1212-1217. Wyatt R 1983 Reproductive biology of the granite outcrop endemic Sedum pusillum (Crassulaceae). Systematic Botany 8:24-28. Wyatt R 1984 The evolution of self-pollination in granite outcrop species of Arenaria (Caryophyllaceae) I Morphological correlates. Evolution 38:804-816. Wyatt R 1986 Ecology and evolution of self-pollination in Arenaria uniflora (Caryophyllaceae). Journal of Ecology 74:403-418. Wyatt R, Evans E A & Sorenson J D 1992 The evolution of self- pollination in granite outcrop species of Arenaria (Caryophyllaceae) VI Electrophoretically detectable genetic variation. Systematic Botany 17:201-209. 129 Journal of the Royal Society of Western Australia, 80:131-139, 1997 The geobiological interface: Granitic outcrops as a selective force in mammalian evolution M A Mares Oklahoma Museum of Natural History and Department of Zoology, University of Oklahoma, Norman, Oklahoma 73019 USA email: mamares@ou.edu Abstract Granite outcrops appear with regularity in various parts of the world. They may be associated with mountains, or they may appear in areas that today have little other topographic relief than the outcrops themselves. Wherever granitic agglomerations are found, however, their influence on the biota of a region is profound. Rocks influence both microhabitat and macrohabitat for plants and animals. In some cases, the rocky habitats permit a more mesic microhabitat to develop within a generally more xeric region. Such microclimatic influences may permit the establishment of trees and the maintenance of green leaves, even during droughts. For mammals, in addition to the increased availability and predictability of food, the rock habitat itself provides shelter, amelioration of climatic variables and, in some cases, a resource that is defensible by the mammals. However, the rock habitat is one that also demands special adaptations by mammals that specialize toward a rupicolous existence. Such specializations include morphological, ecological, and behavioral traits that develop in direct response to the selective forces required to inhabit rocks. These influences may extend even to mating systems, where the unusual habit of harem polygyny develops among small mammals whose males manage to defend the rock resource against other males, thereby accruing females. Because of their influence on macro- and microhabitat, granite outcrops also have a regional influence on biodiversity, increasing significantly the number of species that occur in areas that are frequently rather depauperate due to their semiarid or arid nature. The responses of mammal faunas in Africa and South America to granite outcrops is used to compare and assess the importance of these rock outcrops on aspects of mammalian biology. Introduction The diversity of mammals in any particular region has been related to many factors, including climate, age of the fauna, past formation of refugia, productivity, environmental predictability, competitors and predators, size of the area, and other factors (see RosenzwTeig 1995 for review). In general, the effect of physical structure of the environment (as opposed to the biotic structure) on patterns of diversity and richness has been given less attention, although there has been some work on the effects of dunes on small mammal species richness (Brown 1973). Even in this study, however, the dunes have been considered mainly as habitats that supported fewer species of mammals than the surrounding better- vegetated areas. In a more detailed analysis of regional patterns of species diversity. Brown (1975), found that in a xeric North American desert system, overall species diversity of rodents is strongly influenced not only by a- diversity (species occurring in a single locality), but by (3- diversity (species turnover) as well, since many species have restricted distributions in deserts. Of course, at a macroscale, the influence of mountain ranges and other major topographic features has long been known to have a major influence on species richness through increased habitat diversity, altitudinal effects, refugial formation, and other factors ( e.g . Simpson 1964; Wilson 1974; Brown 1978; Mares & Ojeda 1982; Mares 1992; Willig & Sandlin © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 1991). At a smaller scale, the influence of structural features of the environment on patterns of mammalian diversity and richness are less apparent. One regularly occurring structural feature of habitats scattered throughout the world is granite outcrops. As has been shown in other papers in this issue, the influence of these outcrops on plants, microorganisms, animals, and people are extensive. There has not been a great deal of research on the influence of granite outcrops on mammals, although there are some notable exceptions, particularly in examining the influence of rocks on the biology of small mammals in southern Africa and South America. Hoeck (1975, 1982) and Hoeck et al (1982), for example, studied hyraxes in the Serengeti and found that the rocks themselves were a critical resource. The males of these rock specialists defended the rock-piles in such a way that they were able to attract and keep multiple females as mates. The females also required access to the rocks for food, water, protection, and nesting space. Thus the male hyraxes, by virtue of being able to control access to the rocks, formed harems; harem polygyny is an unusual behavioral trait in a small mammal. Lacher (1981) studied a caviid rodent Kerodon rupestris (guinea pig family) in eastern Brazil, that was also highly specialized for living on granitic outcrops within a larger dry tropical scrubland, the Caatinga. He found that Kerodon, which strongly resembled a hyrax externally, also defended rock-piles and, like the hyraxes of Africa, also accrued females to form harems. In a multivariate comparison of morphological, ecological, and behavioral characteristics of rock-dwelling 131 Journal of the Royal Society of Western Australia, 80(3), September 1997 small mammals from throughout the world, Mares & Lacher (1987) showed that mammals inhabiting rocks have developed a host of similar adaptations, including specific morphological, ecological, and behavioral traits, such as specialized foot structure, general body form, dietary similarities, similar activity periods, alarm calls, communal defaecation, and a propensity to form harems. The formation of harems seems to result from the ability of the males to defend portions of the rock-piles, thus controlling a resource that is required by the females and their offspring. In order for a female to have access to the critical rock resource, and all of the ecological factors associated with the rocks, she must enter into the harem of the male. Mares & Lacher interpreted this to be an example of resource defense polygyny (Emlen & Oring 1977), a pattern that has also been documented for such non-rock specialists as tree-dwelling bats (the Neotropical emballonurid, Saccopteryx bilineata) that control access to cavities in the trees by defending them against other males, thus developing harems of females that require access to the hollow tree in order to raise their young (Bradbury & Emmons 1974). Mares & Lacher (1987) also made clear that becoming a rock specialist requires a broad array of adaptations. Moreover, these adaptations are similar from one part of the world to another, regardless of the phylogenetic history of the organisms in question (they examined 24 species of mammals in 17 genera, 12 families and 4 orders). At a larger scale, rocks have also been shown to have an influence on the regional species richness of small mammals, especially in southern Africa, where this phenomenon has been examined in some detail. Coetzee (1969), for example, noted that "The distribution of animal life (especially the small mammals) is .... largely influenced by rocky outcrops''; he provided a list of 28 species of mammals that frequent rocky outcrops in Namibia. He later (Coetzee 1983) supported these observations with an analysis of mammalian geographic ranges. In a more detailed ecological study in Namibia, Withers (1979) examined the small mammals of an inselberg (isolated rocky outcrop) community and found that both rock specialists and more-generalist species comprise the community. The community ecology of the mammals was greatly affected by microclimate effects of the inselberg and its associated vegetation, as opposed to the surrounding sparse desert habitat. Although he did not conduct a regional analysis of distributional patterns. Withers showed clearly that the rock-piles were a major factor in small mammal species richness in the Namib Desert. Among the areas used by Mares & Lacher (1987) in their broadly comparative study of rock-dwelling mammals were the Namib Desert, the Caatinga, and the Andean region of Argentina. Mares et al (1981) and Mares et al (1985a) had shown that rock-piles in the Caatinga had a significant influence on both species richness and persistence, and Mares (1993) provided information showing that rocky habitats were important components of generic richness for small mammals in Andean deserts. Here, I extend the comparisons of the Namib Desert, Brazilian Caatinga, and Argentine Andes to consider the effect of granite outcrops on the structure of the entire mammalian community in these three areas. I first provide a brief comparative review of the regions and their history to permit a greater understanding of how mammals have adapted to rock-piles, a significant abiotic structural component of their habitat in three markedly different drylands on two southern continents. Figure 1. The interface of sand, river, and rock in the Namibian Desert near Gobabeb. The light-colored material in the foreground is sand at the edge of the riverine vegetation of the Kuiseb River. On the other side of the river are granitic outcrops. Barely visible in the background beyond the rocks is the gravel plain of the Namib. Both the rocky areas and the river support vegetation that is quite different from either of the two major habitats (the dune fields and the plains). The mammals of the rocks are also distinct from those of the other major habitats. 132 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 2. A broad granite surface exposure (lajeiro) in the Caatinga of northeastern Brazil. In the background are rocky outcrops (serras). Lajeiros collect water and form micro-communities of plants (evidenced by the cacti, shrubs and grasses in the center of the photo). Serras support shrub and tree communities that are distinct in physiognomy and persistence of green vegetation from the surrounding Caatinga scrub forest. The mammals of the lajeiros, serras, serrotes (larger outcrops that form small hills), and chapadas (outcrops that are larger still, forming small mountains) are very distinct from those of the non-rocky areas. Methods Study Areas Namibia. The Namib Desert, a coastal desert, is one of the oldest deserts on earth (Seely 1987; Ward & Corbett 1990). It is also among the world's driest deserts, with much of the precipitation being received as fog (Walter 1986). Over much of the area there is only sparse vegetation, whether on the gravel plains or on the enormous dune "sea" that characterize the region. There are, however, areas that support much more vegetation than the habitats that surround them, and these include rivers (e.g. Kuiseb) and granite outcrops (inselbergs and other outcrops), where trees, shrubs, and grass may be present (e.g. Theron et al. 1980; Seely 1987). In some places, such as near the Desert Research Station at Gobabeb, the various habitats abut one another. Here fields of moving dunes pour over granitic outcrops that line the Kuiseb River valley, north of which an abrupt transition to gravel plains occurs (Fig 1). The history of the Namib region, as well as the general ecology of the desert and its flora, fauna, and habitats has been fairly well studied in recent years (e.g. Coetzee 1969, 1983; Stuart 1975; Joubert & Mostert 1975; Werger 1978; Henschel et al. 1979; Withers 1979; Tilson 1980; Tilson et al. 1980; Walter 1986; Seely 1987, 1990). Although seldom examined specifically, researchers have found that inselbergs play a major role in the maintenance of mammal species richness in this desert by virtue of the microhabitat provided by the rocks, which includes floristic and climatological influences. Caatinga. The Caatinga of northeastern Brazil has also received a great deal of attention from biologists in recent years (e.g. Markham 1972; Eiten 1974; Webb 1974; Reis 1976; Ab'Saber 1977; Lacher, 1981; Willig, 1983; Mares et al. 1981, 1985a; Andrade-Lima 1982). This region of more than 650 000 km2 lies well within the tropics and undergoes periodic pronounced droughts at irregular intervals during which most of the vegetation loses its leaves. Its vegetation is much like a thorn scrub in aspect. Populations of small mammals die back during the droughts, and there is great misery among the human inhabitants of the regions as livestock die and water for human use becomes scarce. During wet years, which are more common than dry years, the area is a lush, low, green forest and mammals and other wildlife species are abundant. During droughts, the Caatinga resembles the semiarid chacoan thorn scrub of western Argentina, including the presence of many types of tree cacti ( Cereus ) and other cacti that frequent the areas of open granite (termed "lajeiros" in Portuguese). The Caatinga also has granite outcrops varying in size from isolated boulder piles to small hills (Fig 2). It is in these outcrops that the area's only endemic mammal is found, the caviid rodent, Kerodon rupestris, a rupicolous species specialized for life on the isolated rock-piles. Mares et al. (1985a) hypothesized that these rock-piles play a major role in maintaining mammal diversity in the Caatinga by offering a microhabitat refuge to mammals during the periodic droughts. They suggested that after severe droughts, the Caatinga, especially the core of the region, is in large part re-colonized by small mammal populations that managed to persist in the more mesic microclimate associated with the rock-piles. Argentine Andes. The Andes of westcentral Argentina also support vast areas of granite outcrops, interspersed with other types of rock (Videla & Suarez 1991), and with flat plains located between mountain massifs (Fig 3). The rocky areas of the Andes are also home to a number of rock specialists among small mammals (Mares 133 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 3. Rock outcrops in an inter-Andean valley near the town of Uspallata, Mendoza Province, Argentina at an elevation of about 2,000 m elevation. The scrub desert of the gravel plains, which mainly supports creosote bush (Larrea), gives way to different communities of plants on the rock outcrops. The area supports a number of mammals that are specialized for life in the rocks. & Lacher 1987; Ojeda & Mares 1989; Mares 1993; Braun & Mares 1996). Much of the high Andes of Argentina includes arid habitats that support either sparse grasses or scattered shrubs (Mares et al. 1985b). Although the mammals of this region have not been as well studied as those of the prev ious two areas, the rocky areas appear to function as important microhabitats for species persistence through their effect on microclimate, plant distribution, and as places for shelter ( e.g . Pearson 1948; Pearson & Ralph 1978). Data Published data on mammals and their habitats were examined for the three areas described above. In particular, information was sought that specifically referenced the use of rocks by the mammals in these regions. Species lists were examined for the areas in question, and information was obtained on surveys of the mammals of the areas that included information on the entire mammal fauna, rather than being limited to small mammals. References utilized are included throughout this report. Additionally, my students and I conducted extensive research on mammals of the Caatinga in the 1970s and on mammals in the Andean and pre- Andean mountains during the last three decades (e.g. Mares et al. 1989, 1997; Barquez et al. 1991). Additionally, I was able to spend a brief field period in Namibia in 1995. Effects of Rock-piles Few studies have considered the ecological effects of rock-piles on mammals. While not designed to measure the ecological effect of rocks on mammals per se, Withers (1979) found that the microclimate of the rocky inselbergs he examined supported different plant species from the surrounding desert, and this microhabitat-plant association was related to the abundance of small mammals, as well as to small mammal species richness. Lacher (1981) and Streilein (1982a,b,c) also found that the rocky habitats provided refugia for small mammals during periods of drought, when the green vegetation and occasional pools of water offered escape from the harsh aridity of the surrounding habitat. Water availability has been shown to be an important factor in reproduction of small mammals in both Namibia and the Caatinga (Christian 1976, 1980; Withers 1979; Streilein 1982c,d), although no research has shown how water might effect small mammals in the Andes. Below, I discuss various effects of rocky habitats that are related to mammal occurrence. A taxonomic listing of mammals in these three areas, as well as indications as to whether or not the taxa are rock specialists, is given in Table 1, whereas a list of rock specialists in the three areas is given in Table 2. Microhabitat effects of rocks As noted, many studies have shown that rocky areas provide a microhabitat that differs significantly from the surrounding region. Clearly, the microhabitat effects of rocks on climate, plants, and animals are interrelated in complex ways, but it is instructive to tease apart the various factors that interact to influence mammals in order to understand more clearly the role that rocks play in patterns of mammalian richness and persistence. In the three arid to occasionally semi-arid habitats considered in this report, the rocky substrate contributes in the following ways to mammalian persistence and richness. Structure. Rocks form a major structural component of the ecosystem, especially in areas that would have little topographic relief were it not for the rocks themselves. The structural effect of rocks can be considered to be a first-level effect of the rocks on the community of 134 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 1 Orders, families, and genera of mammals inhabiting granitic outcrops in Namibia (in the general environs of Gobabeb); in the Brazilian Caatinga (near Exu, Pernambuco); and in the Andes of central Argentina (Mendoza Province). Rock specialists (species that are highly adapted to inhabit rocks and that are generally limited to rocky outcrops) are denoted by an asterisk (*). O = order; F - family. Common names are in parentheses. Based on information for Africa in Coetzee (1969), Stuart (1975), Withers (1979), Skinner & Smithers (1990); for Brazil in Mares et at. (1981, 1985), Streilein (1982a, b,c); for Argentina in Barquez et al. (1991), Mares (1993), Mares et al. (1985b, 1997), and Mares (unpublished data). Namibia O. Macrosceloidea F. Macroscelididae (elephant shrews) Elephantulus rupestris* Elephantulus intufi Macroscelides proboscideus 0. Chiroptera F. Vespertilionidae Eptesicus hottentotus (Long-tailed house bat) F. Molossidae Sauromys petrophilus* (Flat-headed free¬ tailed bat) O. Primates F. Cercopithecidae Papio ursinus (Chacma baboon) Cercopithecus aethiops (Vervet monkey) 0. Hyracoidea F. Procaviidae Procavia capensis* (Rock dassie) O. Rodentia F. Petromuridae Petromus typicus (Dassie rat) F. Muridae Aethomys namaquensis* (Namaqua rock mouse) Petromyscus collinus* (Pygmy rock mouse) F. Hystricidae Hystrix africaeuustralis (Porcupine) O. Lagomorpha F. Leporidae Proriolagus rupestris* (Smith's red rock rabbit) O. Artiodactyla F. Bovidae Oreotragus oreotragus* (Klipspringer) Oryx gazella (Gemsbok) O. Carnivora F. Canidae Vulpes chama (Cape fox) Cams mesomelas (Black-backed jackal) F. Mustelidae ktonyx striatus (Striped polecat) Helogale parvula (Dwarf mongoose) F. Flyaenidae Crocuta crocuta (Spotted hyaena) Hyaena brunnea (Brown hyaena) F. Felidae Felis libyca (African wild cat) Felis nigripes (Small spotted cat) Panthera pardus (Leopard) Caatinga O. Didelphimorphia F. Didelphidae Monodelphis domestica (Gray short-tailed opossum) Didelphis albiventris (White-eared opossum) O. Chiroptera F. Molossidae Molossops mattogrossensis* (Flat-headed free-tailed bat) O. Rodentia F. Caviidae Kerodon rupestris* (Rock cavy) Galea spixii (Cavy) F. Dasyproctidae Dasyprocta prymnolopha (Agouti) F. Echimyidae Thrichomys apereoides* (Punare) O. Carnivora F. Canidae Cerdocyon thous (Forest fox) F. Mustelidae Galictis vittata (Grison) Conepatus semistriatus (Hog-nosed skunk) F. Felidae Felis concolor (Puma) Felis yagouaroundi (Jaguarundi) Felis onca (Jaguar) Argentine Andes O. Didelphimorphia F. Didelphidae Thylamys pusilla (Mouse opossum) O. Rodentia F. Abrocomidae Abrocoma vaccarum* (Chinchilla rat) F. Chinchillidae Chinchilla brevicaudata* (Chinchilla) lagidium viscascia * (Mountain vizcacha) F. Muridae Akodon andinus Eligmodontia typus Graojnys griseojlavus Phyllotis darwinii* O. Carnivora F. Canidae Pseudalopex culpaeus (Andean red fox) F. Mustelidae Galictis vittata (Grison) F. Felidae Lynchailurus pajeros (Pampas cat) Felis concolor (Puma) mammals {i.e. an effect on the abiotic components of the habitat). The rocks contribute to environmental complexity in a number of ways. They offer sites for shelter, where both small and large mammals may place their nests or dens to raise their young in a stable microclimate that is relatively secure from predators and competitors. For example, in a remarkable example of convergent evolution, the granite outcrops of Namibia support a small rock-specialized molossid bat, Sauromys petrophilus , that has an extremely flat body and is adapted to live in the cracks of exfoliating granite (Skinner & Smithers 1990). In the Caatinga, there is also a flattened molossid bat, Molossops mattogrossensis, that is specialized for living in exfoliating granite (Willig & 135 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 2 Genera of rock-specialist mammals inhabiting granitic outcrops in Namibia (in the general environs of Gobabeb); in the Brazilian Caatinga (near Exu, Pernambuco); and in the Andes of central Argentina (Mendoza Province). These species, representing among the three areas 6 orders and 11 families, and including 4 endemic families and 9 endemic genera (endemic to the rocky areas in the region) would not be in the arid area in question if the rock resource were not present. Namibia Caatinga Andes Elephantulus* Molossops* Abrocoma* Sauromys* Kerodon* Lagidium* Pronolagus* Procavia* Petromus* Thrichomys* Chinchilla* Petromyscus* Oreotragus* Jones 1985). By virtue of their elevation above the surrounding habitat, rock outcrops also offer sites for observation posts, where predators can be detected, or territorial communication can occur. For example, the hyraxes of Namibia, the caviids of the Brazilian Caatinga, and the chinchillid rodent ( Lagidium ) of the Argentine Andes (Pearson 1948) are all diurnal herbivores that use the rocks for observation points to detect predatory attacks from the air or from the ground. All use sentinels that give warning calls as predators are sighted (Mares & Lacher 1987). Similar behavior in using the structure of the rocky habitat has been shown for the petromurid rodent, Petromus, in Namibia, and many other species as well (Mares & Lacher 1987). Even large mammals, such as the spotted hyaena (Family Hyaenidae, Crocnta crocuta) and the klipspringer (Family Bovidae, Oreotragus oreotragus) use the rocks for dens (hyaena) or for lookout posts (antelope). Additionally, the rocks offer sites for chemical communication, where communal defaecation piles, a common trait of rock-dwelling mammals (Mares & Lacher 1987), may function in territorial and other communication by elevating the piles of faeces high in the air where the scent can be carried for long distances; see, for example, Branch (1993) for information on how the plains vizcacha (Chinchillidae, Lagostomus maximus), a relative of the mountain vizcacha, Lagidium, uses communal sites for pheromonal communication. The rocky substrate also functions as sites for capturing water. Granitic outcrops may capture and hold rainfall for extended periods, and this water is important to mammals in arid regions. Additionally, the rocks may act to permit condensation of fog in some areas (e.g. Namibia), thus increasing water availability to mammals (Withers 1979). Rocks also provide sites that are defensible by males, or by groups of related individuals. Mares & Lacher (1987) showed that rock-piles, and the shelter and other resources they offer, are defensible by males of several species, thus permitting them to accrue harems of females by apportioning the rock resource differentially on the basis of sex. How extensive harem formation may be in rock mammals is not yet clear, but it is clearly not uncommon among species inhabiting rocky areas, although it does not appear to be a common reproductive pattern among small mammals that do not defend rocks. Microclimate. Rocky areas that occur within a more-xeric region offer mesic refugia through microclimate effects. These can also be considered first-level effects of the rocks on the mammal community. Water accumulated by rocks would contribute to micro-changes in humidity. The shelter offered by the rocks, especially in areas where run-off water or fog condensation may accumulate (Withers 1983), will also tend to increase humidity, and thus reduce water loss for mammals that utilize the shelters during arid periods. Similarly, the rocks offer shade during the heat of the day, and a cool refuge during summer or during the dry season. Alternatively, the rocks store heat, thus modifying the microclimate of shelters during cold periods by keeping temperatures elevated above that of the cold desert air. The rocks provide elevated areas where temperature increases rapidly in the morning sun, thus providing points of rapid warming, where mammals can increase their body temperatures quickly before beginning the day's activities. Most rock mammals have been observed to bask in the early morning (Mares & Lacher 1987). Rocks, particularly those that offer deep shelters, provide temperature stability from the outside air which, in arid areas, can undergo pronounced fluctuations. For example, Mares (1975) reported air temperature changes of up to 38 DC over a 24-hr period in the Andes of northwestern Argentina. Rock-piles also provide a diel factor to microclimate, permitting animals to choose from a selection of microclimates throughout the day and night. Because of the structure of rock-piles, animals can utilize different portions of the rocks at different times of day to increase the climatic variability (or stability) of their habitat. For example, sunny portions of the rocks can be used for warming up for some animals, while at the same time others are cooling down in shaded portions. During very wet weather, dry portions of the rock habitat can be sought by mammals when other microsites are too wet for the species in question. Effects of rock-piles on vegetation. Perhaps one of the most evident second-level effects of rock-piles on mammal communities in an arid region is the effect of the rocks on vegetation. In an area such as the Caatinga, for example, rock-piles support small forests of trees that largely remain green throughout the extended droughts. During wet periods, these refugial forests may be less noticeable, imbedded as they are within the brilliant green scrub forest of the Caatinga, but during droughts, almost the only green vegetation remaining over vast areas is that associated with the rock-piles (Mares et al. 1985a). Similarly, in Namibia, the rocks have been shown to support more complex vegetation than much of the surrounding desert (Seely 1987, 1990; Withers 1979). The vegetation of the rocks is often greener for longer periods due to increased water availability afforded by the rocks. Additionally, the granitic outcrops support perennial trees in zones where most of the other vegetation may be low scattered shrubs or short-lived ephemeral plants. The vegetation permits the browsing habit to develop among small mammals (and larger mammals, as well). Thus such disparate species as hyraxes (Order Hyracoidea), rabbits (Order Lagomorpha), and rodents (Order Rodentia) all have evolved rock specialists that browse on the trees and shrubs associated with the rock- piles. Of course, the vegetation contributes enormously 136 Journal of the Royal Society of Western Australia, 80(3), September 1997 to both habitat complexity, habitat stability, and microclimate. In the Caatinga, the rock-piles are assumed to offer refugia for rock specialists and a wide variety of species that are not rock specialists during periods of drought (Mares et al. 1985a). Plants associated with rocks thus offer food, shelter, and other major components of mammal niches throughout the year. Effect of rocks on mammal species richness As noted, the rocks have significant effects, both first- and second-level, on mammal communities and on major components of mammal niches, including food, shelter, and microclimate. The rocks, their plants, and their special microclimate, make it possible for many species either to specialize on the rocks and their associated resources (Mares & Lacher 1987), or to live in the rocks when the surrounding habitat is marginal due to drought or other factors (e.g. Withers 1979; Streilein 1982a; Mares et al. 1985a). The rocks and their vegetation make possible, for example, a richer invertebrate community, including many insects and arthropods, thus making this group of organisms available to mammalian predators. Thus the existence of certain communities of mammals are only possible because of the rock component of the habitat (Table 2), a fact that can be considered a third- level effect of the rocks on species richness. This community may consist of diurnal browsing rock specialists, such as hyraxes and petromurids in Namibia, caviids and echimyids ( Thrichomys ) in the Caatinga, or chinchillids and abrocomids ( Abrocoma ) in Argentina. However, the rock habitat influences much more than just rock specialists. In Namibia, among small mammals, the rocks harbor Aethomys, an insectivorous murid rodent, and Elephantulus (Order Macrosceloidea), an insectivorous elephant shrew (Withers 1979). In the Caatinga, the rocks support an invertebrate-feeding marsupial, Monodelphis domestic a (Streilein 1982a), and in the Andean mountains at lower elevations (up to 2 000 m) insectivorous mouse opossums (genus Marmoset) also inhabit the rocks (Mares et al 1989). Thus the rocks add to the complexity of the small mammal community, as well as to the vegetative, structural, and climatological complexity of the habitat itself. Larger mammals, such as the klipspringer (noted above, in Namibia; Tilson 1980), the huemul deer ( Hippocamelus ) in the Andes (Merkt 1987; Barquez et al. 1991), and a deer (Mazama gouazoubira) in the Caatinga frequent the rocky habitats, especially during dry weather (Mares et al. 1985; T E Lacher Jr, pers. comm .). As the bats, herbivores, microomnivores, and insectivores inhabit the rocks, the habitat becomes increasingly attractive to carnivores, who can forage on the inhabitants of the rocks and adjacent areas. This can be considered a fourth-level effect of the rocks on mammalian species richness. (Of course, carnivores also respond to the rocks for shelter, microclimate, and water, as do other mammals.) In Namibia, for example, the rocks harbor many carnivores, including the leopard Panthera pardus , the spotted hyaena, the African wild cat Felis lybicus, the black-backed jackal Canis mesomelas, and others (Coetzee 1969). In the Caatinga of Brazil, a canid, two mustelids, and several cats frequent the rocky habitat in order to prey on the species of vertebrates and invertebrates that are associated with the rocks. In Argentina, the same families, canids, mustelids, and cats. frequent the rocky outcrops, many obtaining some of their principal foods from species living in the rocks. Discussion The influences of granitic outcrops on patterns of local and regional mammal diversity and ecology are pervasive. In comparing this very similar habitat in quite different environments, it becomes clear that rocks make it possible for a wide variety of rupicolous species to evolve. These species bring with them a high degree of specialized adaptations (and genetic information) to the rupicolous habit that is not present in species in the surrounding habitat. In fact, the rocks provide an important niche component that can only be readily exploited by rock specialists which have developed the special abilities required to live largely or only in rocks (Mares & Lacher 1987). This means that these species, from numerous orders and many families, evolutionary opt to exploit the rock resource and the microenvironment associated with it. By selecting this evolutionary path, the species are able to persist in regions that would otherwise be more challenging (and support fewer species) due to the generally arid environment (Namibia, the high Andes), or because of the periodic devastating droughts (the Caatinga). Those species that are able to specialize on the rock resource in these regions also appear to move into longer-term ecological strategies, where life spans exceed one year (unlike most boom-and-bust) small mammals, and where microclimatic stability leads to complex morphological, ecological, and behavioral traits to persist in the specialized rock habitat. Once this community of mammals has developed, it becomes possible for species associated with them (e.g. carnivores) also to become regular components of the rock-associated mammal community. Thus at a local level, granitic outcrops lead to an increase not only in the numbers of species supported in an area, but in higher-order diversity as well. More families and even orders of mammals are able to occur in an area because of the presence of the rocks. Additionally, those species that have evolved to live in rocks are often so specialized that they are endemic to the rocks; no rocks, no species and, in many cases, no genera, families or orders either. This has significant implications for research in conservation biology. Although much attention has been paid to the presence or absence of species as an important measure of a habitat's importance, habitats that contain endemic species (or families or orders) are much more important to maintain than habitats that merely support many species (Mares 1992). Mares made this argument for a comparison between drylands and tropical forests in the Neotropics, but the same argument can be extended to the unique fauna of rocky outcrops as well. Because rocks open up a series of major new niches in an area, and because to specialize on the rock resource demands a host of unusual adaptations, those species that have taken this evolutionary step become quite different genetically from their nearest relatives. So different, in fact, that their nearest relatives are often in other genera or even families. This is the higher order diversity that was discussed by Mares (1992). When comparing a 137 Journal of the Royal Society of Western Australia, 80(3), September 1997 tropical and two temperate regions on two different continents, the faunistic evolutionary pattern that appears is similar. Rocks lead to specializations among mammals, which leads to large series of unique adaptations, often quite distinct from their nearest relatives. The rocks also lead to the formation of communities of these specialized mammals that differ greatly in their suites of adaptations from species in the surrounding non-rock habitat. Because the rocks are an ancient habitat and have likely impelled this evolutionary process to begin long ago, rock specialists are phylogenetically unique as well, and the degree of endemism is high at the species level, and at higher taxonomic levels. This results in faunas that contain a great deal of unique genetic material in areas that are quite limited geographically. These should be areas of intense conservation interest. These influences are apparent at both the local and regional level. Indeed, it would be interesting to compare the contribution of rock¬ dwelling mammals to global diversity not only at the species level, but at levels above the species. Initial indications are that the contribution of rock-piles to global mammal diversity would be quite high indeed. References Ab'Saber A N 1977 Espac^os ocupados pela expansao dos periodos glacais quaternaries. Paleciimas 3:1-19. Andrade-Lima D de 1982 Present-day forest refuges in northeastern Brazil. In: Biological Diversification in the Tropics (ed G T Prance). Columbia University Press, New York, 245-251. Barquez R M, Mares M A & Ojeda R A 1991 The Mammals of Tucuman (Los Mamiferos de Tucuman). Special Publication, Oklahoma Museum of Natural History, Norman, Oklahoma. Bradbury J W & Emmons L H 1974 Social organization of some Trinidad bats. 1. Emballonuridae. Zeitschrift fur Tierpsychologie 36:137-183. Branch L C 1993 Intergroup and intragroup spacing in the plains vizcacha, Lagostomus maximus. Journal of Mammalogy 74:890-900. Braun J K and Mares M A 1996 Observations on unusual morphological and behavioral traits in Abrocoma (Rodentia: Abrocomidae) from Argentina. Journal of Mammalogy 77:891-897. Brown J H 1973 Resource utilization and coexistence of seed¬ eating desert rodents in sand dune habitats. Ecology 54:788-797. Brown J H 1975 Geographical ecology of desert rodents. In: Ecology and Evolution of Communities (ed M L Cody & J M Diamond). Belknap Press of Harvard University, Cambridge, Massachusetts, 315-341. Brown J H 1978 The theory of insular biogeography and the distribution of boreal birds and mammals. Great Basin Naturalist Memoirs 2:209-227. Christian D P 1976 Physiological correlates of demographic patterns in three sympatric Namib Desert rodents. Physiological Zoology 52:329-339. Christian D P 1980 Vegetative cover, water resources, and microdistributional patterns in a desert rodent community. Journal of Animal Ecology 49:807-816. Coetzee C G 1969 The distribution of mammals in the Namib Desert and adjoining inland escarpment. Scientific Papers of the Namib Desert Research Station 10:23-36. Coetzee C G 1983 An analysis of the distribution patterns of the Namibian terrestrial mammals (bats excluded). Annales Musee Royal de L'Afrique Centrale, Tervuren, Belgique 237:63-73. Eiten G 1974 An outline of the vegetation of South America. Proceedings of the 5th Congress of International Primatological Society, Nagoya, Japan, 529-545. Emlen S T & Oring L W 1977 Ecology, sexual selection, and the evolution of mating systems. Science 197:215-244. Henschel J R, Tilson R & von Blottnitz F 1979 Implications of a spotted hyaena bone assemblage in the Namib Desert. South African Archeology Bulletin 34:127-131. Hoeck H 1975 Differential feeding behavior of the sympatric hyrax Procavia johnstoni and Heterohyrax brucei. Oecologia 22:15-47. Hoeck H 1982 Population dynamics, dispersal, and genetic isolation in two species of hyrax ( Heterohyrax brucei and Procavia johnstoni) on habitat islands in the Serengeti. Zeitschrift fur Tierpsychologie 59:177-210, Hoeck H, Klein, H & Hoeck P 1982 Flexible social organization in hyrax. Zeitschrift fur Tierpsychologie 59:265-298. Joubert E & Mostert P K N 1975 Distribution patterns and status of some mammals in South West Africa. Madoqua 9:5-44. Lacher T E Jr 1981 The comparative social behavior of Kerodon rupestris and Galea spixii and the evolution of behavior in the Caviidae. Bulletin of the Carnegie Museum of Natural History 17:1-71. Mares M A 1975 Observations of Argentine desert rodent ecology, with emphasis on water relations of Eligmodontia typus. In: Rodents in Desert Environments (eds I Prakash & P K Ghosh). Junk BV, The Hague, Netherlands, 155-175. Mares M A 1992 Neotropical mammals and the myth of Amazonian biodiversity. Science 255:976-979. Mares M A 1993 Heteromyids and their ecological counterparts: A pandesertic view of rodent ecology. In: The Biology of the Family Heteromyidae (eds H H Genoways & J H Brown). Special Publication 10, American Society of Mammalogists, Lawrence, Kansas, 652-713. Mares M A & Lacher T E Jr 1987 Ecological, morphological, and behavioral convergence in rock-dwelling mammals. Current Mammalogy 1:307-348. Mares M A & Ojeda R A 1982 Patterns of diversity and adaptation in South American hystricognath rodents. In: Mammalian Biology in South America (eds M A Mares & H H Genoways). Pymatuning Laboratory of Ecology, Special Publication 6, Linesville, Pennsylvania, 393-432. Mares M A, Morello | & Goldstein G 1985b The Monte Desert and other subtropical semiarid biomes of Argentina, with comments on their relation to North American arid areas. In: Ecosystems of the World 12A: Hot Deserts and Arid Shrublands (eds M Evenari, I Noy-Meir & D W Goodall). Elsevier, Amsterdam, 203-237. Mares M A, Ojeda R A, Braun J K & Barquez R M 1997 Observations on the mammals of Catamarca Province, Argentina: systematics, distribution, and ecology. Special Publication, Museum of Southwestern Biology, University of New Mexico, Albuquerque, 3, 89-141. Mares M A, Ojeda R A & Barquez R M 1989 Guide to the Mammals of Salta Province, Argentina. The University of Oklahoma Press, Norman, Oklahoma. Mares M A, Willig M R & Lacher T E Jr 1985a The Brazilian Caatinga in South American zoogeography: Tropical mammals in a dry region. Journal of Biogeography 12:57-69. Mares M A, Willig M R, Lacher T E Jr & Streilein, K E 1981 The mammals of northeastern Brazil: A preliminary assessment. Annals of the Carnegie Museum 50:80-137. Markham C G 1972 Aspectos Climatologicos da Seca no Brasil- Nordeste. Sudene Assessouria Tecnica, Divisao do Documentacao, Recife. Merkt J R 1987 Reproductive seasonality and grouping patterns 138 Journal of the Royal Society of Western Australia, 80(3), September 1997 of the north Andean deer or taruca ( Hippocamelus antisensis) in southern Peru. In: Biology and Management of the Cervidae (ed C M Wemmer). Smithsonian Institution Press, Washington, DC, 388-401. Ojeda, R A & Mares, M A 1989 A zoogeographic analysis of the mammals of Salta Province, Argentina: patterns of community assemblage in the Neotropics. Special Publications, Texas Tech University 27:1-66. Pearson O P 1948 Life history of the mountain viscacha in Peru. Journal of Mammalogy 29:345-374. Pearson O P & Ralph C P 1978 The diversity and abundance of vertebrates along an altitudinal gradient in Peru. Memorias del Museo de Historia Natural "Javier Prado" 18:1-97. Reis A C de Souza 1976 Clirna da Caatinga. Anales Academia Brasileira de Ciencia 48:325-335. Rosenzweig M L 1995 Species Diversity in Space and Time. Cambridge University Press, Cambridge. Seely M 1987 The Namib: Natural History of an Ancient Desert. Shell Namibia Limited, Windhoek. Seely M 1990 Namib Ecology: 25 Years of Namib Research. Transvaal Museum Monograph 7, Pretoria. Simpson G G 1964 Species density of North American Recent mammals. Systematic Zoology 13:57-73. Skinner J D & Smithers R H N 1990 The Mammals of the Southern African Subregion. University of Pretoria, Pretoria. Streilein, K E 1982a Ecology of small mammals in the semiarid Brazilian Caatinga. I. Climate and faunal composition. Annals of the Carnegie Museum 51:79-107. Streilein, K E 1982b Ecology of small mammals in the semiarid Brazilian Caatinga. III. Reproductive biology and population ecology. Annals of the Carnegie Museum 51:251-269. Streilein, K E 1982c Ecology of small mammals in the semiarid Brazilian Caatinga. IV. Habitat selection. Annals of the Carnegie Museum 51:331-343. Streilein, K E 1982d Ecology of small mammals in the semiarid Brazilian Caatinga. II. Water relations. Annals of the Carnegie Museum 51:109-126. Stuart C T 1975 Preliminary notes on the mammals of the Namib Desert Park. Madoqua, Series II, 4:5-68. Theron G K, van Rooyen N & van Rooyen M W 1980 Vegetation of the lower Kuiseb River. Madoqua 11:327^135. Tilson R 1980 Klipspringer ( Oreotragus oreotragus) social structure and predator avoidance in a desert canyon. Madoqua 4:303-314. Tilson R, von Blottnitz F & Henschel J 1980 Prey selection by spotted hyaena (Crocuta crocuta ) in the Namib Desert. Madoqua 12:41-49. Videla M A & Suarez J A 1991 Mendoza Andina: Precordillera Alta Cordillera. Editorial Adalid, Mendoza, Argentina. Walter H 1986 The Namib Desert. In: Ecosystems of the World 12B: Hot Deserts and Arid Shrublands (eds M Evenari, I Noy-Meir & D W Goodall). Elsevier, Amsterdam, 245-282. Ward J D & Corbett I 1990 Towards an age for the Namib. In: Namib Ecology: 25 years of Namib Research (ed M K Seely). Transvaal Musuem Monograph 7, Pretoria, 17-26. Webb K 1974 The Changing Face of Northeast Brazil. Columbia University Press, New York. Werger M J A (ed) 1978 Biogeography and Ecology in Southern Africa. W Junk BV. The Hague, Netherlands. Willig M R 1983 Composition, microgeographic variation, and sexual dimorphism in Caatingas and Cerrado bat communities from Northeast Brazil. Bulletin of the Carnegie Museum of Natural History 23:1-131. Willig M R & Jones J K Jr 1985 Neoplatymops mattogrossensis. Mammalian Species 244:1-3. Willig M R & Sandlin E A 1991 Gradients of species density and turnover in New World bats: A comparison of quadrat and band methodologies. In: Latin American Mammalogy: History, Biodiversity, and Conservation (ed M A Mares and D J Schmidly). University of Oklahoma Press, Norman, Oklahoma, 81-96. Wilson J W III 1974 Analytical zoogeography of North American mammals. Evolution 28:124-140. Withers P C 1979 Ecology of a small mammal community on a rocky outcrop in the Namib Desert. Madoqua Series II, 12:229-246. Withers P C 1983 Seasonal reproduction by small mammals of the Namib Desert. Mammalia 47:195-204. 139 Journal of the Royal Society of Western Australia, 80:141-158, 1997 Plants of Western Australian granite outcrops S D Hopper1' A P Brown2 & N G Marchant3 'Kings Park and Botanic Garden, West Perth WA 6005 email: steveh@kpbg.wa.gov.au •Western Australian Threatened Species & Communities Unit, Department of Conservation and Land Management, PO Box 51, Wanneroo WA 6065 email: andrewbr@calm.wa.gov.au Western Australian Herbarium, Department of Conservation and Land Management, PO Box 104, Como WA 6152 Abstract Outcropping granite rocks in Western Australia span a considerable climatic range, from the mediterranean south-west to inland desert and northern arid subtropics and tropics. At least 1320, and possibly 2000, plant taxa occur on Western Australian granite outcrops. Outcrop plant life is most diverse in the South West Botanical Province, with individual outcrops having up to 200 species, including many endemics not found in surrounding habitats. Species richness and local endemism declines with increasing aridity, to the point where Kimberley and Pilbara outcrops show little discontinuity in species from the surrounding landscape matrix. Outcrops are dominated by woody and herbaceous perennials, especially of the Myrtaceae, Orchidaceae, and Mimosaceae, and have an unusually rich diversity of annuals (Asteraceae, Stylidiaceae, Poaceae, Amaranthaceae etc.) compared with the flora as a whole. An unusual life form is found in resurrection plants capable of extreme desiccation and rehydration (e.g. Borya, Cheilanthes). Among woody perennials, bird pollination is frequent, and some outcrops harbour a high proportion of obligate seeder species due to the refuge from fire provided by bare rock barriers. The diversity of microhabitats and soil moisture regimes on outcrops has enabled the persistence of species beyond their main range in the face of climatic fluctuations. It has also facilitated the evolution of many endemics in the south-west. Major threatening processes facing outcrop plant communities include weed invasion, grazing by stock and feral animals, too-frequent fire, clearing, loss of shrub layer, salinity, and dieback. Conservation of these rich rare habitats needs the support of local communities. Introduction Western Australia is the largest Australian State, extending 2 391 km north-south and 1 621 km east-west, and occupying 2 525 500 km2, or about a third of the nation. Climates range from temperate and mediterranean in the south-west to the monsoonal tropics in the extreme north, with arid conditions prevailing over much of the State from the Nullarbor Plain north through the inland deserts to the north-west Pilbara and Kimberley regions. Most of the State is a plateau 300-600 m above sea level, with the highest point (Mt Meharry) only 1 251 m. Western Australia's numerous granite inselbergs and outcrops, consequently, are significant landscape features, even though most emerge less than 100 m above surrounding terrain. Granite (or granitoid) rocks are mainly found on the Yilgarn and Pilbara Cratons, which together occupy about a third of the State's landmass, and along adjacent south coastal and Kimberley orogens (Myers 1997). The outcrops thus span a considerable climatic range, from the mediterranean south-west to inland desert and northern arid subtropics and tropics. Western Australian outcrops vary in area from about © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 one tenth of a hectare to many hectares. While most are clearly isolated and disjunct, sometimes separated from the next outcrop by tens of kilometres, others are part of extensive inselberg systems covering several square kilometres (e.g. outcrop systems in Cape Le Grand, Porongurups and Cape Arid National Parks on the south coast, at Oudabunna Station near Paynes Find, and at Woodstock Station in the Pilbara). Peak Charles, 100 km north-west of Esperance, stands highest above surrounding country among the State's inselbergs, reaching 500 m from a low plain. While the Kimberley, Pilbara and desert granites are in regions with depauperate floras not unlike those of much of tropical and arid Australia, the granite of the Yilgarn Craton underlies one of the world's regions richest in plant life, the South West Botanical Province (Marchant 1973; Hopper 1979, 1992; Beard 1980, 1981, 1990). As many as 8000 vascular plant species occur in this region, with about 75% endemism. Consequently, the botanical exploration of south-western inselbergs has been particularly rewarding, and much remains to be done. Here, we provide an overview of current knowledge of plant life on Western Australian granite outcrops. Our aim was to compile a baseline, including a preliminary list of plants (vascular and cryptogam) collected from Western Australian outcrops, and a list of key references, to inform future research and provide a context for more specific studies. We conclude with some thoughts on the 141 Journal of the Royal Society of Western Australia, 80(3), September 1997 conservation of this special component of Western Australian plant life. Botanical Exploration Western Australian granite outcrops have been important to aboriginal people ever since they occupied the land, but little has been documented of this long¬ standing relationship (Bindon 1997). Recorded European knowledge of granite outcrop plants in the State commenced with Archibald Menzies, surgeon and naturalist aboard the HMS Discovery, under the command of Captain George Vancouver. The Discovery was anchored in King George's Sound from September 28 to October 11, 1791. Menzies made a "copious collection of ... vegetable productions, principally the genus Banksia, which are here very numerous" from various sites onshore in the vicinity of present-day Albany (Maiden 1909). Given the prominence of granite headlands, hills and islands at Albany, Menzies undoubtedly collected from them, but details are difficult to trace. European botanists first named granite outcrop plants from Western Australia in 1800. The Frenchman Jacques J H de Labillardiere was naturalist on Admiral Bruny d'Entrecasteaux's expedition to Australia, which sought refuge from a storm near the modem town of Esperance over the period 9-17 December 1792. Labillardiere's specimens included a plant he named in 1800 as Eucalyptus cornuta, collected from the granite Observatory Island on 13 December 1792. E. cornuta occupies the apron and deeper soil pockets on granite outcrops throughout the Esperance area and adjacent islands, extending westwards to the Cape Naturaliste area (Fig 1). Labillardiere named other granite outcrop plants from the expedition, including the bizarre genus Borya (Fig 4), a pincushion lily and resurrection plant, named in 1805 and typified by the species B. nitida which occurs on shallow soils on outcrops along the south coast from Albany to east of Esperance. Robert Brown, on the expedition to circumnavigate Australia led by Matthew Flinders in the Investigator, anchored at King George's Sound on December 8, 1801, and collected extensively until January 5, 1802. Between January 10-18, the expedition also explored and collected near Esperance, landing at Lucky Bay (in the present-day Cape Le Grand National Park) for five days, and on islands in the Recherche Archipelago. Brown (1810) also named many granite outcrop plants, including another species of Borya (B. spaherocephala ) that occurs with B. nitida on several south coast outcrops. The arrival of James Drummond at the Swan River with Captain Stirling's colonising party on the Parmelia in 1829 heralded a major advance in knowledge of the south-west's flora, including that of granite outcrops (Erickson 1969). Drummond, a Scot and keen nurseryman with an excellent eye for new taxa, collected widely throughout the forests and present day wheatbelt, his specimens destined for subscribers in Great Britain and Europe. He was followed by many others up to modern times. Beard (1981) provides a useful historical review. The floral wealth of the south¬ west is such that significant numbers of new taxa continue to be discovered and described each year (Green 1985; Hopper 1992), including granite outcrop endemics. The ecology and biogeography of Western Australian granite outcrop plants were observed by the earliest collectors, but drew specific mention by the plant geographer Ludwig Diels (1906), who noted the "dwarflike" vegetation. Smith (1962) provided an account of the vegetation of the granite Porongurup Range, and many others have mentioned granite outcrop plants in broader vegetation surveys in the south-west (e.g. Beard 1981; Newbey & Hnatiuk 1985; Newbey et al. 1995; Wardell Johnson & Williams 1996; Brooker & Margules 1996). Main (1967) gave a delightful account of the natural history of Yorkrakine Rock through an annual cycle. Marchant (in Erickson et al. 1973) summarised botanical features of south-western outcrops in an essay for a general readership. Schweinfurth (1978) described the vegetation of exposed granite headlands along the south coast, while Abbott & Watson (1978) and Abbott (1980) explored these headlands and adjacent islands in greater detail, showing the flora of most islands had a depauperate subset of that found onshore, and displayed little endemicity nor genetic differentiation from mainland populations. Hopper (1981) highlighted the importance of winter¬ flowering woody perennials as a nectar resource used by a diversity of honeyeaters in the central wheatbelt (Fig 3). Pate & Dixon (1982) gave a comprehensive account of the biology of tuberous and cormous plants, including many found on granite. Ornduff (1987) documented the flora of herbfields including, and centrewards from, the Bojya zone on nine outcrops near Perth and Hyden. Burgman (1987) sampled five outcrops inland from Esperance and explored aspects of rarity in granite communities. Hopper et al. (1990) illustrated 29 endangered endemics of granite outcrops and highlighted conservation issues. Pignatti & Pignatti (1994) reported on herbfields from shallow seasonally wet soils on a selection of south-west outcrops. Ohlemuller (1997) documented plants in shallow soil depressions on 26 outcrops throughout the south-west and goldfields. In parallel with these ecological investigations, genetic studies of Western Australian granite outcrop plants were pioneered by James (1965), who's penetrating work on the herb Isotoma petraea (Lobeliaceae) has become a classic in the literature on plant evolution (Bussell & James 1997). Studies on the population genetics of endemic granite outcrop eucalypts have also provided useful insights into the evolution of highly fragmented small populations (Moran and Hopper 1983; Sampson et al. 1988). The breeding systems of granite outcrop endemics have been documented for Villarsia (Menyanthaceae; Ornduff 1986, 1996) and Anthocercis gracilis (Solanaceae; Stace 1995). Sampling Methods Rather than laboriously search the increasing number of botanical publications that include plant taxa recorded 142 Journal of the Royal Society of Western Australia, 80(3), September 1997 on Western Australian granite outcrops, we compiled the plant list from specimens housed in the Western Australian Herbarium. The database WAHERB stores full label details of 413 000 specimens of vascular plants and cryptogams collected throughout the State, and afforded an opportunity to list those taxa for which habitat details on labels included the words "granite outcrop". Initial trials indicated that searching on "granite" or "granitic" alone picked up too many taxa that occupied regions underlain by granite rock, but favoured deep soils well removed from rock outcrops. In our list, taxa are mainly species, but infraspecific subspecies, varieties and forms, together with hybrids, are all counted. This ensures that all named plant biodiversity is included, and allows for discrepancies in assigning rank among taxonomists. The above approach to listing plant taxa was based on an understanding that most authors involved in botanical survey of Western Australian granite outcrops have collected voucher specimens and deposited them in the Western Australian Herbarium ( e.g . Ornduff 1987; Burgman 1987; Hopper 1981; Hopper ct al. 1990; Hopper & Brown, unpublished orchid research ; Newbey & Hnatiuk 1985; Newbey et ah 1995). In addition, as an independent check, the senior author has compiled field herbaria in notebooks for ca 150 outcrops throughout Western Australia over the past two decades. Most outcrops sampled occur in the South West Botanical Province (e.g, Fig 1) and adjacent pastoral region, while a few have been examined in the Pilbara (e.g. Fig 2) and Kimberley. Precisely defining the boundaries of a granite outcrop is difficult, especially for mobile components of the biota (Main 1997). However, the sedentary nature of adult plants, especially perennials, affords a reliable and measurable biological indicator of the limits of the outcrop system. In fieldwork, we have used the distribution of plants that are locally endemic to outcrops to define the extent of outcrop communities sampled. Thus, all species centrewards from the outermost individuals of granite outcrop local endemics were included as present on an outcrop. This approach ensures that all endemics are sampled, but has the effect of picking up some species from surrounding deeper soils that do not extend onto soil pockets on the outcrop itself. Other authors have restricted their sampling to such soil pockets (e.g. Houle 1990; Porembski et al. 1995; Ohlemuller 1997), or only to herbfields (e.g. Ornduff 1987). Experience showed that perennial vegetation fringing the base of Western Australian outcrops had to be sampled to ensure a comprehensive inventory (see below). We followed Newbey and Hnatiuk (1985) and Newbey et al (1995) in collecting on random walks stratified by microhabitat rather than within quadrats i.e. the random stratified walk technique. A comparison of the senior author's data with the quadrat-based data of Burgman (1987) for Mt Ney inland from Esperance showed that four 5m x 5m quadrats spaced at 25 m intervals on NE scree slope and sheet rock yielded 61 taxa, whereas 192 taxa were documented through random stratified wralks covering cardinal compass points. Microhabitats recognised and deliberately searched for on random stratified walks included bare rock, cryptogamic crusts, gnammas (rock pools or weather pits, seasonal and permanent), soil-filled crevices, caves/ tafoni /shade of boulders or exfoliated slabs, herbfields on shallow well-drained soil, herbfields on shallow seasonally waterlogged soil (ephemeral flush vegetation), shrublands and woodlands on deep well-drained soil, shrublands, woodlands and forests on deep seasonally waterlogged soil, permanent springs, major watercourses, and salt lakes. Floristics A total of 1320 Western Australian granite outcrop plant taxa were represented and labelled as such in the collections of the Western Australian Herbarium in May 1996 (Appendix 1). Of these, 1097 taxa (83.0%) were described, 94 (7.1%) were undescribed but well-delimited with manuscript names, and 129 (9.9%) were of uncertain taxonomic delimitation. Only 4 (0.3%) were hybrids, well below the proportion (4%) estimated for the whole WA flora (Hopper 1995). The list includes 11 macrofungi, 34 lichens, 46 mosses, 9 liverworts, 9 ferns, 7 fern allies, 2 cycads, 2 conifers, and 1200 angiosperms (284 monocotyledons, 916 dicotyledons). While these statistics provide a useful baseline, it is pertinent to note that sampling of most cryptogam groups is poor. For example, we recorded only 34 lichens with vouchered granite outcrop specimens for the whole of WA, whereas Pigott & Sage (1997) list 36 lichen taxa for Yilliminning Rock alone. Macrofungi undoubtedly are orders of magnitude more diverse than presently known. The moss flora is comparatively better documented at the species level (Stoneburner et al. 1993; Stoneburner & Wyatt 1996), but much more collecting is needed to fill in knowledge of distribution and ecology. Some confidence may be placed in the list of vascular plant taxa, which accounts for 1220 (92.4%) of the total of 1320. Based on our knowledge of rock outcrop floras across the State, the majority of south-western vascular plant taxa on granite outcrops are listed. However, some conspicuous omissions of taxa that are common (e.g. Hakea petiolaris, Corymbia calophylla), uncommon (e.g. Grevillea magnifica, Phylloglossum drummondii and Isoetes drummondii ) and rare (e.g. Drummondita hassellii var longifolia, Hibbertia bracteosa, Villarsia calthifolia , Pleurophascum occidentale) indicate that revision upwards is required. Moreover, a comparison of the list of 187 taxa for Yilliminning Rock near Narrogin (Pigott & Sage 1997) with Appendix 1 shows that 78 Yilliminning taxa (42%) are not present in our list for the State. This discrepancy would be even more so for northern outcrops. For example, absent from the present list are the dominant spinifex hummock grasses on granite outcrops (Triodia spp), as well as common lemon grasses (Cymbopogon spp), and scattered but widespread subtropical woody shrubs and small trees such as Terminalia canescens. It would be no surprise to us, consequently, if the number of vascular plant taxa on Western Australian granite outcrops approached 2000 in the future. In the present list, major vascular plant families include the Myrtaceae (162 taxa), Orchidaceae (159), Mimosaceae 143 Journal of the Royal Society of Western Australia, 80(3), September 1997 (82), Asteraceae (72), Papilionaceae (66), Proteaceae (54), and Stylidiaceae (33). Genera richest in taxa on granite outcrops include Acacia (82 taxa), Caladenia (70), Eucalyptus (44), Stylidium (28), Melaleuca (26), Grevillea (25), Drosera (21), Verticordia (20), Eremophila (18) and Thelymitra (16). Thus, woody perennial shrubs and trees of myrtles, wattles, peas and Proteaceae are taxon-rich on WA granite outcrops, reflecting trends in the flora as a whole (Green 1985; Hopper et al. 1996). However, the granite woody perennials are depauperate in Epacridaceae, which is rich in other habitats. The granite outcrop flora is clearly divergent from the flora at large in the unusually high taxon richness of groups such as the tuberous perennial terrestrials (orchids, sundew's, lilies), and herbaceous, often annual, daisies and triggerplants. Systematic sampling of shallow-soil herbfields on south-western outcrops by Ornduff (1987) and Ohlemuller (1997) yields a useful comparison with our data on herbaceous families. Ornduff recorded 160 taxa belonging to 37 families (he didn't sample Orchidaceae for conservation reasons), of which the largest were Asteraceae, Poaceae, Liliaceae sens, lat., and Stylidiaceae. Ohlemuller recorded 134 species from 42 families, with the largest being Asteraceae, Orchidaceae, Poaceae, Apiaceae, Stylidiaceae and Centrolepidaceae. Major herbaceous families in our list are Orchidaceae (Fig 5), Asteraceae, Stylidiaceae, Anthericaceae (or Liliaceae sens. lat.), Poaceae, Droseraceae, Cyperaceae, Goodeniaceae, Amaranthaceae and Centrolepidaceae. The Amaranthaceae is a family richest in the north of Western Australia, so it is not surprising that Ornduff (1987) and Ohlemuller (1997) found it to be depauperate in their south-wTest study areas. Orchidaceae undoubtedly emerges as the most taxon-rich herbaceous family in our list because two of us (SH and AB) have made a detailed and prolonged study of granite outcrop orchids, and found that repeat surveys over different seasons and years are needed to develop comprehensive orchid inventories for individual rocks. Other differences between our ranking and those of Ornduff (1987) and Ohlemuller (1997) are minor. This general concordance suggests that major trends and statistics for the vascular plants in Appendix 1 are accurate, at least for south-west outcrops. Plants of Special Interest Granite outcrop habitats have elicited remarkable parallel evolution in many plants and animals (Porembski et al. 1997; Wyatt 1997; Mares 1997). The combination of high solar radiation, rapid runoff of rainfall, and shallow soils on a rocky substrate provide microhabitats of accentuated seasonal and diurnal stresses. Conditions may vary over a few metres from cool permanently moist shaded caves with water seepage to drought- afflicted shallow soils and rock surfaces fully exposed to all the elements. Few' organisms can tolerate the harshest of these rock- surface habitats, which tend to be occupied by cryptogamic crusts of cosmopolitan cyanobacteria, lichens and mosses such as Grimmia laevigata. Western Australian outcrops conform to this trend. However, on south-west outcrops experiencing high rainfall, moss cushions become prominent, with the luxuriant brown carpets of Breutelia affinis and the verdant green Campylopus bicolor and C. australis noteworthy. Coping with seasonal or unpredictable drought is, undoubtedly, the most significant survival strategy faced by granite outcrop plants. Among Western Australian herbaceous perennials on outcrops, pincushion lilies (Bon/a spp, Boryaceae) are abundant and remarkable in their capacity as resurrection plants (Gaff 1981) to withstand desiccation to less than 5% of normal leaf moisture content, turning orange in the process, and rehydrating to normal green leaves within a day of rainfall (Fig 4). There are at least six species of Borya in the South West Botanical Province (B. sphaerocephala is a variable taxon awaiting detailed study and possible subdivision). These plants do not extend to the Pilbara outcrops. Other resurrection plants commonly seen on WA granites include the rock ferns Cheilanthes spp and Pleurosorus rutifolius. Persisting underground as a tuber during drought is a common strategy among Western Australian granite outcrop herbs in the south-west and adjacent pastoral regions, especially lilioids ( e.g . Wurmbea spp, Colchicaceae), orchids ( Caladenia , Thelymitra , Pterostylis, Diuris , Prasophyllum) and sundews (Drosera), as well as the unusual fern ally Phylloglossum (Pate & Dixon 1982). On northern outcrops, tuberous Cyperaceae tend to occupy this niche. Succulence is not prominent in the Western Australian granite outcrop perennial floras, but has evolved in a few taxonomically-unrelated species e.g. Spiculaea ciliata (Orchidaceae; Fig 5), Carpobrotus spp (Aizoaceae). Some perennial shrubs of south coast granites have unusually thick leaves (e.g. Hakea clavata, Proteaceae; Anthocercis viscosa , Solanaceae), and at least one has tuberous roots (Calothamnus tuberosus, Myrtaceae). Sclerophyllous leaves characterise the vast majority of woody perennials found on granite outcrops, as well as in the Western Australian flora at large. The graminoid habit has evolved in several unrelated groups in the south-west, with common perennials forming tussocks including diverse Cyperaceae (especially Lepidosperma spp), blind grass (Stypandra spp, Phormiaceae), the sedge-like grass Spartochloa scirjioidea (Poaceae) and lemon grass Cymbopogon spp (Poaceae). Desert, Pilbara and Kimberley outcrops are dominated by uniquely Australian hummock grasses with inrolled pungent leaves (Triodia spp, Poaceae). Drought avoidance through an annual life history is a major feature of Western Australian granite herbs in families such as the Asteraceae, Stylidiaceae, Poaceae, Goodeniaceae, Amaranthaceae and Centrolepidaceae. Most of these annuals have relatively small inconspicuous flowers suggestive of inbreeding (Ornduff 1987), but some are brightly coloured and adundant. Most outcrops in southern Western Australia are bedecked with colourful swards of annual triggerplants (Stylidium, Stylidiaceae) and everlastings (e.g. Rhodanthe, Waitzia, Asteraceae) that enliven the herbfields. Succulence is a feature of some annuals, especially those that occupy the dry spectrum of shallow soils (e.g. Calandrinia spp, Portulacaceae; Crassula spp, Crassulaceae). 144 Journal of the Royal Society of Western Australia, 80(3), September 1997 1. Coastal granite outcrops east of Albany on Two Peoples Bay Nature Reserve, with Eucalyptus cornuta (pale canopy), described in 1800 by French scientific explorer Labillardiere. Photograph by SD Hopper. 2. Granite sheet, boulders and distant inselbergs in the arid Pilbara, Spear Hill, south-west of Marble Bar. Photograph by SD Hopper. 3. Bird-pollinated plants are a feature of south-west Australian granite outcrops - New Holland Honeyeater on Eucalyptus caesia , Boyagin Rock, north-west of Pingelly. Photograph by SD Hopper. 4. The resurrection plant, pincushion lily Borya const ricta, with leaves 1 cm long desiccated (orange) but rehydrating (green) following rain at the end of summer drought, Chiddarcooping Rock, north-east of Merredin. Photograph by SD Hopper. 5. Terrestrial orchids are the second most species-rich plant family on Western Australian granite outcrops. Here, flowers of the elbow orchid (Spiculaea ciliata) 1 cm long splay from the succulent stem that feeds them in early summer, and a male thynnid wasp attempts to copulate with the flower's sexual decoy, an insectiform hinged labellum, Boulder Rock, Brookton Highway. Photograph by BA & AG Wells. 6. Arguably the rarest Western Australian granite outcrop plant, the annual aquatic millfoil Myriophyllum lapiiiicola in full flower (with round floating leaves), and known from just two gnammas (rock pools), alongside the invasive aquatic South African weed, Crassula natans var minus. Photograph by SD Hopper. 145 Journal of the Royal Society of Western Australia, 80(3), September 1997 Annuals are also found commonly in seasonally waterlogged shallow soils or ephemeral flush communities (Pignatti & Pignatti 1994). These annuals include species of Centrolepidaceae ( Centrolepis and Aphelia), Apiaceae (Hydrocotyle), Juncaginaceae (' Triglochin ), Cyperaceae ( Schoenus , Isolepis, Cyperus), Asteraceae (Quinetia, Millotia, Rutidosis, Siloxerus, Toxanthes, Hyalosperma, Podotheca etc.) and Stylidiaceae (Sty lidiu m, Leven hookia) . Deeper soils enable survivorship of woody perennials, of which the largest on Western Australian outcrops include eucalypts (Eucalyptus and Corymbia, Myrtaceae), wattles ( Acacia , Mimosaceae), she-oaks (Alloc asuarina, Casuarinaceae), and rock figs (Ficus, Moraceae). A noteworthy feature of the woody perennials is the high proportion of bird-pollinated species (Hopper 1981; Fig 3) in families such as Myrtaceae (e.g. Eucalyptus, Calothamnus, M elaeuca, Verticordia), Proteaceae (Grevillea, Eiakea, Banksia), Myoporaceae (Eremophila), Papilionaceae (Kennedya, Crotolaria), Viscaceae (Amyema, Lysiaua) and Sterculiaceae (Brachy chiton). Also, south-western outcrops have an unusually high number of woody perennials that are obligate seeders - plants that are killed by fire and recruit only from seed. For example, 77% of the perennials in a granite community at Chiddarcooping Nature Reserve north-east of Merredin were obligate seeders (Hopper et at., unpubi). Gnammas on Western Australian outcrops have a unique flora comprising annual species of quillworts (Isoetes, lsoetaceae), mudmats (Glossostigma, Scrophulariaceae), millfoils (Myriophyjlum, Haloragaceae; Fig 6) and others, including the introduced South African annual Crassula natans (Crassulaceae). Weeds Ornduff (1987) and Ohlemiiller (1997) found that introduced weed species, mainly annuals, comprised 23.7% and 17.0% of the granite herbfield floras sampled respectively. These figures are high compared with the ca 9% that weeds represent of the WA flora as a whole (Green 1985; Keighery 1995). They are even more significant when considered with granite outcrop herbfield floras elsewhere on earth, which have far fewer invasive weeds (Porembski ct al. 1997; Wyatt 1997). Weed growth is most pronounced in full sun on outcrops, especially where soil has been disturbed and enriched by rabbit dung or agricultural activity. In these situations, annual grasses such as Briza maxima, Avena fatua and Ehrharta longifolia often dominate and replace native annuals throughout much of the south-west. Weeds are rare only where a dense shrub layer or low forest of native woody perennials persist. It is clear that a persistent seed rain of weeds occurs over vast regions of the south-west, as weeds appear in open areas on outcrops where little or no disturbance is evident and native plants dominate. Hopper (1997) has attributed the high invasibility of disturbed Western Australian plant communities to the absence of major glacial soil stripping as an evolutionary force acting on the flora. Native species are unable to compete against weeds from habitats where soil disturbance is a regular pertubation. Further experimental, study of weeds in granite herbfields could test this hypothesis, and assist attempts at restoration of invaded outcrops. Biogeography and Endemism Western Australian granite outcrops display plant biogeographic patterns that mirror that of the whole flora (Hopper 1979, 1992; Hopper et al. 1996); species-richness and endemism are pronounced in the transitional rainfall zone (wheatbelt) of the south-west, and attenuate as rainfall decreases through the pastoral country to the deserts, Pilbara and Kimberley. For example, the total number of vascular plant taxa recorded by the senior author and colleagues on a sample of rocks ranged from 142 (Point Matthew, near Augusta) and 201 (Mt Frankland) in the highest rainfall forests, 192 (Mt Ney) and 187 (Yilliminning Rock near Narrogin; Pigott and Sage 1997) in the transitional rainfall zone, to 85 (Daggar Hills, near Yalgoo) in the pastoral zone, and 90 (Moolyella Rocks, east of Marble Bar) and 80 (Spear Hill, west of Marble Bar; Fig 2) in the arid Pilbara. There are no vascular species shared between northern (Kimberley, Pilbara) outcrops and those in the south-west. Moreover, the northern outcrop floras are virtually identical with that from the matrix of surrounding terrain, save for outcrop specialists like rock figs and rock ferns. Walters & Wyatt (1982) similarly recorded low endemism and little discontinuity between vascular plants on granite outcrops and adjacent landforms of the arid Central Mineral Region of Texas. In contrast, south-western and adjacent pastoral zone rocks have higher levels of local endemism, especially in the high rainfall forest region where the outcrops present the most striking difference in habitat to the surrounding vegetation matrix (e.g. Wardell Johnson & Williams 1996; Brooker & Margules 1996). Biogeographical relationships of outcrop floras across the south-west are under ongoing study by the senior author. As a precursor. Hopper & Brown (unpublished) documented the distribution of 126 orchid taxa on 41 outcrops ranging from the highest rainfall forests through the transitional zone wheatbelt to the arid zone. Each rock outcrop was treated as a site in a classification of the orchid data. The study highlighted a number of significant trends. A primary division occurred between the 15 rocks found in the forested High Rainfall Zone ( >800 mm p.a.) and the rest ranging from the Transitional Rainfall Zone of the wheatbelt into the Arid Zone. Subsequent divisions established as much difference among forest rocks as among the wheatbelt/ arid rocks, even though the forest rocks were confined to a much smaller area. Moreover, remarkably, closely adjacent rocks were widely separated in the classification, indicating significant differences in their orchids (e.g. 10.3 km Rock and 10.9 km Rock of Ornduff' s (1987), separated by just 600 m of jarrah forest on Albany Highway). Conversely, rocks separated geographically often had similar orchid floras (e.g. Boyagin Rock and Pingaring Rock on the western and eastern sides of the south-central wheatbelt respectively). These patterns suggest significant barriers to orchid 146 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 1 Lists of Western Australian orchid taxa that occur on granite outcrops. All taxa Caladcnia attingens subsp a tt ingens ms Caladenia attingens subsp gracillima ms Caladcnia brevisura ms Caladenia brownii ms Caladenia caesarea subsp maritima ms Caladenia caesarea subsp tran&iens ms Caladenia citrina ms Caladenia denticulata Caladenia dimidia ms Caladenia discoidea Caladenia doutchiae Caladenia exstans ms Caladenia falcata Caladenia Jilifera Caladenia flaccida subsp flaccida ms Caladenia flaccida subsp pulchra ms Caladenia flaua subsp flava ms Caladenia flava subsp maculata ms Caladenia flava subsp sylvestns ms Caladenia footeana ms Caladenia granitora ms Caladenia heberleana ms Caladenia hirta subsp rosea ms Caladenia hoffmanii subsp gramticola ms Caladenia incensa ms Caladenia incrassata ms Caladenia infundibularis Caladenia Integra Caladenia latifolia Caladenia lobata Caladenia longicauda subsp clivicola ms Caladenia longicauda subsp eminens ms Caladenia longicauda subsp longicauda ms Caladenia longicauda subsp rigidula ms Caladenia longiclavata Caladenia macrostylis Caladenia marginata Caladenia microchila ms Caladenia multiclavia Caladenia nivalis ms Caladenia pachychila ms Caladenia pholcoidea ms Caladcnia polychroma ms Caladenia radialis Caladenia replans subsp impensa ms Caladenia replans subsp reptans ms Caladenia rhomb oidif or mis Caladenia roei Caladenia saccharata Caladenia s erotina ms Caladenia sigmoidea Caladenia splendens ms Caladenia hiemalis ms Caladenia horistes ms Caladenia pendens subsp pendens ms Caladenia remota subsp remota ms Caladenia pendens subsp talbotii ms Caladenia voigtii ms Corybas recurvus Cyanicula ampfexans ms Cyanicula ashbyae ms Cyanicula caerulea subsp apertala ms Cyanicula deformis ms Cyanicula fragrans ms Cyanicula gemma ta ms Cyanicula sericea ms Cyrtostylis huegelii Cyrtostylis robusta Diuris aff longifolia Diuris brumalis Diuris conspicillata Diuris laevis Diuris laxiflora Diuris longifolia Diuris maculata Diuris picta Diuris pulchella Diuris recurva Diuris setacea Drakonorchis barbarossa ms Drakonorchis drakeoides ms Drakonorchis mesocera ms Elythranthera brunonis Elythranthera emarginata Eriochilus dilatatus subsp dilatatus ms Eriochilus dilatatus subsp multiflorus ms Eriochilus dilatalus subsp undulatus ms Eriochilus helonomos ms Eriochilus pulchellus ms Eriochilus scaber Genoplesium nigricans Leptoceras menziesii Lyperanthus serratus Microtis aff parviflora Microtis atrata Microtis brownii Microtis eremaea Microtis graniticola Microtis media subsp eremicola Microtis media subsp media Monadenia bracteata Paracaleana nigrita Paracaleana triens ms Prasophyllum aff parvifolium Prasophyllum brownii Prasophyllum cucullatum Prasophyllum elatum Prasophyllum fimbria Prasophyllum gibbosum Prasophyllum gracile Prasophyllum parvifolium Prasophyllum ringens Pterostylis aff nana Pterostylis aff rufa Pterostylis allantoidea Pterostylis aspera Pterostylis barbata Pterostylis el ega ntis i m a Pterostylis hamiltonii Pterostylis mutica Pterostylis recurva Pterostylis roensis Pterostylis sanguinea Pterostylis sargentii Pterostylis sea bra Pterostylis vittata Pyr orchis nigricans Spiculaea ciliata Thelymitra aff holmsii Thelymitra aff longifolia Thelymitra aff nuda Thelymitra aff pauciflora Thelymitra antennifera Thelymitra benthamiana Thelymitra crinita Thelymitra cucullata Thelymitra aff dedmaniarum Thelymitra flexuosa Thelymitra macrophylla Thelymitra spiralis Taxa that predominantly occur on granite outcrops Caladenia caesarea subsp maritima ms Caladenia exstans ms Caladenia granitora ms Caladenia hoffmanii subsp graniticola ms Caladenia Integra Caladenia longicauda subsp clivicola ms Caladenia longicauda subsp rigidula ms Caladenia multiclavia Caladenia nivalis Caladenia remota subsp remota ms Cyanicula ashbyae ms Cyanicula fragrans ms Diuris conspicillata Diuris picta Diuris pulchella Eriochilus pulchellus ms Microtis eremaea Microtis graniticola Microtis media subsp eremicola Spiculaea ciliata Thelymitra aff nuda Thelymitra aff dedmaniarum Declared Rare orchid taxa of granite outcrops Caladenia caesarea subsp maritima ms Caladenia exstans ms Caladenia hoffmanii subsp graniticola ms Caladenia voigtii ms Thelymitra aff dedmaniarum dispersal, particularly between forest rocks, and high levels of local extinction and stochastic events underlying the presence of orchids on individual rocks in the south¬ west and adjacent arid zone. There have been dynamic climatic fluctuations across the south-west for several million years as Australia drifted northwards and arid conditions overtook much of central Australia (Hopper 1979; Hopper et al 1996). The diversity of microhabitats on granite outcrops provided refuge for plants adapted to both dry or wet conditions as the surrounding matrix waxed and waned climatically (Marchant 1973; Main 1997). Survivorship in small populations on granite refuges undoubtedly was a matter of chance in the face of such repeated climatic turmoil. 147 Journal of the Royal Society of Western Australia, 80(3), September 1997 Interestingly, the above conditions of small disjunct outcrop populations undergoing recurrent stresses is predicted as ideal for genetic divergence and speciation (Grant 1981). Is this prediction borne out by studies of Western Australian outcrop plants? Table 1 provides a recently updated list of 141 orchid taxa recorded from Western Australian granite outcrops. Of these, 22 (16%) are more or less endemic The endemics have geographical ranges from widespread on outcrops throughout the south-west (e.g. Spiculaea ciliata; Fig 5) to highly restricted to a few adjacent outcrops less than 10 km apart (e.g. Caladenia caesarea subsp maritima, Thelyrnitra aff dedmaniarum). In terms of evolutionary origins, these endemics display at least three patterns (Hopper and Brown, unpublished); • relictual, with no obvious close relatives and therefore likely to have been on granites for a long period of time (e.g. Spiculaea ciliata, a monotypic genus); • derived by speciation from allopatric congeners of habitats other than granite (e.g. C. granitora, from coastal granites east of Albany, sister species to C. infimdibularis of western high rainfall forests and coastal heaths; C. hoffmanii subsp graruticola, of east central wheatbelt outcrops, sister to C. hoffmanii subsp hoffmanii of lateritic loams well to the north¬ west in the Northampton region); and • derived by speciation from allopatric congeners of other granite outcrops (e.g. C. exstans, from outcrops east of Esperance, sister to C. integra of western wheatbelt outcrops). Thus, the orchid data do indeed support the hypothesis that conditions on south-west granite outcrops have facilitated genetic divergence and speciation. Genetic studies of a few other granite taxa lend further support, e.g. the herb Isotoma petraea (Bussell & James 1997), and eucalypts endemic to granite (Hopper & Burgman 1983; Sampson et al. 1988). Clearly, more work along these lines is needed. There is a large number of granite endemics in south¬ western Australia, especially among the perennials that dominate the woody vegetation and herbfields. We have already shown that 16% of orchids on outcrops are endemic. For eucalypts, around 24% are endemic (Hopper, unpublished). The level of endemism for the whole granite outcrop flora is difficult to determine without more penetrating research, but there is no doubt that south-western Australia has higher levels than any other system documented (e.g. Walters & Wyatt 1982; Porembski et al. 1995, 1997). The refugial opportunities offered by south-west granite outcrops are also evident in species that have highly disjunct outliers well removed from the main geographical distribution. These include populations on wet outcrop sites in much lower rainfall areas than the main species7 stand (e.g. the Jilakin Rock stand of jarrah Eucalyptus marginata, the Twine Rock stand of E. wandoo, and the Kuendar stand of E. rudis). Conversely, arid-adapted species penetrate high rainfall areas on dry north-facing slopes of granites (e.g. populations of Eucalyptus drummondii west of Margaret River). Conservation Granite outcrops occupy a very small proportion of most Western Australian landscapes in which they occur. Especially in the south-west, the outcrop plant communities are, therefore, by definition, rare, and likely to contain rare species. Hopper et al. (1990) found that endangered granite outcrop plants numbered 29 (12.2%) of the 238 plants declared in 1989 as Rare Flora under the Wildlife Conservation Act. These endangered plants ranged from large mallees (e.g. Eucalyptus crucis subsp crucis) and small trees (Acacia denticulosa, Banksia verticillata), through compact shrubs (e.g. Drummondita hasselii var longifolia, Verticordia staminosa ) and climbers (Kennedia beckxiana, K. macrophylla) to diminutive herbs (Tribonanthes purpurea) and annual aquatics ( Myriophyllum petraeum). While some endangered taxa have several populations spread across a number of disjunct outcrops, some are confined to very few localities (e.g. Myriophyllum lapidicola , known from just two gnammas). Accidental destruction of such populations could be catastrophic. Conservation in the wild in such cases needs to be backed up by off-site activites such as germplasm storage and artificial propagation (underway at Kings Park and Botanic Garden for M. lapidicola and other critically endangered granite endemics; Dixon 1994). Apart from endangered species, the diverse communities on south-western granite outcrops are noteworthy in the rapidity with which they change within and between rocks. They are complex, ever- changing, and rare in their own right. While many granite outcrops have been spared direct clearing due to their unsuitability for agriculture, and some are included within conservation reserves, most face threatening processes that need management if the native biota is to persist. Such processes include replacement or damage by invasive weeds, feral animals, grazing, inappropriate fire regimes, clearing, loss of shrub layer, salinity and dieback disease. We have briefly addressed the issue of weeds above, and highlighted the importance of maintaining undisturbed soil and dense shrub layers to control weeds effectively. Such restoration activities require an ongoing presence and commitment to a given outcrop. Local communities are vital in this context. Western Australians are fortunate in being custodians of a unique and diverse suite of granite outcrop plants. We hope that this brief review will stimulate others to study and conserve what is a remarkable heritage. Acknowledgements: To all those colleagues who helped with field work, discussion and ideas, our sincere thanks. L Sage assisted with the analysis of granite orchid biogeography. R Wyatt and R Omduff hosted SH on sabbatical in the USA in 1990, and greatly facilitated development of an international perspective on granite outcrop plants. S Porembski has recently broadened this international perspective. We are grateful for their interest and ideas. References Abbott 1 1980 The transition from mainland to island, illustrated by the flora and landbird fauna of headlands, peninsulas and islands near Albany, Western Australia. Journal of the Royal Society of Western Australia 63:79-92. 148 Journal of the Royal Society of Western Australia, 80(3), September 1997 Abbott I & Watson JR 1978 The soils, flora, vegetation and vertebrate fauna of Chatham Island, Western Australia. Journal of the Royal Society of Western Australia 60:65-70. Beard JS 1980 A new phytogeographic map of Western Australia. Western Australian Herbarium Research Notes 3:37-58. Beard JS 1981 Vegetation of Western Australia. 1:1 000 000 Series. Swan. University of Western Australia, Perth. Beard JS 1990 Plant Life of Western Australia. Kangaroo Press, Sydney. Bindon P 1997 Aboriginal people and granite domes. Journal of the Royal Society of Western Australia 80:173-179. Brooker MG & Margules CR 1996 The relative conservation value of remnant patches of native vegetation in the wheatbelt of Western Australia: I. Plant diversity. Pacific Conservation Biology 2:268-278. Brown R 1810 Prodromus Florae Novae Hollandiae et Insulae Van-Diemen exhibens characteres plantarum quas annis 1802-1805. Taylor, London. Burgman MA 1987 An analysis of the distribution of plants on granite outcrops in southern Western Australia using Mantel tests. Vegetatio 71:79-86. Bussell JD & James SH 1997 Rocks as museums of evolutionary processes. Journal of the Royal Society of Western Australia 80:221-229. Diels L 1906 Die Pflanzenwelt von West-Australien sudlich des Wendekreises. Englemann, Leipzig. Dixon KW 1994 Towards integrated conservation of Australian endangered plants - the Western Australian model. Biodiversity and Conservation 3:148-159. Erickson R 1969 The Drummonds of Hawthornden. Lamb Paterson, Perth. Erickson R, George AS, Marchant NG & Morcombe MK 1973 Flowers and Plants of Western Australia. Reed Books, Chatswood, NSW. Gaff DF 1981 The biology of resurrection plants. In: The Biology of Australian Plants (eds JS Pate & AJ McComb). The University of Western Australia Press, Nedlands, 114- 146. Grant V 1981 Plant Speciation. Colombia University Press, New York. Green JW 1985 Census of the Vascular Plants of Western Australia. Western Australian Herbarium, Department of Agriculture, Perth. Hopper SD 1979 Biogeographical aspects of speciation in the south west Australian flora. Annual Review of Ecology and Systematics 10:399-422. Hopper SD 1981 Honeyeaters and their winter food plants on granite rocks in the central wheatbelt of Western Australia. Australian Wildlife Research 8:187-97. Hopper SD 1992 Patterns of diversity at the population and species levels in south-west Australian mediterranean ecosystems. In: Biodiversity of Mediterranean Ecosystems in Australia (ed RJ Hobbs). Surrey Beatty & Sons, Sydney, 27- 46. Hopper SD 1995 Evolutionary networks: natural hybridization and its conservation significance. In: Nature Conservation 4: the Role of Networks (eds DA Saunders, JL Craig & EM Mattiske). Surrey Beatty & Sons, Chipping Norton, NSW, 51- 66. Hopper SD 1997 An Australian perspective on plant conservation biology in practice. In: Conservation Biology for the Coming Decade (eds PL Fiedler & P Kareiva). Chapman & Hall, New York. In press. Hopper SD & Burgman MA 1983 Cladistic and phenetic analyses of phylogenetic relationships among populations of Eucalyptus caesia. Australian Journal of Botany 31:161-172. Hopper SD, Harvey MS, Chappill JA, Main AR & Main BY 1996 The Western Australian biota as Gondwanan Heritage - a review. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (eds SD Hopper, JA Chappill, MS Harvev & AS George). Surrey Beatty & Sons, Chipping Norton, NSW, 1-46. Hopper SD, van Leeuwen S, Brown AP & Patrick SJ 1990 Western Australia's Endangered Flora. Department of Conservation and Land Management, Perth. Houle G 1990 Species-area relationship during primary succession in granite outcrop plant communities. American Journal of Botany 77:1433-1439. James SH 1965 Complex hybridity in Isotoma petraea I. The occurrence of interchange heterozygosity, autogamy and a balanced lethal system. Heredity 20:341-53. Keighery GJ 1995 How many weeds? In: Invasive Weeds and Regenerating Ecosystems in Western Australia (ed G Burke). Institute of Science & Technology Policy, Murdoch University, Perth, 8-12. Maiden JH 1909 Records of Western Australian Botanists. Journal of the Western Australian Natural History Society 2:5-26. Main BY 1967 Between Wodjil and Tor. Jacaranda Press, Brisbane, and Landfall Press, Perth. Main BY 1997 Granite outcrops: A collective ecosystem. Journal of the Royal Society of Western Australia 80:113-122. Marchant NG 1973 Species diversity in the south-western flora. Journal of the Royal Society of Western Australia 56:23-30. Mares MA 1997 The geobiological interface: Granitic outcrops as a selective force in mammalian evolution. Journal of the Royal Society of Western Australia 80:131-139. Moran GF & Hopper SD 1983 Genetic diversity and the insular population structure of the rare granite rock species, Eucalyptus caesia Benth. Australian Journal of Botany 31:161- 172. Myers JS 1997 Geology of granites. Journal of the Royal Society of Western Australia 80:87-100. Newbey KR & Hnatiuk RJ 1985 Vegetation and flora. In: The Biological Survey of the Eastern Goldfields. Part 3. Jackson- Kalgoorlie Study Area. Records of the Western Australian Museum Supplement 23:11-38. Newbey KR, Keighery GJ & Hall NJ 1995 Vegetation and flora. In: The Biological Survey of the Eastern Goldfields. Part 11. Boorabbin-Southem Cross Study Area. Records of the Western Australian Museum Supplement 49:17-30. Ohlemiiller R 1997 Biodiversity patterns of plant communities in shallow depressions on Western Australian granite outcrops (inselbergs). Diplomarbeit zur Erlangung des Grades eines Diplom-Biologen, University of Bonn. Ornduff R 1986 Comparative fecundity and population composition of heterostylous and non-heterostylous species of Villarsia (Menyanthaceae) in Western Australia. American Journal of Botany 73:282-286. Ornduff R 1987 Islands on islands: plant life on the granite outcrops of Western Australia. Harold L. Lyon Arboretum Lecture Number Fifteen. Universitv of Hawaii Press, Honolulu. Ornduff R 1996 An unusual floral monomorphism in Villarsia (Menyanthaceae) and its proposed origin from distvly. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (eds SD Hopper, JA Chappill, MS Harvey & AS George). Surrey Beatty & Sons, Chipping Norton, NSW, 212-222. Pate J.S & Dixon KW 1982 Tuberous, Cormous and Bulbous Plants. The University of Western Australia Press, Nedlands. Piggott PJ & Sage LW 1997 Remnant fegetation, priority flora and weed invasions at Yilliminning Rock. Journal of the Royal Society of Western Australia 80:201-208. Pignatti E & Pignatti S 1994 Centrolepidi-Hydrocotyletea alatae, a new class of ephemeral communities in Western Australia. Journal of Vegetation Science 5:55-62. 149 Journal of the Royal Society of Western Australia, 80(3), September 1997 Porembski S, Brown G & Barthlott W 1995 An inverted latitudinal gradient of plant diversity in shallow depressions on Ivorian inselbergs. Vegetatio 117:151-163. Porembski S, Seine R & Barthlott W 1997 Inselberg vegetation and the biodiversity of granite outcrops. Journal of the Royal Society of Western Australia 80:193-199. Sampson JF, Hopper SD & James SH 1988 Genetic diversity and the conservation of Eucalyptus crucis Maiden. Australian Journal of Botany 36:447-460. Schweinfurth U 1978 Beobachtungen an exponierten Standorten der sudwestaustralischen Kuste. Bot. Jahrb. Syst. 99:168-187. Smith GG 1962 The flora of granite rocks of the Porongurup Range, South Western Australia. Journal of the Royal Society of Western Australia 45:18-23. Stace HM 1995 Protogyny, self-incompatibility and pollination in Anthocercis gracilis (Solanaceae). Australian Journal of Botany 43:451-459. Stoneburner A, Wyatt R, Catcheside DG & Stone IG 1993 Census of the mosses of Western Australia. The Bryologist 96:86-101. Stoneburner A & Wyatt R 1996 Ecological and phytogeographical characteristics of the mosses of Western Australia. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (eds SD Hopper, JA Chappill, MS Harvey & AS George). Surrey Beatty & Sons, Chipping Norton, NSW, 109-119. Walters TW & Wyatt R 1982 The vascular flora of granite outcrops in the Central Mineral Region of Texas. Bulletin of the Torrey Botanical Club 109:344-364. Wardell Johnson G & Williams M 1996 A floristic survey of the Tingle Mosaic, south-w'estern Australia: applications in land use planning and management. Journal of the Royal Society of Western Australia 79:249-276. Wyatt R 1997 Reproductive ecology of granite outcrop plants from the south-eastern United States. Journal of the Royal Society of Western Australia 80:123-129. Appendix 1 Alphabetical list of Western Australian plants (vascular and cryptogams combined) represented by specimens in the Western Australian Herbarium and for which habitat descriptions on labels included the words "granite outcrop". List is current to 28 May 1996, from the WAHERB database of 413 000 specimens collected from all habitats throughout W A. Nomenclature follows Green (1985) and subsequent amendments included in the Western Australian Herbarium's WACENSUS database. Taxa are listed alphabetically by generic, specific and infraspecific names, with entry concluded by family name. Numbers are those assigned to vascular plant families as in Green (1985). Taxa who's family names are preceded by a letter rather than a number are cryptogams as follows: B - moss (bryophyte); H - liverwort (hepatic); L - lichen; F - macrofungus. Uncertainty as to taxonomic delimitation is given by off, sp , ?, " unsorted ", sens, lat., or phrase name (e.g. Form 'n', or "walyahmnningensis" . Well-delimited but unpublished taxa are identified by ms (for manuscript name). Full authorship of all names, published and unpublished, is given in Green (1985) and subsequent amendments included in WACENSUS. Acacia aciphylla 163 Mimosaceae Acacia acuaria 163 Mimosaceae Acacia acuminata subsp acuminata ms 163 Mimosaceae Acacia acutata 163 Mimosaceae Acacia aff cyperophylla 163 Mimosaceae Acacia aff rhodophloia 163 Mimosaceae Acacia andrewsii 163 Mimosaceae Acacia aneura var aneura 163 Mimosaceae Acacia aphylla 163 Mimosaceae Acacia applanata 163 Mimosaceae Acacia assimihs subsp assimilis 163 Mimosaceae Acacia ayersiana var latifolia 163 Mimosaceae Acacia browtuana var browniatia (typical variant) 163 Mimosaceae Acacia broumiana var obscuta 163 Mimosaceae Acacia cochlear is 163 Mimosaceae Acacia conniana 163 Mtmosaceae Acacia constablci 163 Mimosaceae Acacia coolgardiensis subsp coolgardiensis 163 Mimosaceae Acacia coolgardiensis subsp effusa (Pedunculate variant) 163 Mimosaceae Acacia costimana 163 Mimosaceae Acacia cotoaniana 163 Mimosaceae Acacia cracentis ms 163 Mimosaceae Acacia cremilata ms 163 Mimosaceae Acacia cuneifolia ms 163 Mimosaceae Acacia cyclops 163 Mimosaceae Acacia denticulosa 163 Mimosaceae Acacia drurnmondii subsp elegans Porongurup variant (RJ Cumm ms) 163 Mimosaceae Acacia ephedroidcs 163 Mimosaceae Acacia encifolia 163 Mimosaceae Acacia extensa 163 Mimosaceae Acacia fauntlervui 163 Mimosaceae Acacia flavipila var flavipila 163 Mimosaceae Acacia graniticola ms 163 Mimosaceae Acacia hast u lata 163 Mimosaceae Acacia hetmteles 163 Mimosaceae Acacia heterochta subsp hetcroclita ms 163 Mimosaceae Acacia heteroclita subsp valida ms 163 Mimosaceae Acacia huegelii 163 Mimosaceae Acacia jennerae 163 Mimosaceae Acacia jibberdingensis 163 Mimosaceae Acacia lasiocalyx 163 Mimosaceae Acacia leptopetala 163 Mimosaceae Acacia leptostachya 163 Mimosaceae Acacia tigulata 163 Mimosaceae Acacia Itnophylla 163 Mimosaceae Acacia littorea 163 Mimosaceae Acacia lystphloia 163 Mimosaceae Acacia merinthophora 163 Mimosaceae Acacia merrallii 16.3 Mimosaceae Acacia microbotrya 163 Mimosaceae Acacia mnnticola 163 Mimosaceae Acacia niultisiliqua 163 Mimosaceae Acacia myrti/olia 163 Mimosaceae Acacia olgana 163 Mimosaceae Acacia pram n 163 Mimosaceae Acacia pravijoha 163 Mimosaceae Acacia pulchella subsp pulchella 163 Mimosaceae Acacia pulchella var goadbyi 163 Mimosaceae Acacia pundiculala ms 163 Mimosaceae Acacia ramulosa 163 Mimosaceae Acacia redolent 163 Mimosaceae Acacia restiacea 163 Mimosaceae Acacia rigens 163 Mimosaceae Acacia sahgna 163 Mimosaceae Acacia scleroclada 163 Mimosaceae Acacia sessilispica 163 Mimosaceae Acacia sp 163 Mimosaceae Acacia sp P58 (BR Maslin 3625) 163 Mimosaceae Acacia sp P94 (JS Beard 6178) 163 Mimosaceae Acacia stenoptera 163 Mimosaceae Acacia sloioardii 163 Mimosaceae Acacia subcaerulea 163 Mimosaceae Acacia subflexuosa subsp subflexuosa 163 Mimosaceae Acacia tarculensis 163 Mimosaceae Acacia tetragonophylla 163 Mimosaceae Acacia trachycarpa 163 Mimosaceae 150 Journal of the Royal Society of Western Australia, 80(3), September 1997 Acacia tratmaniana 163 Mimosaceae Acacia trigonophylla 163 Mimosaceae Acacia triptycha 163 Mimosaceae Acacia urophylla 163 Mimosaceae Acacia verricula 163 Mimosaceae Acacia viscifolia 163 Mimosaceae Acacia wifldemu'iana 163 Mimosaceae Acacia ? glaucisSima 163 Mimosaceae Acacia ? quadrimargmea 163 Mimosaceae Acacia Sect Pluri (microneurae, terete) 163 Mimosaceae Acaraspora cilrina L Acarosporaceae Acaulon integrifalium B Pottiaceae Adinobole oldfieldiaua 345 Asteraceae Actinotus leucocephalu s 281 Apiaceae Adenantho s obovatu s 090 Proteaceae Agaricus sp F Agaricaceae Agon is hypericifolia 273 Mvrtaceae Agonis linearifolia 273 Mvrtaceae Agonis marginala 273 Mvrtaceae Agonis parviceps 273 Mvrtaceae Agrastocrinum scab rum 054F Anthericaceae Air a cupaniana 031 Poaceae Allocasuarina campestm 070 Casuarinaceae Aflocasuarma decussata 070 Casuarinaceae Allocasuarina eriochlamys subsp grossa 070 Casuarinaceae Allocasuarina huegeliana 070 Casuarinaceae Allocasuarina humilis 070 Casuarinaceae Allocasuarina thuyoides 070 Casuarinaceae Allocasuarina trichodon 070 Casuarinaceae Alyogyne hakeifoha 221 Malvaceae Alyogyne huegelii var huegelii ms 221 Malvaceae Alyogyne huegelii var wrayae ms 221 Malvaceae Alyogyne pinoniana 221 Malvaceae Amanita sp F Amanitaceae Amaranthus mitchdhi 106 Amaranthaceae Arnaranthus pallidiftorus 106 Amaranthaceae Amaranthus sp (Silent Grove) RJC 6532 106 Amaranthaceae Amperea simulant 185 Euphorbiaceae Amphipogon carictnus 031 Poaceae Amyema gibberula var tatci 097 Loranthaceae Anagallis amends 293 Primulaceae Anarthria prolifera 039 Restionaceae Andersoma aff auriculata 288 lipacndaceae Andersonia aff lehnmmann 288 Epacndaceae Andersonia anstata 288 Epacndaceae Andersonia lehmanniana 288 Epacridaceae Andersonia parvifolia 288 Epacndaceae Andersonia sp 288 Epacridaceae Andersonia sprengelioides 288 Epacndaceae Angianthus tomentosus 345 Asteraceae Anthocerci s anisantha subsp anisantha 315 Solanaceae Anthocercis genistoides 315 Solanaceae Anthocercis gracilis 315 Solanaceae Anthocercis littorea 315 Solanaceae Anthocercis viscosa subsp caudata 315 Solanaceae Anthocercis viscosa subsp niscosa 315 Solanaceae Anthurus sp F Clathraceae Aphelta brizula 040 Centrolepidaceae Aphelia cyperoides 040 Centrolepidaceae Aphelia nutans 040 Centrolepidaceae Aristida contorta 031 Poaceae Arthropodium curoipes 054F Anthericaceae Arthropodium dyeri 054F Anthericaceae Asplenium aethiopicum 01 IE Aspleniaceae Asplenium flabellifolium 011E Aspleniaceae Astartea ambigua 273 Mvrtaceae Astartea fascicularis 273 Myrtaceae Astartea sp Mt Johnston (AR Annels 5645) 273 Myrtaceae Asterella drummondu H Hepaticae Asterolasia pallida subsp "unsorted " 175 Rutaceae Astroloma ciliatum 288 Epacridaceae Astrolorna epacridis 288 Epacndaceae Astroloma pallidum 288 Epacridaceae Astroloma prostratum 288 Epacridaceae Astroloma serratifolium 288 Epacridaceae Astroloma sp 288 Epacridaceae Baeckea " walyahmoningensis " 273 Myrtaceae Baeckea behrii 273 Myrtaceae Baeckea camphorosmae 273 Myrtaceae Baeckea crispiflora 273 Myrtaceae Baeckea crispiflora ms 273 Myrtaceae Baeckea maidenu 273 Myrtaceae Baeckea margarethae 273 Myrtaceae Baeckea margarethae ms 273 Myrtaceae Baeckea polyandra 273 Myrtaceae Baeckea recurva ms 273 Myrtaceae Baeckea sp Payne s Find (S Patrick 1095) 273 Myrtaceae Baeckea tenuiramea 273 Myrtaceae Baeckea tetragona 273 Mvrtaceae Banksia seminuda subsp remanam 090 Proteaceae Banksia verticil lata 090 Proteaceae Barbula calycina B Pottiaceae Barbula crinata B Pottiaceae Bartsia trixago 316 Scrophulariaccae Beaufortia empetrifoha 273 Mvrtaceae Beaufortia incana 273 Myrtaceae Beaufortia sdmieri 273 Myrtaceae Blennospora drummondii 345 Asteraceae Boron ia albiflora 175 Rutaceae Boronia algida 175 Rutaceae Boronia busselliana 175 Rutaceae Boronia coerulescens subsp "unsorted" 175 Rutaceae Boronia crassifolia 175 Rutaceae Boronia crenulata var "unsorted" 175 Rutaceae Boronia crenulata var crenulata 175 Rutaceae Boronia defohata 175 Rutaceae Boronia gracilipes 175 Rutaceae Boronia inornnta subsp mornata 175 Rutaceae Boronia sp 175 Rutaceae Boronia subsessilis 175 Rutaceae Boronia tetrandra 175 Rutaceae Borya constricta 054F Anthericaceae Borya constricta F Anthericaceae Borya laciniata 054F Anthericaceae Borya longiscapa 054F Anthericaceae Borya mtida 054F Anthericaceae Borya scirpoidea 054F Anthericaceae Borya sphaerocephala 054F Anthericaceae Borya sphaerocephala sens lat 054F Anthericaceae Bossiaea dentata 165 Papilionaceae Bossiaea eriocarpa 165 Papilionaceae Bossiaea linophyUa 165 Papilionaceae Brachychiton grandifloms 223 Sterculiaceae Brachychiton gregorti 223 Sterculiaceae Brachymenium preissianum B Brvaceae Brachyscome ciliaris 345 Asteraceae Brachyscome iberidifolia 345 Asteraceae Brachyscome perpusilla 345 Asteraceae Brachyscome sp 345 Asteraceae Breutelia affinis B Bartramiaceae Briza maxima 031 Poaceae Briza minor 031 Poaceae Bromus rubens 031 Poaceae Bruchia brevipes B Dicranaceae Bryum argenteum B Bryaceae Bryum billardieri B Bryaceae Bryum caespiticium B Bryaceae Bryum campylothecium B Bryaceae Bryum capillare B Bryaceae Bryum cheelii B Bryaceae Bryum chrysoneuron B Bryaceae Bryum dichotomum B Bryaceae Bryum pachytheca B Bryaceae Bryum sp B Catcheside B Bryaceae Bryum torquescens B Bryaceae Bulbine semibarbata 054G Asphodelaceae Burdmdia multiflora 054J Colchicaceae Bursaria occidentals 152 Pittosporaceae Caesia micrantha 054F Anthericaceae Caladenia atf ingens subsp att ingens ms 066 Orchidaceae Caladenia attmgens subsp gracillima ms 066 Orchidaceae Caladenia brmsura ms 066 Orchidaceae Caladenia broumii ms 066 Orchidaceae Caladenia caesarca subsp maritnna ms 066 Orchidaceae Caladenia caesarca subsp transiens ms 066 Orchidaceae Caladenia citrina ms 066 Orchidaceae Caladenia denticulata 066 Orchidaceae Caladenia dilatata dilatata 066 Orchidaceae Caladenia dimidia ms 066 Orchidaceae Caladenia discoidea 066 Orchidaceae 151 Journal of the Royal Society of Western Australia, 80(3), September 1997 Caladenia doutchiae 066 Orchidaceae Caladenia exstans ms 066 Orchidaceae Caladenia falcata 066 Orchidaceae Caladenia filifera 066 Orchidaceae Caladenia flaccida subsp flaccida ms 066 Orchidaceae Caladenia flaccida subsp pulchra ms 066 Orchidaceae Caladenia Jlava subsp flaw ms 066 Orchidaceae Caladenia flaw subsp maculata ms 066 Orchidaceae Caladenia Jlava subsp sylvestris ms 066 Orchidaceae Caladenia footeana ms 066 Orchidaceae Caladenia granitora ms 066 Orchidaceae Caladenia heberleana ms 066 Orchidaceae Caladenia hirtd subsp mea ms 066 Orchidaceae Caladenia hoffmami subsp graniticola ms 066 Orchidaceae Caladenia incensa ms 066 Orchidaceae Caladenia incrassata ms 066 Orchidaceae Caladenia infundibularis 066 Orchidaceae Caladenia Integra 066 Orchidaceae Caladenia latifolia 066 Orchidaceae Caladenia lobata 066 Orchidaceae Caladenia longicauda subsp clivicola ms 066 Orchidaceae Caladenia longicauda subsp eminens ms 066 Orchidaceae Caladenia longicauda subsp longicauda ms 066 Orchidaceae Caladenia longicauda subsp rigid ula ms 066 Orchidaceae Caladenia longiclavata 066 Orchidaceae Caladenia macmtylis 066 Orchidaceae Caladenia marginata 066 Orchidaceae Caladenia microchila ms 066 Orchidaceae Caladenia multiclavia 066 Orchidaceae Caladenia nivalis ms 066 Orchidaceae Caladenia pachychila ms 066 Orchidaceae Caladenia pendens ms 066 Orchidaceae Caladenia pholcoidea ms 066 Orchidaceae Caladenia polychroma ms 066 Orchidaceae Caladenia radial is 066 Orchidaceae Caladenia reptans subsp impensa ms 066 Orchidaceae Caladenia rqitans subsp reptans ms 066 Orchidaceae Caladenia rhomboidifomis 066 Orchidaceae Caladenia roei 066 Orchidaceae Caladenia saccharata 066 Orchidaceae Caladenia serotina ms 066 Orchidaceae Caladenia sigmoidea 066 Orchidaceae Caladenia splendens ms 066 Orchidaceae Caladenia hiemalis ms 066 Orchidaceae Caladenia horistes ms 066 Orchidaceae Caladenia pendens subsp pendens ms 066 Orchidaceae Caladenia remota subsp remota ms 066 Orchidaceae Caladenia pendens subsp talbotii ms 066 Orchidaceae Caladenia voigtii ms 066 Orchidaceae Calandnma calyptrata 111 Portulacaceae Calandrinia ptychosperma 111 Portulacaceae Calectasia grandiflora 054C Dasypogonaceae Callilriche stagnalis 186 Callitrichaceae Callitris glaucophylla 018 Cupressaceae Calhtris preissn subsp "unsorted" 018 Cupressaceae Calocephalus viultiflorus 345 Asteraceae Caloplaca sp l Teloschistaceae Calothamnus quadrifldus var "unsorted" 273 Myrtaceae Calothnrnnus quadrifidus var Esperance District (AE Orchard 1676) 273 Myrtaceae Calothamnus rupestris 273 Myrtaceae Crt/of/jrtHimis schaueri 273 Myrtaceae Gi/of/wtfwws sp 273 Myrtaceae Calothamnus tuberosus 273 Myrtaceae Calothamnus villosus 273 Myrtaceae Calycopeplus paucifolius 185 Euphorbiaceae Calytrix acutifoha 273 Myrtaceae Calytrix angulata 273 Myrtaceae Calytnx decandra 273 Myrtaceae Calytrix depressa 273 Myrtaceae Calytnx fraseri 273 Myrtaceae Calytrix glutmosa 273 Myrtaceae Calytrix leschenaultu 273 Myrtaceae Calytnx tetragona 273 Myrtaceae Campylopus australis B Dicranaceae Campylopus bicolor B Dicranaceae Campylopus flindersii B Dicranaceae Campylopus mtroflexus B Dicranaceae Carpobrotus aequilaterus 110 Aizoaceae Cassytha glabella forma dispar 131 Lauraceae Cassytha racemosa 131 Lauraceae Centaurium erythraea 303 Gentianaceae Centaurium sp 303 Gentianaceae Centranthus ruber 334 Valenanaceae Centrolepis alepyroides 040 Centrolepidaceae Centrolepis aristata 040 Centrolepidaceae Centrolepis drumnondiana 040 Centrolepidaceae Centrolepis eremica 040 Centrolepidaceae Centrolepis glabra 040 Centrolepidaceae Centrolepis pilosa 040 Centrolepidaceae Centrolepis polygyna 040 Centrolepidaceae Centrolepis stngosa subsp rupestris 040 Centrolepidaceae Centrolepis stngosa subsp rupestris ms 040.Centrolepidaceae Centrolepis strigosa subsp stngosa 040 Centrolepidaceae Cephaloziella arctica subsp subantarctica H Cephaloziellaceae Chamaesnlla corymbosa var corymbosa 054F Anthericaceae Chamaescilla corymbosa var latifolia 054F Anthericaceae Chamelaucium cilialum 273 Myrtaceae Chamelaucium flortferum ms 273 Myrtaceae Chamelaucium floriferum subsp diffusion ms 273 Myrtaceae Chamelaucium floriferum subsp floriferum ms 273 Myrtaceae Chamelaucium forrestii 273 Myrtaceae Chamelaucium forrestii subsp forrestii ms 273 Myrtaceae Chamelaucium sp Yalgoo (Y Chadwick 1816) ms 273 Myrtaceae Chamonitia sp f Hymenogastraceae Cheilanthes aff austrotenuifoUa 007 Adiantaceae Cheilanthes austrotenuifoUa 007 Adiantaceae Cheilanthes distorts 007 Adiantaceae Cheilanthes lasiophylla 007 Adiantaceae Cheilanthes sieberi subsp sieberi 007 Adiantaceae Cheiranthera fllifolia var fllifolia 152 Pittosporaceae Chiloscyphus semiteres H Geocalycaceae Chloris truncata 031 Poaceae Chorizandra enodis 032 Cyperaceae Chorizema aciculare subsp aciculare 165 Papilionaceae Chorizema cordatum 165 Papilionaceae Chorizema dicksonii 165 Papilionaceae Chorizema nervosum 165 Papilionaceae Chorizema retrorsum 165 Papilionaceae Chrysocephalum apiculatum 345 Asteraceae Chrysocephalum ramosissimum 345 Asteraceae Chrysocephalum semicalvum 345 Asteraceae Chthonocephalus pseudevax 345 Asteraceae Cladia aggregata L Cladiaceae Cladia ferditiandii l. Cladiaceae Cladia sullivanii L Cladiaceae Cladonia cajntellata L Cladoniaceae Clematicissus angustissima 216 Vitaceae Clematis pubescens 119 Ranunculaceae Ckmit' tetrandra var tetrandra 137A Capparaceae Clconte une fera 137A Capparaceae Coccocarpia erythroxili l Coccocarpiaceac Cochlospermum gilhvraei 241 Cochlospermaceae Comesperma calymega 183 Polygalaceae Cqmespenm confertum 183 Polygalaceae Comesperma integer rimunt 183 Polygalaceae Comesperma spinosum 183 Polygalaceae Comesperma volubile 183 Polygalaceae Commersonia crispa 223 Sterculiaceae Conospermum caeruleum 090 Proteaceae Conostylis aculeata subsp aculeata 055 Hacmodoraceae Canostylis bealiuM 055 Haemodoraceae Conostulis prolifera 055 Haemodoraceae Conostylis setigera 055 Haemodoraceae Canostylis setigera subsp setigera 055 Haemodoraceae Conostylis set os a 055 Haemodoraceae Corchorus etachocarpus 220 Tiliaceae Corybas recumis 066 Orchidaceae Cot ula coronopifnlia 345 Asteraceae Cofula cot abides 345 Asteraceae Craspedia var tab ills 345 Asteraceae Crassula calorata var acuminata 149 Crassulaceae Crassula decumbens vardecuntbens 149 Crassulaceae Crassula exserta 149 Crassulaceae Crassula natans var minus 149 Crassulaceae Crassula pedicellosa 149 Crassulaceae Crassula peduncular is 149 Crassulaceae Crassula siebmana subsp tetramera 149 Crassulaceae Crassula sp 149 Crassulaceae Crossidium davidai B Pottiaceae Crotalaria sp 165 Papilionaceae 152 Journal of the Royal Society of Western Australia, 80(3), September 1997 Cryptandra arbutiflora var arbutiflora 215 Rhamnaceae Cryptandra congesta 215 Rhamnaceae Cryptandra pungnts 215 Rhamnaceae Cryptandra yrilsonii 215 Rhamnaceae Cryptostylis ovata 066 Orchidaceae Cuscuta epithymum 307A Cuscutaceae Cyanicula amplexans ms 066 Orchidaceae Cyanicula ashbyac ms 066 Orchidaceae Cyanicula caeruka subsp apertala ms 066 Orchidaceae Cyanicula deformis ms 066 Orchidaceae Cyanicula fragrant ms 06h Orchidaceae Cyanicula gemmata ms 066 Orchidaceae Cyanicula sericea ms 066 Orchidaceae Cycas pruinosa 016 Cycadaceae Cynoglossum oust rale var ''unsorted" 310 Boraginaceae Cy per ut centralis 032 Cyperaceae Cyperus cunninghamii subsp cunninghamii 032 Cyperaceae Cyperut tenellus 032 Cyperaceae Cyphanthera nncrophylla 315 Solanaceae Cyrtostylis huegelii var 066 Orchidaceae Cyrtostylis robusta renifomis 066 Orchidaceae Dampiera alata 341 Goodeniaceae Dampiera altissima 341 Goodeniaceae Dampiera fasciculata 341 Goodeniaceae Dampiera haematotricha subsp dura 341 Goodeniaceae Dampiera lavandulaceae 341 Goodeniaceae Dampiera linearis 341 Goodeniaceae Dampiera sp 341 Goodeniaceae Dampiera stenostachya 341 Goodeniaceae Danthonia caespitosa 031 Poaceae Danthonia sp 031 Poaceae Darwima citriodora 273 Myrtaceae Daminia diosmoides 273 Myrtaceae Danoinia thymoides 273 Myrtaceae Darunnia vestita 273 Myrtaceae Daucus glochidiatus 281 Apiaceae Daviesia benthamii subsp acanlhoclona ms 165 Papilionaceae Davietia crOniniana 165 Papilionaceae Daviesia decurrens 165 Papilionaceae Daviesia horrida 165 Papilionaceae Daviesia incrassata subsp reversifolia 165 Papilionaceae Davietia inflata 165 Papilionaceae Desmatodon canvolutus B Pottiaceae Desmatodon recurvatus B Pottiaceae Desmatodon sp nov.X B Pottiaceae Desmocladus aspera ms 039 Restionaceae Dichondra repens 307 Convolvulaceae Dichopogon capilhpes 054F Anthencaceae Dicranoloma diaphanoneururit B Dicranaceae Didymodon torquatus B Pottiaceae Digitaria brownu 037 Poaceae Dillwynia sp A Perth Flora (R Covenv 8036) 165 Papilionaceae Dioscorea hastifolia 059 Dioscoreaceae Diplocyclos palmntus 337 Cucurbitaceae Diplolaena andreivsii 175 Rutaceae Diplolaena gramticola ms 175 Rutaceae Diplolaena microcephala 175 Rutaceae Diplolaena velutina ms 175 Rutaceae Diploschistes sp L Thelotremataceae Diuris aff longifolia 066 Orchidaceae Diuris brumalis 066 Orchidaceae Diuris conspicillata 066 Orchidaceae Diuris laevis 066 Orchidaceae Diuris laxiflora 066 Orchidaceae Diuris longifolia 066 Orchidaceae Diuris maculata 066 Orchidaceae Diuris picta 066 Orchidaceae Diuris pulchella 066 Orchidaceae Diuris recurva 066 Orchidaceae Diuris setacea 066 Orchidaceae Dodonaea bursarufoliu 207 Sapmdaceae Dodonaea caespitosa 207 Sapmdaceae Dodonaea ceratocarjia 207 Sapmdaceae Dodonae-a tnaecjuifolia 207 Sapmdaceae Dodonaea lobulata 207 Sapmdaceae Dodonaea microzyga var acrolobata 207 Sapindaceae Dodonaea petwlaris 20 7 Sapmdaceae Dodonaea pinifolia 207 Sapindaceae Dodonaea ptarmicaefolia 207 Sapindaceae Dodonaea stenozyga 207 Sapindaceae Dodonaea viscosa subsp angustissima 207 Sapindaceae Dodonaea viscosa subsp mucronata 207 Sapindaceae Dodonaea viscosa subsp spatulata/angustissima 207 Sapindaceae Dodonaea viscosa subsp spatulata 207 Sapindaceae Drakonorclus barbdrossa ms 066 Orchidaceae Drakonorclus drakeoides ms 066 Orchidaceae Drakonorclus mesocera ms 066 Orchidaceae Drosera andersoniana 143 Droseraceae Drosera bulbosa subsp bulbosa 143 Droseraceae Drosera erythrogync 143 Droseraceae Drosera gigantea subsp gigantea 143 Droseraceae Drosera glandullgcra 143 Droseraceae Drosera heteropln/lla 143 Droseraceae Drosera leucoblasta 143 Droseraceae Drosera lownei 143 Droseraceae Drosera macrantha subsp macrantha 143 Droseraceae Drosera menziesii 143 Droseraceae Drosera menziesii subsp menziesii 143 Droseraceae Drosera menziesii subsp penicillaris 143 Droseraceae Drosera microphylla 143 Droseraceae Drosera pelt at a 143 Droseraceae Drosera ramellosa 143 Droseraceae Drosera sp 143 Droseraceae Drosera stolonifera 143 Droseraceae Drosera stolonifera subsp rupicola 143 Droseraceae Drosera stolonifera subsp stolonifera 143 Droseraceae Drosera subhirtella subsp moorei 143 Droseraceae Drosera subhirtella subsp subhirtella 143 Droseraceae Dryandra armata 090 Proteaceae Dryandra armata var armata 090 Proteaceae Dryandra cuneata 090 Proteaceae Dryandra mvea subsp nivea ms 090 Proteaceae Dysphania glomulifera subsp eremaea 105 Chenopodiaceae Eccremidium arcuatum B Ditrichaceae Eccremidium pulchellum B Ditrichaceae Ecdeiocolea monos tachya 039 Restionaceae Elatine gratioloides 235 Elatinaceae Elythranthera brunonis 066 Orchidaceae Elylhranthera emargmata 066 Orchidaceae Enchylaena tomentosa var toMentoSa 105 Chenopodiaceae Endocarpon sp ( crassisporum vel macrosporum) L Verrucariaceae Ephebe lanata L Lichmaceae Ephemerum cristatum B Ephcmcraceae Eremophila altenufolia 326 Myoporaceae Eremophila clarkei 326 Myoporaceae Eremophila denpiens subsp "unsorted" 326 Myoporaceae Eremophila deetpiens subsp decipiens ms 326 Myoporaceae Eremophila deserti 326 Myoporaceae Eremophila exilifolia 326 Myoporaceae Eremophila fbrrestu subsp "unsorted" 326 Myoporaceae Eremophila glabra subsp "unsorted" 326 Myoporaceae Eremophila glutmosa 326 Myoporaceae Eremophila lehmanniana 326 Myoporaceae Eremophila oldfieldii subsp angustifoha ms 326 Myoporaceae Eremophila oppositifblia subsp angustifolia ms 326 Myoporaceae Eremophila pachyphylla 326 Myoporaceae Eremophila phillipsh 326 Myoporaceae Eremophila platycalyx subsp platycalyx ms 326 Myoporaceae Eremophila s copana 326 Myoporaceae Eremophila serrulata 326 Myoporaceae Eremophila sp 326 Myoporaceae Eriachne pulchella subsp pulchella 031 Poaceae Eriochilus dilatatus subsp dilatatu s ms 066 Orchidaceae Eriochilus dilatatus subsp muHiflarus ms 066 Orchidaceae Eriochilus dilatatus subsp undulatus ms 066 Orchidaceae Eriochilus helanomos ms 066 Orchidaceae Eriochilus pulchellus ms 066 Orchidaceae Eriochilus scaber 066 Orchidaceae Eriostemon brucei subsp brucei 175 Rutaceae Eriostemon linearis 175 Rutaceae Erodium cicuiarium 167 Geraniaceae Eucalyptus acies 273 Myrtaceae Eucalyptus aff lane-poolei 273 Myrtaceae Eucalyptus angulosa 273 Myrtaceae Eucalyptus ariichnuea 273 Mvrtaceae Eucalyptus aspratihs 273 Myrtaceae Eucalyptus brachyphylla 273 Myrtaceae Eucalyptus brtnupes 273 Myrtaceae Eucalyptus caesia subsp caesia 273 Myrtaceae Eucalyptus calycogona var calycogona 273 Myrtaceae 153 Journal of the Royal Society of Western Australia, 80(3), September 1997 Eucalyptus cornuta 273 Myrtaceae Eucalyptus crucis subsp crucis 273 Myrtaceae Eucalyptus crucis subsp lanceolata 273 Myrtaceae Eucalyptus decipiens 273 Myrtaceae Eucalyptus decipiens subsp adesmophloia 273 Myrtaceae Eucalyptus decipiens subsp chalara 273 Myrtaceae Eucalyptus doratoxylott 273 Myrtaceae Eucfl/vpfus drummondii 273 Myrtaceae Eucalyptus eremophila 273 Myrtaceae Eucalyptus eremophila hybrid 273 Myrtaceae Eucalyptus flocktoniae 273 Myrtaceae Eucalyptus gratiae 273 Myrtaceae Eucalyptus incrassata 273 Myrtaceae Eucalyptus insularis 273 Myrtaceae Eucalyptus kruseana 273 Myrtaceae Eucalyptus lehmamw 273 Myrtaceae Eucalyptus loxophleba subsp lissophloia 273 Myrtaceae Eucalyptus megacarpa 273 Myrtaceae Eucalyptus rnelanoxylon 273 Myrtaceae Eucalyptus orbifolia 273 Myrtaceae Eucalyptus patens 273 Myrtaceae Eucalyptus petraea 273 Myrtaceae Eucalyfdus phaenophylhi subsp interjacens 273 Myrtaceae Eucalyptus pileata 273 Myrtaceae Eucalyptus polysciada 273 Myrtaceae Eucalyptus prava ms 273 Myrtaceae Eucalyptus salulms 273 Myrtaceae Eucalyptus sheathiana 273 Myrtaceae Eucalyptus sporadica subsp sporadica ms 273 Myrtaceae Eucalyptus suggratidis subsp alipes 273 Myrtaceae Eucalyptus suggrandis subsp biwing (PJW 588) 273 Myrtaceae Eucalyptus tephrodes ms 273 Myrtaceae Eucalyptus tetraptera 273 Myrtaceae Eucalyptus uncinata 273 Myrtaceae Eucfl/i/p/us v irginae ms 273 Myrtaceae Euphorbia coghlanii 185 Euphorbiaceae Euphorbia sp 185 Euphorbiaceae Eutaxia microphylla 165 Papilionaceae Eutaxia obovata 165 Papilionaceae Ficus platypoda var minor 087 Moraceae FilagogalUca 345 Asteraceae Fimbristyhs dichotoma (desert form) 032 Cyperaceae Fimbristylis neilsonii 032 Cyperaceae Fissidens asplenioides B Fissidcntaceae Fissidens taylorii B Fissidentaceae Fissidens vittatus B Fissidentaceae Flavoparmelia haysomii L Parmeliaceae Flavoparmelia rutidota L Parmeliaceae Fossonibronia sp H Hepaticae Funaria apophysata B Funariaceae Funaria helmsii B Funariaceae Funaria producta B Funariaceae Calenna sp F Cortinariaceae Galium aparine 331 Rubiaceae Gastrolobium brownii 165 Papilionaceae Gastrolobium callistachys 165 Papilionaceae Gastrolobium calycmum 165 Papilionaceae Gastrolobium involutum 165 Papilionaceae Gastrolobium laytunn 165 Papilionaceae Gastrolobium ovalifolium 165 Papilionaceae Gastrolobium parviflorum 165 Papilionaceae Gastrolobium sptnosum var spmosum 165 Papilionaceae Gastrolobium spinosum var trilobum 165 Papilionaceae Genoplt'sium nigricans 066 Orchidaceac Genus, sp 032 Cyperaceae Genus sp 185 Euphorbiaceae Genus sp Shannon (PC Wilson 1237B) ms 281 Apiaceae Cigaspermum repens B Gigaspermaceae Ghschrocaryon aureum var "unsorted" 276 llaloragaceae Glischrocaryon aureum var angustifolium 276 Haloragaceae Ghschrocaryon aureum var aureum 276 Haloragaceae Glischrocaryon flavescens 276 Haloragaceae Ghschrocaryon roei 276 Haloragaceae Glischrocaryon sp 276 Haloragaceae Glossosti'gma diandrum 316 Scrophulariaceae Glossostigma drummondii 316 Scrophulariaceae Glossostigma sp 316 Scrophulariaceae Glycine canescens 165 Papilionaceae Glycine clandestina 165 Papilionaceae Gnaphalium indutum 345 Asteraceae Gnephosis tenuissmia 345 Asteraceae Gompholobium kmghtianum 165 Papilionaceae Gompholobium marginatum 165 Papilionaceae Gompholobium polymorphum 165 Papilionaceae Gompholobium venustum 165 Papilionaceae Gonocarpus nodulosus 276 Haloragaceae Gonocarpus paniculfittis 276 Haloragaceae Gonocarpus scordioides 276 Haloragaceae Gonocarpus sp 276 Haloragaceae Gooden la affinis 341 Goodemaceae Goodenia caerulea 341 Goodemaceae Goodema concinna 341 Goodemaceae Goodenia decurswa 341 Goodemaceae Goodenia havilandu 341 Goodeniaceae Goodenia helmsii 341 Goodeniaceae Goodenia incana 341 Goodeniaceae Goodema micrantha 341 Goodeniaceae Goodema muelleriana 341 Goodeniaceae Goodema pulchella 341 Goodeniaceae Goodema quadrilocularis 341 Goodeniaceae Goodema scapigera 341 Goodeniaceae Goodema stenophylln 341 Goodemaceae Goodema tripartita 341 Goodeniaceae Granitites intangendus 215 Rhamnaceae Grevillea aff concinna 090 Proteaceae Grevillea aslencosa 090 Proteaceae Grevillea bipin natifida 090 Proteaceae Grevillea centristigma 090 Proteaceae Greinllea christineaa 090 Proteaceae Grevillea arsufolia 090 Proteaceae Greinllea comma subsp concinna 090 Proteaceae Grevillea concinna subsp lemannuma 090 Proteaceae Grevillea dwersifoha subsp subtersericata 090 Proteaceae Grevillea dolichopoda 090 Proteaceae Grevillea fuscolutea 090 Proteaceae Greinllea granulosa 090 Proteaceae Greinllea huegelii 090 Proteaceae Grevillea leptobotrys 090 Proteaceae Grevillea manglesii subsp manglesii 090 Proteaceae Grrvillea nann subsp nana 090 Proteaceae Grrvillea pamculata (Form 'n') 090 Proteaceae Grevillea parallela 090 Proteaceae Grevillea pauciflora subsp psilophylla 090 Proteaceae Grevillea pityophylla 090 Proteaceae Grevillea plurijuga 090 Proteaceae Grrvillea pulchella subsp pulchella ms 090 Proteaceae Grevillea rigida subsp rigida 090 Proteaceae Grevillea tetrapleura 090 Proteaceae Grevillea umbellulata subsp urnbellulata 090 Proteaceae Grimmia larvigata B Grimmiaceae Guichenutia macrantha 223 Stercuhaceae G ymnopilu! s sp F Cortinariaceae Gyrostemon subnudus 108 Gyrostemonaceae Haemodorum sparsiflorum 055 Haemodoraceae Hakea artda 090 Proteaceae Hakea clavata 090 Proteaceae Hakea erinacea 090 Proteaceae Hakea myrtoides 090 Proteaceae Hakea ublicjua 090 Proteaceae Hakea prostrata 090 Proteaceae Hakea recurva 090 Proteaceae Hakea siurveolens 090 Proteaceae Hakea undulata 090 Proteaceae Haloragodendron glandulosum 276 Haloragaceae Haloragodendnm racemvsum 276 Haloragaceae Hanmifordia bissillii 223 Sterculiaceae Hannafordm sp 223 Sterculiaceae Heliotropium tenuifalium 310 Boraginaceae Helipterum craspedioides 345 Asteraceae Helipterum heteranthum var minor 345 Asteraceae Hemiandra pungens 313 Lamia ceae Flemigema aff scncea 313 Lamiaceae Hemigema incana 313 Lamiaceae Flemigema podalynna 313 Lamiaceae Hemigema rigida 313 Lamiaceae Hemigema sericea 313 Lamiaceae Hemigema sp 313 Lamiaceae Hemigema sp Albany (GJ Keighery 8712) 313 Lamiaceae Hemigenia sp Albany (GJ Keighery 8712) ms 313 Lamiaceae Heterodea muelleri L Heterodeaceae 154 Journal of the Royal Society of Western Australia, 80(3), September 1997 Heterodermia ohscurata L Verruca riaceae Heteropogon contort us 031 Poaceae Hexagona apiaria F Pulvporaceae Hibbcrtia acerosa 226 Dilleniaceae Hibbertia aff gracilipes (glabrous carpels) 226 Dilleniaceae Hibbcrtia aff gracilipes 226 Dilleniaceae Hibbertia cunningham'ii 226 Dilleniaceae Hibbertia glomerosa 226 Dilleniaceae Hibbertia gracilipes 226 Dilleniaceae Hibbertia graruticola 226 Dilleniaceae Hibbertia hypericoides 226 Dilleniaceae Hibbertia mucronata 226 Dilleniaceae Hibbertia pungens 226 Dilleniaceae Hibbertia racemosa 226 Dilleniaceae Hibbertia rhadinopoda 226 Dilleniaceae Hibbertia rupicola 226 Dilleniaceae Hibbertia spicatn 226 Dilleniaceae Hibiscus leptocladus 221 Malvaceae Homaloscmdium homalocarpum 281 Apiaceae Hovea acanthoclada 165 Papilionaccae Hovea pungens 165 Papilionaccae Hyalochlamys globifera 345 Asteraceae Hyalosperma glutinosum subsp glutmosum 345 Asteraceae Hyalosperma pusillurn 345 Asteraceae Hybanthus floribundus subsp floribundus 243 Violaceae Hybanthus monopetalus 243 Violaceae Hydrocotyle alata 281 Apiaceae Hydrocotylc callicarpa 281 Apiaceae Hydrocotyle diantha 281 Apiaceae Hydrocotyle puberuk ms 281 Apiaceae Hydrocotyle scutellifera 281 Apiaceae Hypericum japorveum 233 Clusiaceae Hypocalymma angustifolium 273 Myrtaceae Hypocalyrnma uncinatum ms 273 Myrtaceae Hypoclmeris glabra 345 Asteraceae Hypogymnia subphysodes L Hypogymniaceae Hypoxis glabella var glabella 056A Hypoxidaceae Hypoxis occidentals var i juadriloba 056A Hypoxidaceae Indigofera australis 165 Papilionaceae Indigofera monophylla 165 Papilionaceae Ipomoea nil 307 Convolvulaceae Ischyrodon lepturus B Brachvtheciaceae Isoetes australis 004 Isoetaceae Isoetes caroli 004 Isoetaceae Isoetes inflata 004 Isoetaceae Isoetes sp 004 Isoetaceae Isolepis congrua 032 Cyperaceae Isolepis marginata 032 Cyperaceae Isopogon attenuatus 090 Proteaceae Isopogon divergent 090 Proteaceae Isopogon formosus 090 Proteaceae Isotoma hypocrateriformis 340 Lobeliaceae Isotoma petraea 340 Lobeliaceae Isotoma scapigera 340 Lobeliaceae lsotropis atropurpurea 165 Papilionaceae lsotropis cuneifoha 165 Papilionaceae lsotropis sp 165 Papilionaceae Jacksonia alata 165 Papilionaceae Jacksonia furcellata 165 Papilionaceae Jacksonia spinosa 165 Papilionaceae Jamcsoniella col ora t a H Hepaticac I uncus bufonius 052 Juncaceae Juncus capitatus 052 Juncaceae Juncus pauciflorus 052 Juncaceae Kennedia beckxiana 165 Papilionaceae Kennedia carinata 165 Papilionaceae Kennedia glabrata 165 Papilionaceae Kennedia macrophylla 165 Papilionaceae Kennedia prostrala 165 Papilionaceae Kennedia stirlingii 165 Papilionaceae Keraudrenia integrifolia 223 Sterculiaceae Kunzea aff pulchella 273 Myrtaceae Kunzea af finis 273 Myrtaceae Kunzea baxteri 273 Myrtaceae Kunzea niicromera 273 Myrtaceae Kunzea pulchella 273 Myrtaceae Labichea lanceolata subsp brevifolia 164 Caesalpiniaceae Labichea teretifolia subsp grandistipulata 164 Caesalpiniaceae Laccaria laccata F Tricholomataceae Lambertia inermis var inermis 090 Proteaceae Lampranthus glaucus 1 10 Aizoaceae Lasiopetalum compaction 223 Sterculiaceae Lasiopetalum cordifolium 223 Sterculiaceae Lawrencella rosea 345 Asteraceae Lawrencia diffusa 221 Malvaceae Laxmannia minor 054F Anthericaceae Laxmannia sp 054F Anthericaceae Laxmannia squarrosa 054F Anthericaceae Lechenaultia formosa 341 Goodeniaceae Lemooria burkittu 345 Asteraceae Lepidosperma angustatum 032 Cyperaceae Lepidosperma brunonianum 032 Cyperaceae Lepidosperma resirtosum 032 Cyperaceae Lepidosperma sp 032 Cyperaceae Lepidosperma sp A2 Island Flat (Keighery 7000) 032 Cyperaceae Lepidosperma sp K 032 Cyperaceae Lepidosperma tuberculatum 032 Cyperaceae Lepiota sp F Agancaceae Leporella fimbriata 066 Orchidaceae Leprocaulon microscopicum L fungi Imperfect i Leproloma membra naceum L ?Pannariaceae Leptocarpus aff nwi "white'' 039 Restionaceae Leptocarpus kraussii ms 039 Restionaceae Leptocarpus roycci ms 039 Restionaceae Leptoceras menziesii 066 Orchidaceae Leptodontium paradoxum B Pottiaceae Leptomena cunninghanui 092 Santalaceae Leptomeria squarrulosa 092 Santalaceae Leptosema aphyllum ms 165 Papilionaceae Leptospermum aff roet 273 Myrtaceae Leptospermum erubescens 273 Myrtaceae Leptospermum incanum 273 Myrtaceae Leptospermum macgilhvrayi 273 Myrtaceae Leptospermum roei 273 Myrtaceae Leptospermum senceum 273 Myrtaceae Leptospermum sp 273 Myrtaceae Lethocolea squamata H Hepaticac Leucopogon aff reflexus 288 Fpacridaceae Leucopogon australis 288 Epacridaceae Leucopogon bracteolans 288 Epacridaceae Leucopogon glabellus 288 F.pacridaceae Leucopogon pendulus 288 Epacridaceae Leucopogon reflexus 288 Epacridaceae Leucopogon revolutus 288 Epacridaceae Leucopogon rotundifolms 288 Epacridaceae Leucopogon sp 288 Epacridaceae Leucopogon sprengehoides 288 Epacridaceae Leucopogon strictus 288 Epacridaceae Leucopogon unilateralis 288 Epacridaceae Levenhookia dubia 343 Stylidiaceae Levenhookia Icptantha 343 Stylidiaceae Levenhookia pauciflora 343 Stylidiaceae Levenhookia pusilla 343 Stylidiaceae Levenhookia stipitata 343 Stylidiaceae Lobelia alata 340 Lobeliaceae Lobelia gibbosa 340 Lobeliaceae Lobelia heterophylla 340 Lobeliaceae Lobelia sp 340 Lobeliaceae Logan ia serpylhfolta subsp serpytlifolia 302 Loganiaceae Logania sp 302 Loganiaceae Logania stenophylla 302 Loganiaceae Lolium perenne 031 Poaceae Lomandra Integra 054C Dasypogonaceae Lomandra rigida 054C Dasypogonaceae Lotus angus’tissinrus 165 Papilionaceae Lycoperdon sp F Lycoperdaceae Lyperanthus senatUs 066 Orchidaceae Macrozamia riedlei 016 A Zamiaceae Maireana glomerifolia 105 Chenopodiaceae Malleostemon tuberculatus 273 Myrtaceae Marsdenia graniticola 305 Asdepiadaceae Marsilea drummondii 013 Marsiieaceae Melaleuca adnata 273 Myrtaceae Melaleuca ctenoides 273 Myrtaceae Melaleuca cuticularis 273 Myrtaceae Melaleuca diosrmfolia 273 Myrtaceae Melaleuca eleuterostachya 273 Myrtaceae Melaleuca elliptica 273 Myrtaceae Melaleuca eximia ms 273 Myrtaceae Melaleuca fulgens subsp fulgens 273 Myrtaceae 155 Journal of the Royal Society of Western Australia, 80(3), September 1997 Melaleuca fulgens subsp steedmanii 273 Myrtaceae Melaleuca glaberrima 273 Myrtaceae Melaleuca globifera 273 Myrtaceae Melaleuca hannilosa 273 Myrtaceae Melaleuca holosericea 273 Myrtaceae Melaleuca lasiandra 273 Myrtaceae Melaleuca laxiflora 273 Myrtaceae Melaleuca letocarpa 273 Myrtaceae Melaleuca macronychia subsp macronychia 273 Myrtaceae Melaleuca pentagotui var s ubulifalia 273 Myrtaceae Melaleuca pulchella 273 Myrtaceae Melaleuca pungens var "unsorted" 273 Myrtaceae Melaleuca radula 273 Myrtaceae Melaleuca s cobra 273 Myrtaceae Melaleuca sp 273 Myrtaceae Melaleuca uncinata 273 Myrtaceae Melaleuca urceolaris 273 Myrtaceae Melaleuca r mined 273 Myrtaceae Melilotus indicus 165 Papilionaceae Menegazzia fbraminulosa L Hypogymniaceae Mmegazzia subpertusa L Hypogymniaceae Mentha spicata 313 Lamiaceae Mesumelaena tetragona 032 Cyperaceae Micromyrtus serrulata 273 Myrtaceae Micromyrtus sulphurea 273 Myrtaceae Microtis aff parviflora 066 Orchidaceae Microti 5 atrata 066 Orchidaceae Microtis broumii 066 Orchidaceae Microtis media subsp "inland race” 066 Orchidaceae Microtis media subsp media 066 Orchidaceae Microtis rara 066 Orchidaceae Millotia depauperata 345 Asteraceae Millotia myosotidifolia 345 Asteraceae Millotia tenuifolia var tenuifolia 345 Asteraceae Mirbeha longifolia 165 Papilionaceae Mirbelia sp 165 Papilionaceae Mitrasacme nudicaulis 302 Loganiaceae Mitrasacme sp 302 Loganiaceae Monadenia bracteata 066 Orchidaceae Monotaxis gratidiflora 185 Euphorbiaceae Muehlenbeckia adpressa 103 Polygonaceae Muelleranthus trifolwlatus 165 Papilionaceae Mukia maderaspatam 337 Cucurbitaceae Myriocejjhalus guerinae 345 Asteraceae Myriocephalus nudus 345 Asteraceae Myriocephalus Occident alls 345 Asteraceae Myriocephalus pygmaeus 345 Asteraceae Mynophyllum balladaniense 276 Haloragaceae Myriophyllum laptdicola 276 Haloragaceae Mynophyllum petraeum 276 Haloragaceae Myriophyllum sp 276 Haloragaceae Nelsonia campestns 325 Acanthaceae Nemcia acuta 165 Papilionaceae Neofuscelia pulla L Parmeliaceae Neurachne alopecuroidea 031 Poaceae Nicotiana cavicola 3l5Solanaceae Nicotiana occidental is subsp obliqua 315 Solanaceae Nicotiana rot undifolia 315 Solanaceae Oleana aff paucidentata 345 Asteraceae Olearia algida 345 Asteraceae Oleana muellen 345 Asteraceae Oleana paucidentata 345 Asteraceae Olearia propmqua 345 Asteraceae Olearia stuartii 345 Asteraceae Omphnlappula concava 310 Boraginaceae Opercularia vagin at a 331 Rubiaceae Ophioglossum lusitanicum 005 Ophioglossaceae Ophioglossum sp 005 Ophioglossaceae Osffospcrwum clandestinum 345 Asteraceae Oudemansiella sp / Tncholomataceae Ozothammis alpinu^ 345 Asteraceae Ozof/wnmus ramosus 345 Asteraceae Pandorea fwndorana 317 Bignoniaceac Pamcaleana mgrita 066 Orchidaceae ParacaJeana trims ms 066 Orchidaceae Paraserumthes lophantha subsp lophantha 163 Mimosaceae Parentucellia latifolia 316 Scrophulariaceae Parent ucellia vrscosa 316 Scrophulanaceae Parmelia sp L Parmeliaceae Paspalidium clementii 031 Poaceae Patersonia occidentals 060 Indaceae Pelargonium australe 167 Geraniaceae Pelargonium cf. littorale 167 Geraniaceae Pelargonium drummondii 167 Geraniaceae Pelargonium havlasae 167 Geraniaceae Pelargonium sp 167 Geraniaceae Pentaschistis airoides 031 Poaceae Pertcalymma elhpticum var floridum ms 273 Myrtaceae Persoonia amahae 090 Proteaceae Versoonia falcata 090 Proteaceae Persoonia pentasticha ms 090 Proteaceae Persoonia quinquenervis 090 Proteaceae Petalostylis cassioides 164 Caesalpiniaceae Petrophile divan cat a 090 Proteaceae Petroplnle squamata 090 Proteaceae Petrophile striata 090 Proteaceae Petrorhagia vehitina 113 Caryophyllaceae Phebalium elegans ms 175 Rutaceae Phebalium filifolium 175 Rutaceae Phcbalnim rhytidophyllum 175 Rutaceae Plulotheca langei 175 Rutaceae Philydrella pygmaea subsp "unsorted" 050 Philydraceae Philydrella pygmaea subsp pygmaea 050 Philydraceae Phyllangium divergens 302 Loganiaceae Phyllangium paradoxum ms 302 Loganiaceae Phyllantlms calycinus 185 Euphorbiaceae Phyllanthus sp 185 Euphorbiaceae Pimelea brachyphylla 263 Thymelaeaceae Pimelea ciliata 263 Thymelaeaceae Pimelea ferruginea 263 Thymelaeaceae Pimelea forrestiana 263 Thymelaeaceae Pimelea graniticola 263 Thymelaeaceae Pimelea imbncata var imbricata 263 Thymelaeaceae Pimelea imbncata var piligera 263 Thymelaeaceae Pimelea microcephala subsp nucroccphala 263 Thymelaeaceae Pimelea preissii 263 Thymelaeaceae Pimelea sp 263 Thymelaeaceae Pimelea spiculigera var thesioides 263 Thymelaeaceae Pimelea suaveolens subsp suaveolens 263 Thymelaeaceae Pimelea sulphurea 263 Thymelaeaceae Pittosporum phylliraeoides 152 Pittosporaceae Pityrodia dilatata 31 1 A Chloanthaceae Pityrodia teckiana 311 A Chloanthaceae Pityrodia terminalis 3^1 A Chloanthaceae Plantago debihs 329 Plantaginaceae Platysace compressa/filiformis 281 Apiaceae Platysace compressa 281 Apiaceae Pleuridium nervosum B Ditrichaceae Pteurosorus rulif ilius 01 IE Aspleniaceae Pleurosorus subglandulosus 01 IE Aspleniaceae Pleurotus nidiformis F Polyporaceae Poa annua 031 Poaceae Poa poifonius 031 Poaceae Podolcpis canescens 345 Asteraceae Podolepis capillaris 345 Asteraceae Podolcpis lesson ii 345 Asteraceae Podolepis rugata 345 Asteraceae Podotheca angustifolta 345 Asteraceae Podotheca wilsonii 345 Asteraceae Pogonolepis stricta 345 Asteraceae Polycarpaea breviflora var gracilis 113 Caryophyllaceae Polycarpaea corymbosa var "unsorted" 113 Caryophyllaceae Polygala orbicularis 183 Polygalaceae Parana comnuxta 307 Convolvulaceae Parana sericea 307 Convolvulaceae Pottia drummondii B Pottiaceae Prasophyllum aff parvifolium 066 Orchidaceae Prasophyllum brownh 066 Orchidaceae Prasophyllum cucullatum 066 Orchidaceae Prasophyllum datum 066 Orchidaceae Prasophyllum fimbria 066 Orchidaceae Prasophyllum gibbosvtn 066 Orchidaceae Prasophyllum gibbosum subsp gibbosum 066 Orchidaceae Prasophyllum gractlc 066 Orchidaceae Prasophyllum parvifolium 066 Orchidaceae Prasophyllum r ingens 066 Orchidaceae Prasophyllum sp 066 Orchidaceae Prostanthera baxten 313 Lamiaceae Prostanthera campbellii 313 Lamiaceae Prostanthera carrickiana 313 Lamiaceae 156 journal of the Royal Society of Western Australia, 80(3), September 1997 Prostanthera florifera 313 Lamiaceae Prostanthera grylloana 313 Lamiaceae Prostanthera magnified 313 Lamiaceae Prostanthera serpylhfolia subsp microphylla 313 Lamiaceae Prostanthera sp 1 313 Lamiaceae Prostanthera striatiflora 313 Lamiaceae Prostanthera verticillans 313 Lamiaceae Pterochaeta paniculata 345 Asteraceae Pterostylis aff nana 066 Orchidaceae Pterostylis aff rufa 066 Orchidaceae Pterostylis allantoidea 066 Orchidaceae Pterostylis aspera 066 Orchidaceae Pterostylis barbata 066 Orchidaceae Pterostylis elegantisirna 066 Orchidaceae Pterostylis hamHtomi 066 Orchidaceae Pterostylis rnutica 066 Orchidaceae Pterostylis recurva 066 Orchidaceae Pterostylis roensis 066 Orchidaceae Pterostylis sanguined 066 Orchidaceae Fterostylis sargentii 066 Orchidaceae Pterostylis scabra 066 Orchidaceae Pterostylis vittata 066 Orchidaceae Ptilotus corymbosus var cciri/mhosus 106 Amaranthaceae Ptilotus divaricatus var '‘unsorted" 106 Amaranthaceae Ptilotus gaudichaudii var J' unsorted" 106 Amaranthaceae Ptilotus gaudichaudii var pamflorus 106 Amaranthaceae Ptilotus hum His var hunulis 106 Amaranthaceae Ptilotus incanus var inoinus 106 Amaranthaceae Ptilotus obovatus var " unsorted " 106 Amaranthaceae Ptilotus obovatus var obovatus 106 Amaranthaceae Ptilotus polystachyus forma rubriflorus 106 Amaranthaceae Ptilotus spathulatus forma spathulatus 106 Amaranthaceae Pultenaea elachista 165 Papilionaceae Pultenaea encifolia 165 Papilionaceae Patterned obcordata 165 Papilionaceae Puncteha subrudecta L Parmeliaceae Pyrorchis nigricans 066 Orchidaceae Quinetia urvillei 345 Asteraceae Quinqueremulus linearis 345 Asteraceae Ranunculus colonorum 119 Ranunculaceae Ranunculus sp 119 Ranunculaceae Restio crispatus 039 Restionaceae Rhadinothamnus euphenuae 175 Rutaceae Rhagodia eremaea 105 Chenopodiaceae Rhagodia preissii subsp preissn 105 Chenopodiaceae Rhodanthe battii 345 Asteraceae Rhodanthe chlorocephala subsp splendida 345 Asteraceae Rhodanthe citrina 345 Asteraceae Rhodanthe frenchu 345 Asteraceae Rhodanthe manglesii 345 Asteraceae Rhodanthe rubella 345 Asteraceae Rhodanthe spicata 345 Asteraceae Rhodanthe tietkensii 345 Asteraceae Riccia crinita H Ricciaceae Riccia crystallina H Ricciaceae Riccia lamellosa H Ricciaceae Riccia Hmbata H Ricciaceae Ricinocarpos glaucus 185 Euphorbiaceae Ricinocarpos rosmarinifolius 185 Euphorbiaceae Rimelia reticulata L Parmeliaceae Rinzia sp 273 Myrtaceae Rulingia craurnphylla 223 Sterculiaceae Rulingia cygnorum 223 Sterculiaceae Rulingia kempeana 223 Sterculiaceae Rulingia luteiflora 223 Sterculiaceae Russula "small white" (K Syme 633/931 F Russulaceae Rutidosis multiflora M5 Asteraceae Samolus repens subsp "unsorted" 293 Primulaceae Santalum spicatum 092 Santalaceae Scaevola glandulifera 341 Goodeniaceae Scaevola spinescens 341 Goodeniaceae Scaevola theswides subsp filifolia 341 Goodeniaceae Schizaea sp 006 Schizaeaceae Schoenia ayersii 345 Asteraceae Schoenia cassiniana 345 Asteraceae Schoenus aff odontocarpus 032 Cyperaceae Schoenus multiglumis 032 Cypcraccae Schoenus odontocarpus (Small typical variant) 032 Cyperaceae Schoenus sculptus 032 Cyperaceae Schoenus sp 3 (aff odontocarpus ) 032 Cyperaceae Sclerolaena costata 105 Chenopodiaceae Sclerolaena fusiformis 105 Chenopodiaceae Scyphocoronis major 345 Asteraceae Selaginella granllima 003 Selaginellaceae Sematophyllum homomallum B Sematophyllaceae Senecio hispidulus var hispid ulus 345 Asteraceae Senecio leucoglossus 345 Asteraceae Senna artemisioides subsp x artemisioides 164 Caesalpiniaceae Senna artemisioides subsp x s turtii 164 Caesalpiniaceae Senna glutuwsa subsp charlesiana 164 Caesalpiniaceae Sida calyxhymenia 221 Malvaceae Sida hookeriana 221 Malvaceae Siloxerus fihfolius 345 Asteraceae Siphula coriacea L Siphulaceae Sisymbrium irio 138 Brassicaceae Solatium cleistogamum 315 Solanaceae Solarium ferocissimum 315 Solanaceae Solanum lasiophyllum 315 Solanaceae Solanum lucani 315 Solanaceae Solanum nummularium 315 Solanaceae Solanum orbiculatum subsp orbiculatum 315 Solanaceae Solanum phlomoides 315 Solanaceae Solanum symonii 315 Solanaceae Sollya heterophylla 152 Pittosporaceae Sonchus asper sens lat 345 Asteraceae Spartochloa s cirpnidea 031 Poaceae Spergularia rubra 1 13 Caryophyllaceae Sphenotoma capitatum 288 Hpacridaceae Spiculaea ciliata 066 Orchidaceae Stackhousia monogyna 202 Stackhousiaceae Stackhousla muricata 202 Stackhousiaceae Stenanthemum notiale Subsp notiale 215 Rhamnaceae Stipa compressa 031 Poaceae Stipa hemipogon 031 Poaceae Stipa variabilis 031 Poaceae Stropharia sp F Slrophariaceae Stylidium amoenum var "unsorted" 343 Stylidiaceae Stylidium arenicola 343 Stylidiaceae Stylidium assimile 343 Stylidiaceae Stylidium breviscapum var breviscapum 343 Stylidiaceae Stylidium brunonianum subsp "unsorted" 343 Stylidiaceae Stylidium brunonianum subsp brunonianum 343 Stylidiaceae Stylidium bulbtferum var "unsorted" 343 Stylidiaceae Stylidium caespitosum 343 Stylidiaceae Stylidium calcarotum 343 Stylidiaceae Stylidium crassifolium 343 Stylidiaceae Stylidium despectum 343 Stylidiaceae Stylidium dichotomum 343 Stylidiaceae Stylidium dielsianum 343 Stylidiaceae Stylidium divaricatum 343 Stylidiaceae Stylidium diversifolium 343 Stylidiaceae Stylidium ecorne 343 Stylidiaceae Stylidium emarginatum subsp emarginatum 343 Stylidiaceae Stylidium inundatum 343 Stylidiaceae Stylidium longibradeatum 343 Stylidiaceae Stylidium merrallii 343 Stylidiaceae Stylidium negleetum 343 Stylidiaceae Stylidium nungarinense 343 Stylidiaceae Stylidium perpusillum 343 Stylidiaceae Stylidium pukhellum 343 Stylidiaceae Stylidium pygniaeum 343 Stylidiaceae Stylidium sp 343 Stylidiaceae Stylidium spathulatum subsp " unsnrted " 343 Stylidiaceae Stylidium spathulatum subsp acuminatum 343 Stylidiaceae Stypandra glauca 054E Phormiaceae Swainsona gracilis 165 Papilionaceae Swainsona sp 165 Papilionaceae Synaphea sp 090 Proteaceae Templetoma sulcata 165 Papilionaceae Tephrosia sp 165 Papilionaceae Tephrosia sp B Kimberley Flora (CA Gardner 7) 165 Papilionaceae Tetragonia crislata 110 Aizoaceae Tetrapterum cylindricum B Pottiaceae Tetraria capillaris 032 Cyperaceae Tetratheca aff harperi 182 Tremandraceae Tetratheca deltoidea 182 Tremandraceae Tetratheca hirsuta 182 Tremandraceae Tetratheca hispidissima 182 Tremandraceae Tetratheca nuda 182 Tremandraceae Tetratheca paynterae 182 Tremandraceae 157 Journal of the Royal Society of Western Australia, 80(3), September 1997 Thelymitra SP 066 Orchidaceae Thelymitra aff hoelmsit 066 Orchidaceae Thelymitra aff longifolia 066 Orchidaceae Thelymitra aff nuda 066 Orchidaceae Thelymitra aff pauciflora 066 Orchidaceae Thelymitra antemufera 066 Orchidaceae Thelymitra benthamiana 066 Orchidaceae Thelymitra crimta 066 Orchidaceae Thelymitra cucullata 066 Orchidaceae Thelymitra dedmaniarum 066 Orchidaceae Thelymitra flexuosa 066 Orchidaceae Thelymitra macmillanii 066 Orchidaceae Thelymitra macrophylla 066 Orchidaceae Thelymitra nuda 066 Orchidaceae Thelymitra sp 066 Orchidaceae Thelymitra sp 066 Orchidaceae Thelymitra spiralis 066 Orchidaceae Themeda tnandra 031 Poaceae Thomasia foliosa 223 Sterculiaceae Thomasia glutmosa var latifolia 223 Sterculiaceae Thomasia grandiflora 223 Sterculiaceae Thomasia multiflora 223 Sterculiaceae Thomasia pauciflora 223 Sterculiaceae Thomasia petalocalyx 223 Sterculiaceae Thomasia sp 223 Sterculiaceae Thryptomene australis 273 Myrtaceae Thryptomene mucronulata 273 Myrtaceae Thryptomene saxicola 273 Myrtaceae Thryptomene sp 273 Myrtaceae Thuidium sparsum var hastatum B Thuidiaceae Thysanothecium hookeri L Cladoniaceae Thysanotus dichotomus 054F Anthericaceae 77iysflnctfus manglesianus 054F Anthericaceae T/n/samitws sparteus 054F Anthericaceae Thysanotus tenellus 054F Anthericaceae Thysanotus triandrus 054F Anthericaceae Tinospora smilacina 122 Menispermaceae Tolpis bar bat a 345 Asteraceae Tortula antarctica B Pottiaceae Tortula pnncejjs B Pottiaceae Trachymenc cyanopetala 281 Apiaceae Trachymene moorei subsp Tutanning (AS George 12867) ms 281 Apiaceae Trachymene ornata 281 Apiaceae Trachymene pilosa 281 Apiaceae Trachymene sp Pilbara (RA Saffrey 1117) ms 281 Apiaceae Tremandra diffusa 182 Tremandraceae Tremandra s telligera 182 Tremandraceae Tribonanlhes australis 055 Haemodoraceae Tnbonanthcs longipetala 055 Haemodoraceae Tribonanthes purpurea 055 Haemodoraceae Tnbulus terrestris 173 Zygophyilaceae Trichodesma zeylanicum var "unsorted" 310 Boraginaceae Tricoryne elatior 054F Anthericaceae Tnfohum dubtum 165 Papilionaceae Tnglochin aff mmutissimum 026 Juncaginaceae Triglochin calcitrapum subsp incurvum ms 026 Juncaginaceae 7 riglochm centrocarpum 026 Juncaginaceae Triglochin mmutissimum 026 Juncaginaceae Tnglochin sp 026 Juncaginaceae Tnpogon loliiformis 031 Poaceae Tripterococcus brutwms 202 Stackhousiaceae Tnptilodiscus pygmaeus 345 Asteraceae Triquetrella papillata B Pottiaceae Trymalium ledifolium var rosmarinifolium 215 Rhamnaceae Trymahum spatulatum 215 Rhamnaceae Tylophora sp 305 Asclepiadaceae Ursinia anthemoides 345 Asteraceae Usnea confusa L Parmeliaceae Usnea inermis L Usneaceae Utricularia menziesii 323 Lentibulariaceae Utricularia multifida323 Lentibulariaceae Utricularia violacea 323 Lentibulariaceae Velleia cycnopotamica 341 Goodeniaceae Velleia trinervis 341 Goodeniaceae Verticordia acerosa var acerosa 273 Myrtaceae Verticordia acerosa var preissii 273 Myrtaceae Verticordia brownii 273 Myrtaceae Verticordia chrysantha 273 Myrtaceae Verticordia chrysanthella 273 Myrtaceae Verticordia endheheriana var angustifolia 273 Myrtaceae Verticordia habrantha 273 Myrtaceae Verticordia helmsii 273 Myrtaceae Verticordia huegelh var decumbens 273 Myrtaceae Verticordia huegelii var huegelii 273 Myrtaceae Verticordia huegelii var tridens 273 Myrtaceae Verticordia minutiflora 273 Myrtaceae Verticordia penicillaris 273 Myrtaceae Verticordia pemugera 273 Myrtaceae Verticordia plumosa 273 Myrtaceae Verticordia plumosa var grandiflora 273 Myrtaceae Verticordia plumosa var plumosa 273 Myrtaceae Verticordia staminosa subsp cyhndracea var cylindracea 273 Myrtaceae Verticordia staminosa subsp cylindracea var erecta 273 Myrtaceae Verticordia staminosa subsp staminosa 273 Myrtaceae Vimniaria juncea 165 Papilionaceae Vulpia myuros 031 Poaceae Wahlenbergia preissii 339 Campanulaccae Waitzia acuminata var acuminata 345 Asteraceae Waitzia nitida 345 Asteraceae Waitzia podolepis 345 Asteraceae Waitzia suaveolens var flava 345 Asteraceae Weissia brachycarpa B Pottiaceae Weissia controversa B Pottiaceae W&S5M rutilans B Pottiaceae Westringia dampieri 313 Lamiaceae Wurmbea cernua 054J Colchicaceae Wurmbea densiflora 054J Colchicaceae Wurmbea dioica 054J Colchicaceae Wurmbea dioica subsp alba 054J Colchicaceae Wurmbea murchisoniana 054J Colchicaceae Wurmbea pygmaea 054J Colchicaceae Wurmbea tenella 054J Colchicaceae Xanthopanneha digitiformis L Parmeliaceae Xanthoparmcha hypoleia L Parmeliaceae Xanthoparmelia isidiigera L Parmeliaceae Xanthopanneha neotmetina L Parmeliaceae Xanthoparmelia replans L Parmeliaceae Xanthoparmelia substrigosa L Parmeliaceae Xanthoparmelia tasmanica L Parmeliaceae Xanthoria parictina L Teloschistaceae Xanthosia Candida 281 Apiaceae Xanthosia singuliflora 281 Apiaceae Xerolirion divancata 054C Dasypogonaceae 158 Journal of the Royal Society of Western Australia, 80:159-166, 1997 Terrestrial fauna of granite outcrops in Western Australia P C Withers & D H Edward Department of Zoology, The University of Western Australia, Nedlands WA 6907 email: pwithers@cyllene.uwa.edu.au Abstract In our overview of the terrestrial animals associated with granite outcrops in Western Australia, we document relatively few animals which are known to be restricted to granite outcrops, despite the large area over which granite outcrops occur in Western Australia. The spider Teyl and the chironomid fly Archaeochlus are restricted to granite outcrops. Other spiders (e.g. Rebilus), some pseudoscorpions (e.g. Synsphyronus) and the embiopteran web-spinner (Notoligotoma) may be restricted to granite outcrops. No amphibians are restricted to granite outcrops, although many species use them for shelter or breeding, and some species seem to be restricted to the eastern extent of the granites along the south coast of Western Australia. Only four reptiles appear to be restricted to granite outcrops, the dragon (agamid) lizards Ctenophorus ornatus, C. yinnietharra and C. rufescens , and the gecko Gehyra montium, but many other species have been recorded from granite outcrops. Birds, being more mobile than other terrestrial vertebrates, range widely and no species are restricted to granites, although many species are found on granite outcrops. No mammals appear to be restricted to granite outcrops, although a number are rock specialists (e.g. rock wallabies Petrogale spp, long-tailed dunnart Sminthopsis longicauda, rock rats Zyzomys spp, rock ringtail possum Pseudocheirus dahli), and various other species use granite outcrops as well as other habitats. Although relatively few species of terrestrial animals appear to be restricted to granite outcrops, many animals are found in these specialised habitats, and it is clear that granite outcrops are important as a seasonal resource or temporary refuge for the fauna of the surrounding habitats. The fringing apron of granite outcrops is an important site of interaction between granite outcrops and their surrounding habitats, and must be preserved as vigorously as the granite outcrops themselves. Introduction One of the objectives of this symposium, and this issue of The Journal of The Royal Society of Western Australia, was to examine and hopefully explain the pattern of distribution of some animals and plants in Western Australia, with respect to the extensive granite outcrops that are a major geological feature of much of the State. Here, we overview the terrestrial animals associated with granite outcrops in Western Australia. Granite outcrops form a complex part of the varied ecosystems in Western Australia, although they are often perceived as isolated rock forms jutting out of the surrounding landscape. Not only do they form a specialised suite of habitats for animals and plants in their own right (e.g. rock pools, meadows, exfoliating rock sheets, rock crevices) but they also merge with the surrounding habitats and form specialised edge habitats, the fringing apron. Thus, the flora and fauna associated with granite outcrops are not only those species adapted to survive and persist on the granite rock habitats, but include many species from the surrounding habitats that seek temporary or permanent refuge amongst the granite rocks, or on the fringing apron. Even a cursory examination of published accounts of the fauna identified on granite outcrops reveals that an immense number of animals has been reported to be © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 associated with granite outcrops on either a temporary or permanent basis (c.£. the Western Australian Museum Biological Survey reports; Lovell 1978; Dell et ah 1985; Anon, 1991; Hopper 1981), various unpublished reports of granite rock fauna and flora surveys, field guides to Western Australian animals (e.g. Storr et al. 1981, 1983, 1986, 1990; Storr & Johnstone 1985; Tyler et al. 1994) and more general guides to Australian animals (e.g. Strahan 1983; Wilson & Knowles 1988; Cogger 1992). This reflects the opportunistic or seasonal activities of many of these animals rather than any predilection or requirement for granite outcrop habitats. Therefore, for practical reasons, the scope of our investigation here has been limited to the terrestrial fauna restricted to granite outcrops in Western Australia. Distribution of Granites Granite outcrops are distributed widely throughout Western Australia. Although there does not appear to be a specific map of granite outcrops in Western Australia, the distribution of granite and sandstone landforms clearly indicates a widespread distribution of granite outcrop habitats over much of the western half of Western Australia, and sandstone habitats in northern, eastern and west-central Western Australia (Fig 1; see also Myers, 1997). Much of the western part, particularly the Yilgarn cratonic block, is covered by granite, and there are scattered patches of granite in the East and West Pilbara. There is a pronounced "V"-shaped patch of granite in the Kimberley, and scattered patches of granite 159 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 1. Distribution of granitoid (dark) and sandstone (light) landforms in Western Australia. Redrawn from the Hydrogeological Map of Western Australia, Geological Survey of Western Australia (1989); not shown are other Precambrian and Cambrian landforms (shale, limestone, dolomite, undifferentiated sedimentary, metamorphic, basalt, dolerite, volcanic and gneiss and Phanerozoic sedimentary rocks. along the eastern border, particularly around the Musgrave Ranges. The granite outcrops of Western Australia were formed in three major, distinct orogens, representing separate episodes of continental collision (see Myers 1997). The most extensive granite outcrops are part of the Yilgam Craton, and were formed between 2700 and 2600 million years ago, by the melting of continental crust; these include the Pilbara granites, which are mostly buried by volcanic rocks and the Hammersley banded iron formations. The younger granite outcrops of the Kimberley and Gascoyne regions result from the colliding, about 2000 million years ago, of the Pilbara and Yilgam continents (Gascoyne region, Capricorn orogen), and others (Kimberley region. Halls Creek and King Leopold orogens). More recently, granite was formed in the south of Western Australia in two distinct episodes of continental crust collision. The older of these, the Recherche granite, was intruded about 1300 million years ago by deformation of the Mawson Craton (South Australia and East Antarctica) and the West Australian Craton at the Albany-Fraser Orogen. Younger Esperance granite, about 1180 million years old, may be largely derived from another episode of compression between the Mawson and West Australian Cratons. The Musgrave granites formed at about the same time through deformation between the Mawson Craton and the North Australian Craton. It is clear that the times of formation of these different groups of granite outcrops in Western Australia immensely pre-date the presence of existing species of terrestrial vertebrates and invertebrates, and so current 160 Journal of the Royal Society of Western Australia, 80(3), September 1997 distributions of terrestrial animals on granite outcrops must reflect relatively recent (in geological time) evolutionary and dispersal patterns rather than relictual isolation of species on the various groups of granites. Terrestrial Invertebrates Consideration of the various terrestrial invertebrates indicates that only a few arthropods are known to be restricted to granite outcrops. The mygalomorph spider Teyl hiculentus is found in virtually all of the meadows, and also on the fringing apron, of granite outcrops in the wheatbelt and western Goldfields areas of south-western Australia (B York Main, pers. comm.). Teyl (Fig 2) is an ancient trapdoor spider genus and its distribution appears to be relictual, associated with the wetter, boggy meadows and aprons of granite outcrops; T. hiculentus is doubtless a complex of species to be described (B York Main, pers. comm.). Main (1975) listed various granite rocks as habitats for T. hiculentus, but some of those records now relate to new species, some restricted to specific rocks, as the genus is being revised; all Teyl are restricted to granite outcrops or granite-related habitats (apron, granitic soils; B York Main, pers. comm.). Teyl aestivates during the dry periods by encasing itself in its sealed burrow. A number of other mygalomorph spiders have been recorded from the fringing apron or outer edge of granite outcrops, including Kwonkan sp, Merredinia damsonoides, Aganippe sp, Gains sp, Barychelidae sp, Chenistonia tepperi and Aname diversicolor , but are not restricted to granite outcrops (B York Main, pers. comm.). The most common and conspicuous spiders of granites in the wheatbelt and goldfields are the large spiders found in vegetation around the granites e.g. the golden orb weaver ( Nephila edulis), Christmas spider (Gasteracantha minax) and huntsman spiders ( Olios sp). Many spiders are well adapted to living in granite .4* Figure 2. The burrowing mygalomorph spider Teyl luculentus is commonly found in small meadows on granite rocks and the seasonal boggy apron habitats (B York Main). Figure 3. The pseudoscorpions Synsphyronus sp are known from granite outcrops, (D Elford, Western Australian Museum). outcrops, although not restricted to them, being also found in vegetation and under tree bark. The sparassid huntsmen spiders usually live in cracks and crevices of rocks, as well as under loose bark; they have a flattened cephalothorax and abdomen, with legs rotated forwards, making them adept at moving in narrow spaces in rock cracks and between slabs of exfoliating granite. The clubionid and gnaphosid sac spiders, Rebilus , Hemicloea, Miturga and some lycosids are often found under exfoliating granites, as also are redback spiders ( Latrodectus ). The relatively large Rebilus is a common spider under the exfoliating granites of the central wheatbelt in Western Australia; it is also found under bark on trees. In the higher rainfall south-western granites, it is replaced by Miturga , and in the lower rainfall eastern wheatbelt and goldfields it is replaced by the wolf spider Pardosa (Main, 1975). Rebilus has a flattened body, like the sparassid huntsmen spiders. The dorso-ventrally compressed selenopid spiders are found in rock piles and under bark. Some wolf spiders, and the dysderid spider Ariadna, are also found with Teyl in the boggy meadows and aprons of granite outcrops. Ariadna forms a silk tube, from which radiate silk threads. Some species of the pseudoscorpion genus Synsphyronus (Fig 3) may be restricted to granite outcrops; S. elegans Beier is known only from Yorkrakine Hill, S. leo Harvey is known only from Lion Island, Recherche Archipelago, and two undescribed species have been collected from other granite outcrops (M Harvey, pers comm). However, further collecting is needed to confirm that these species are definitely restricted to granite outcrops, since most species of Synsphyronus are found under the bark of trees, in leaf litter, and under other types of rocks (Harvey 1987). The Embioptera (web-spinners), a small order of insects, are mainly tropical species but some occur in warm temperate climates, like the related termites and earwigs (Ross 1991). These small, slender insects, with a large head and eyes, live in silk tunnels which they weave 161 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 4. The distribution of Archaeochlus brundini and A. sp nov is restricted to the ancient granite outcrops in Western Australia (AGSO bedrock data, acid/intermediate intrusions; graphed using Arcview by staff of the Western Australian Museum). using silk glands and spinning glands located on the front tarsi; they feed on vegetation. The embiopteran Notoligotoma may be restricted, in part, to granite outcrops. Ross (1991) states that " Notoligotoma hardyi and a complex of related species or races occur in south¬ western Australia, usually on the undersurfaces of exfoliated rock slabs on granitic outcrops of arid regions. In Perth, N. hardyi is common in crevices of old board fences in residential areas". The aquatic invertebrates that occur in pools on granite outcrops are discussed by Bayly (1982, 1997). Another important aquatic habitat on granite outcrops is the temporary streams formed by seepages from the meadows on the outcrops. In these seepages are the larvae of Archaeochlus, an ancient genus of chironomid fly. Archaeochlus spp are restricted to granite outcrops in south-western Australia (Fig 4) and survive the dry period as a gravid adult female. The genus is also known from temporary streams in the Drakensberg Escarpment of southern Africa, and provides evidence of a Gondwanan connection between Australia and Africa. The two continental lineages are considered to antedate the breakup of Gondwanaland in the Upper Jurassic and have therefore been separated for a minimum of 120 million years (Cranston et al. 1987; Edward 1989). Terrestrial Vertebrates As with invertebrates, it is difficult to discern which terrestrial vertebrate species are restricted to granite outcrops, as a very large number of species of amphibians, reptiles, birds and mammals can be found associated with granite outcrops either permanently, temporarily or seasonally. No fishes, amphibians, snakes, birds or mammals are restricted to granite outcrops, but four lizards (three dragons and one gecko) are apparently restricted to granite outcrops (Table 1). 162 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 1 Summary of the number of species of frogs, lizards and snakes found in Western Australia, indicating the number of species restricted to granite rocks (g) or rocks in general (r; usually sandstone or limestone). Based on Tyler et al. (1994), Storr et al. (1981, 1983, 1986, 1 9 9 O' ) Wilson & Knowles (1988), and Cogger (1992). FROGS 77 (Og, 5r) Leptodactylidae Arenophyrne 1 Crinia 1 Geocrinia 5 Heleioporus 5 Limnodynastes 6 Megistolotus 1 (lr) Metacrinia 1 Myobatrachus 1 Neobatrachus 8 No to den 3 Pseudophryne 3 (lr) Ranidella 5 Uperoleia 12 Hylidae Cyclorana 7 Litoria 18 (3r) SNAKES 80 (Og, Or) Typhlopidae Ramphotyphlops 18 Boidae Aspidites 2 Morelia 7 Acrochordidae Chersydrus 1 Homalopsidae Cerberus 1 Fordonia 1 Myron 1 Colubridae Amphiesma 1 Boiga 1 Detidrelaphis 1 Elapidae Acanthophis 3 Crypt ohis 1 Demansia 7 Denisonia 4 Furina 1 Notechis 5 Oxy uranus 1 Pseuchis 2 Pseudonaja 5 Rhinoplocephalus 6 Vermicella 11 Aclys Aprasia Delma Lialis Pletholax Pygopus Scincidae Bassiana Carlia Cryptoblepharus Ctenotus Egernia Eremiascincus Hemiergis Leris t a Menetia Morethia Notoscincus LIZARDS 270 (4g, 25r) Proablepharus Sphenomorphus Tiliqua Gekkonidae Crenadactylus 1 Agamidae Diplodactylus 28 (2r) Caimanops Gehyra 9 (lg, 2r) Chelosania Heteronotia 3(2r) Chlamydosaurus Nephrurus 6 Ctenophorus Oedura 6 (2r) Diporiphora Phyllodactylus 1 Gemmatophora Pseudothecadactylus 1 (lr) Moloch Rhynchoedura 1 Pogona Underwood isau rus 1 Tympanocryptis Pygopodidae Varanidae Varanus 1 8 10 1 1 2 1 6 4 (lr) 45 (5r) 14 (2r) 2 3 27 (2r) 4 7 2 2 2 4 4 1 1 1 17 (3g, lr) 11 4 1 3 6 18 (5r) Amphibians None of the recorded 77 species of Western Australian amphibians (frogs) appear to be restricted to granite outcrops, based on their distributions or habitats, although a number of species are sometimes associated with rocks (including granites) as either adults or eggs/ tadpoles (see Tyler et al. 1994). For example, adult treefrogs ( Litoria rubella) will shelter in cracks in rocks; eggs and tadpoles of frogs, such as the kunapalari frog ( Neobatrachus kunapalari ), are often found in gnammas on granite outcrops, or in ponds around the apron of outcrops. A notable relationship, however, exists in the close correspondence of the eastern-most extent of granites along the south coast of Western Australia with the eastern-most limit of the distributional range for a number of frogs (L Smith, pers. comm.). Examination of the distribution of Western Australian frogs (see Tyler et al. 1994) shows a reasonable correspondence for over 10 species with the easternmost extent of the granites at about Israelite Bay (Fig 5), which also coincides with the eastern limit of the south-western winter rainfall zone. This coincidence suggests that these species of frog, although not restricted to granites, might be limited in the eastern extent of their distribution by the absence of granite outcrops that either directly (as rock pools) or indirectly (as pools formed by run-off from rocks) provide suitable breeding sites elsewhere. In contrast to no species being restricted to granite rocks, five frog species appear to be restricted to other rock habitats in the Pilbara and Kimberley regions of Western Australia (Table 1). The woodworker frog Megistolotis lignarius is always associated with rocks, being found beneath rock piles, on open escarpments or in caves (Tyler et al. 1994). The habitat of Douglas7 toadlet Pseudophryne douglasi is permanent seeps or deep, shaded pools in deep gorges and canyons (Main 1965). Three treefrogs, the cave-dwelling frog Litoria cavernicola, Copeland's rock frog L. copelandi and the rockhole frog L. meriana, are found only near rock pools, caves or streams over rock in the Kimberley region. Reptiles Although many lizards and snakes have been collected under exfoliating slabs or rocks on granite outcrops (see Storr et al. 1981, 1983, 1986, 1990; Wilson & Knowles 1988; Cogger 1992), only four species of lizards appear to be restricted to granite outcrops. 163 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 5. Relationship between eastern distributional limits for 13 species of Western Australian frogs and the eastern extent of granite outcrops; the eastern limits of the south west winter rainfall zone (annual rainfall 300 and 500 mm isohyets) are also indicated. The species of frogs are; Crinia georgiana, Heleioporus eyrei, Limnodynastes dorsalis, Litoria adelaidensis, Litoria cyclorhynchus, Myobatrachus gouldii, Neobatrachus albipes, Neobatrachus kunapalari , Neobatrachus pelobatoides, Pseudophryne guentheri, Pseudophryne occidentalis, Ranidella insignifera and Ranidella subinsignifera. Distributional data (total sample size is 6925) from records of the Western Australian Museum; frog localities are plotted over the acid /intermediate intrusions (AGSO bedrock data, acid /intermediate intrusions; graphed using Arcview by staff of the Western Australian Museum). The most obvious lizards restricted to granite outcrops are some of the Ctenophorus dragons. The ornate dragon, Ctenophorus ornatus, is commonly found on granite outcrops of the semiarid and subhumid zones of southwestern Western Australia, where it favours expanses of bare rock strewn with exfoliations and loose rocks (Bradshaw 1971). It is restricted to certain granites of the Yilgarn and the Esperance and Recherche regions by cold temperatures and aridity, and does not occur on the Pilbara, Musgrave or Kimberley granites (Fig 6A). Other species of Ctenophorus occur in these granite regions, some of which are also restricted to granites. In the Pilbara, the Yinnietharra rock dragon (C. yinnietharra) is known only from or near low granite outcrops, and the rusty dragon (C. rufescens ) is found on granite outcrops in the Musgrave Range region of Western Australia (Fig 6A). in contrast, other saxicolous Ctenophorus dragons in the Pilbara, particularly the ring¬ tailed dragon (C. caudicinctus ) but also to a lesser extent the Western netted dragon (C. reticulatus ), are found more widely on the other rock forms such as volcanic and banded iron formations, as well as granite outcrops. Only one gecko is apparently restricted to granite outcrops. Gehyra montium (Fig 6B) occurs only on the granite outcrops of the Musgrave Range region of east Western Australia, having a similar distribution as the rusty dragon (C. rufescens). One other gehyra gecko, the spotted dtella G. punctata , is a saxicolous species that is restricted in part to granite rocks, but also occurs on other rock forms (Fig (SB), like the ring-tailed dragon (C. caudicinctus). Two diplodactyline geckos, Diplodactylus wilsoni and D. wombeyi, and Heteronotia spclea are rock- dwellers in the Pilbara region. Another wide-spread gehyra gecko, the spotted dtella G. variegata, is predominantly arboreal, but in the western part of its range (i.e. Western Australia) is both arboreal and saxicolous, under exfoliating rocks on granite outcrops. A similar dichotomy of arboreal and exfoliating granite 164 Journal of the Royal Society of Western Australia, 80(3), September 1997 A: Dragons Ctenophorus yinnietharra Ctenophorus rufescens Ctenophorus ornatus B: Geckos Phyllodactylus marmoratus Gehyra montium Gehyra pilbara Figure 6. A: Distribution of Ctenophorus ornatus (n = 802 records), C. yinnietharra (n = 19) and C. rufescens (n = 25) with respect to granite landforms in Western Australia. B: Distribution of Phyllodactylus marmoratus (n = 1337), Gehyra punctata (n = 696) and Gehyra montium (n = 83) with respect to granite landforms in Western Australia. Distributional data from the Western Australian Museum; localities are plotted over the acid /intermediate intrusions (dark; corresponding roughly to granitoid in Fig 1) and proterozoic sediments (light; corresponding roughly to sandstone in Fig 1). AGSO bedrock data; graphed using Arcview by staff of the Western Australian Museum. habitats is observed for the marbled gecko Phyllodactylus (Christinas) marmoratus (Fig 6B), with the coastal and island populations tending to use limestone and granite rocks more than trees. Similarly, the marbled velvet gecko ( Oedura marmorata) is both arboreal and saxicolous. In contrast to the relatively low number (4) of lizards restricted to granite rocks, a larger number (30) are restricted to sandstone rocks, or rocks in general (Table 1). Compared to one gecko restricted to granites, 9 species are restricted to sandstone or rocks in general. Compared to 3 dragons restricted to granites, only one additional species C. caudicinctus is restricted to rocks in general. Although no skinks were restricted to granites, about 10 species are restricted to other rocks. Finally, a number of goannas are restricted to rocky areas, particularly the sandstones of the Kimberley. Birds Birds, being generally more mobile than other terrestrial vertebrates, are less likely to be restricted to specific habitats, such as granite outcrops. Thus, although many species of bird have been recorded on granite outcrops (e.g. McKenzie et al. 1973; Dell & Johnstone 1976; Dell 1977; Youngson & McKenzie 1977; Hopper 1981; Dell et al. 1985), no species are restricted to them. In fact, the dominant shrubs and mallees in the fringing apron around the rocks appear to be the main attractant to these birds, such as honey-eaters, and so granite rocks are exceptionally rich in honey-eater species compared with most wheatbelt habitats. Granite rocks represent one of the few habitat types that have not experienced extensive habitat destruction for agriculture and this increases their attractiveness to bird species (Hopper 1981). Water, when available in rock pools and gnammas, also attracts a number of bird species to granite outcrops. Mammals As with amphibians and birds, there are apparently no mammals restricted to granite outcrops, although a number of species are restricted to rocky areas. Rock wallabies ( Petrogale spp) are an obvious example, being found in rocky habitats, including but certainly not exclusive to granite outcrops (Strahan 1983). The long¬ tailed dunnart (Sminthopsis longicaudala) is found in rugged, rocky areas of central Western Australia, and its long tail and striated foot pads apparently aid in rock climbing (Burbidge et al. 1983). Rock rats (Zyzomys argurus, Z. woodwardi) always are associated with rocky outcrops, particularly sandstones, of the Pilbara and 165 Journal of the Royal Society of Western Australia, 80(3), September 1997 Kimberley. The rock ringtail possum ( Pseudocheirus dahli) is found exclusively in rock outcrops of the Kimberley (Nelson & Kerle 1983). Conclusions Although granite outcrops occur over a large area of Western Australia, especially the western Yilgarn and Pilbara regions, very few species of terrestrial animals are apparently restricted to them. A few terrestrial arthropods, primarily Teyl spiders and the insect Archaeochlus, are known to be restricted to granites, which are presumably relictual habitats for these ancient genera. Some species of Synsphyronus pseudoscorpion and the embiopteran insect Notoligotoma may also be restricted to granite outcrops. Only four lizards are known to be restricted to granite outcrops, three dragons Ctenophorus ornatus , C. yinnietharra and C. rufescens, and the gecko Gehyra montium. In contrast to the sparse list of species known to be restricted to granite outcrops, a very large number of terrestrial animals have been reported from granite outcrops. It is clear that granite outcrops are important as a seasonal resource for many animals or as a temporary refuge for the fauna of the surrounding habitats. In this regard, the fringing apron of modified habitat surrounding granite outcrops is an especially important aspect of the interaction between granite outcrops and their surrounding habitats, and must be preserved as vigorously as the granite outcrops themselves. Acknowledgements: We thank numerous staff of the Western Australian Museum for their valuable discussions and generous sharing of data and ideas; R How, J Dell, L Smith, M Cowan, K Aplin, R Johnstone, M Harvey. We also are indebted to B York Main for sharing her wealth of knowledge on spiders. References Anon 1991 Ecological survey of Abydos-Woodstock Reserve, Western Australia. Western Australian Museum, Perth. Bayly I A E 1982 Invertebrate fauna and ecology of temporary pools on granite outcrops in southern Western Australia. Australian Journal of Marine and Freshwater Research 33:599-606. Bayly I A E 1997 Invertebrates of temporary waters in gnammas on granite outcrops in Western Australia. Journal of the Royal Society of Western Australia 80:167-172. Bradshaw S D 1971 Growth and mortality in a field population of Amphibolurus lizards exposed to seasonal cold and aridity. Journal of Zoology 165:1-25. Burbidge A A, McKenzie N L & Fuller P J 1983 Long-tailed dunnart Sminthopsis longicaudata. In: The Australian Museum Complete Book of Australian Mammals (Ed R Strahan). Angus & Robertson Publishers, London, 58-59. Cogger H G 1992 Reptiles and Amphibians of Australia. Reed Books, Chatswood. Cranston P S, Edward D H D & Colless D H 1987 Archaeochlus Brundin: A midge out of time (Diptera: Chironomidae). Systematic Entomology 12:313-334. Dell J 1977 Birds of Bendering and West Bendering Nature Reserves. Records of the Western Australian Museum Suppl 5:31-46. Dell J & Johnstone R E 1976 Birds of Tarin Rock and North Tarin Rock Reserves. Records of the Western Australian Museum Suppl 2:69-84. Dell J, How R A, Newbey K R & Hnatiuk R J 1985 The Biological Survey of the Eastern Goldfields of Western Australia. Western Australian Museum, Perth. Edward D H D 1989 Gondwanaland elements in the Chironomidae (Diptera) of south-western Australia. Proceedings of the Xth International Symposium on Chironomidae. Acta Biologica Debrecina Suppl Oecologia Hungarica 2:181-187. Harvey M S 1987 A revision of the genus Synsphyronus Chamberlin Garypidae: Pseudoscorpionida: Arachnida. Australian Journal of Zoology Supplementary Senes 126:1-99. Hopper S D 1981 Honeyeaters and their winter food plants on granite rocks in the central wheatbelt of Western Australia. Australian Wildlife Research 8:187-197. Lovell A F 1978 Biological survey of the Western Australian wheatbelt. Part 6: Durokoppin and Kodj Kodjin Nature Reserves. Western Australian Museum, Perth. Main A R 1965 Frogs of southern Western Australia. Western Australian Naturalists' Club, Perth. Main B Y 1975 The citrine spider: a new genus of trapdoor spider Mygalomorphae: Dipluridae. Western Australian Naturalist 13:73-78. McKenzie N L, Burbidge A A & Marchant N G 1973 Results of a biological survey of a proposed wildlife sanctuary at Dragon Rocks near Hyden, Western Australia. Report Number 12, Western Australia Department of Fisheries and Fauna, Perth. Myers J S 1997 Geology of granite. Journal of the Royal Society of Western Australia 80:87-100. Nelson J E & Kerle J A 1983 Rock ringtail possum. In: The Australian Museum Complete Book of Australian Mammals. The National Photographic Index of Australian Wildlife (ed R Strahan). Angus & Robertson Publishers, London, 132. Ross E S 1991 Embioptera (Embiidina). In: The Insects of Australia. CSIRO, Canberra & Melbourne University Press, Melbourne, 405-409. Storr G M & Johnstone R E 1985 A field guide to the birds of Western Australia. Western Australian Museum, Perth. Storr G M, Smith L A & Johnstone R E 1981 Lizards of Western Australia. I. Skinks. Western Australian Museum, Perth. Storr G M, Smith L A & Johnstone R E 1983 Lizards of Western Australia. II. Dragons and Monitors. Western Australian Museum, Perth. Storr G M, Smith L A & Johnstone R E 1986 Snakes of Western Australia. Western Australian Museum, Perth. Storr G M, Smith L A & Johnstone R E 1990 Lizards of Western Australia. III. Geckos and Pygopods. Western Australian Museum, Perth. Strahan R 1983 The Australian Museum Complete Book of Australian Mammals. The National Photographic Index of Australian Wildlife. Angus & Robertson Publishers, London. Tyler M J, Smith L A & Johnstone R E 1994 Frogs of Western Australia. Western Australian Museum, Perth. Wilson S K & Knowles D G 1988 Australia's Reptiles. A Photographic Reference to the Terrestrial Reptiles of Australia. Collins Publishers Australia, Sydney. Youngson W K & McKenzie N L 1977 The wildlife of the proposed Karroun Hill Nature Reserve, Western Australia. Report 30, Western Australia Department of Fisheries and Wildlife, Perth. 166 Journal of the Royal Society of Western Australia, 80:167-172, 1997 Invertebrates of temporary waters in gnammas on granite outcrops in Western Australia I A E Bayly Department of Biological Sciences, Monash University, Clayton VIC 3168 Abstract Thirty-six flooded gnammas (rock pools), distributed between 17 different granite outcrops, were each sampled once only during the winter of 1990. The water of most pools was acidic (pH < 7.0) and of low conductivity (K25 < 200 pS cm1). A total of 88 invertebrate taxa were found with Boeckella opaqua (Copepoda; 19 occurrences), Cypretta baylyi (Ostracoda; 17) and Neothrix armata (Cladocera; 14) most common. A high proportion of the Crustacea consisted of species endemic to Western Australia. Six new species of Cypricercus (Ostracoda) were discovered. The mean number of taxa was 8.2 per pool. There was a highly significant positive correlation between species richness and the logarithm of both the pool volume and pool area. The nature of the fauna is reviewed within the framework of the type of adaptations employed by the animals for tolerating or avoiding the dry phase. Introduction The word "gnamma" is of Aboriginal origin and refers to a depression that has been weathered out of the surface of an outcrop of bare rock; some overseas workers ( e.g . Smith 1941) have referred to these depressions as " weather pits". Gnammas, or weather pits, are commonly found on granite outcrops and especially on the top of domed inselbergs. Despite the widespread occurrence of gnammas in Australia, the aquatic biota of rain-filled gnammas has received remarkably little attention. This neglect might be partly explicable on the basis that, because gnammas are decidedly at the small end of the size spectrum of lentic waters (pool-pond- lake), they have been regarded as being of little interest or consequence to aquatic ecologists. In fact, small size confers on gnammas a number of methodological advantages, such as ease of sampling and experimental manipulation, as subjects for an ecologist. Sheldon (1984), for example, draws attention to the good potential of rock pools for comparative and experimental studies of the aerial colonization dynamics of adult aquatic insects. This claim for gnammas, along perhaps with phytotelmata (pockets of water held in plants such as the Albany Pitcher Plant; Bayly 1984), as being quintessential aquatic microcosms, is appealing. It may be claimed with some confidence, therefore, that the potential of studies of gnammas for illustrating ecological principles is just as great, if not greater, than that of large lakes. Among the earliest written observations on the gnammas of Western Australian granite outcrops are those of the explorer Carnegie (1898). Interesting comments on the form, distribution and likely origin of granitic gnammas in this State were also recorded by the geologist Jutson (1934). Taxonomic work on invertebrates collected from Western Australian © Royal Society of Western Australia 1997 Granite Outcrops Symposium. 1996 gnammas has been undertaken by Wolf (1911), Fairbridge (1945), Petkovski (1973), Wallwork (1981), De Deckker (1981), Frey (1991, 1998), Lansbury (1995), Smirnov & Bayly (1995) and Benzie & Bayly (1996). Studies of the ecology of these gnammas, or of specific animals living in them, have been made by Jones (1971, 1974) and Bayly (1982). Thus far only a limited attempt has been made to deal with the fauna of these peculiar and distinctive aquatic habitats in a comprehensive manner. The aim of this paper is to provide a fairly complete account of the invertebrate inhabitants of gnammas located in the more southern regions of Western Australia. This is a necessary basis for the future realisation of the above-mentioned potential of gnammas for facilitating ecological studies. Ecological aspects of the present study include the documentation of some salient chemical features of gnamma waters and the examination of species richness in relation to pool size. This account is based on collections made from 36 gnammas, distributed between 17 different granite outcrops, during the winter of 1990. Thirteen of the 17 outcrops were in the form of domical inselbergs while the remainder were small flat surfaces. With one exception, the gnammas were shallow, flat-bottomed pan-gnammas sensu Twidale & Corbin (1963). The exception was a deep pit-gnamma located on War Rock. Materials and Methods The maximum depth, length and width (the greatest width at right angles to the line of maximum length) were measured at each pool. Hydrogen ion concentration and conductivity of water samples collected in a polyethylene bottle were determined with a Metrohm E588 pH-meter and Radiometer CDM2e conductivity meter respectively. Invertebrates were collected with a rectangular-framed net conforming with that described by Bayly (1982), except that the mesh 167 Journal of the Royal Society of Western Australia, 80(3), September 1997 aperture was 150 pm, and preserved in either 10% formalin or 90% ethanol. The product of the maximum length and maximum width of a pool was used to approximate (consistently overestimate) the true areas of the pools. Likewise, the product of the maximum length, maximum width and maximum depth was used to approximate (consistently overestimate) the true volumes of the pools. The approximated areas and volumes of the pools were log- transformed for correlation calculations. Results Data on the location and physical dimensions of the 36 pools as well as their water chemistry are presented in Table 1. With one exception, pH was within the range 4. 6-7.9, but for 31 of the pools it was less than 7.0. With three exceptions, the conductivity (K25) was less than 1000 pS cm1, and for the 33 pools whose conductivity was less than this value the mean was 147 pS cm1. The results of taxonomic studies are summarized in Table 2. A total of 88 taxa were recorded and the mean number was 8.2 per gnamma. Branchiopod crustaceans were found in 18 pools, with Limnadia the most frequently occurring taxon. Cladocera occurred in 26 pools, with Ncothrix armata the commonest species. Copepoda were found in 23 pools, with Boeckella opacjua occurring most frequently. Ostracods occurred in all but one gnamma and were the only animals collected from four gnammas. The commonest ostracod was Ci/pretta bai/lyi. At least six new species of Cypricercus were recorded. Insects were found in 19 pools, with Anisops thienemanni the commonest species. Pools 16 and 21 yielded a disporportionately high number of insect species. Oribatid mites were restricted to two small coastal pools. There is a highly significant, positive correlation between species richness and the logarithm of approximated pool volume (Fig 1; Pearson r = 0.595, n = 36, PcO.001). If one outlier (two taxa from pool 19 with log volume 4.78) is removed from the data set, then the Table 1 Physico-chemical features of Western Australian granite rock-pools sampled June-August 1990. Locality number and name Location (lat S, long E) Sampling date Maximum depth (cm) Max. length x max. width (m2) pH Conductivity (K25 ; pS cm'1 ) 1. Coragina Rock (a) 32° 55', 123° 30' 24.vi.1990 25 16x7 8.7 91 2. Coragina Rock (b) // // " 6 3x2 6.4 59 3. Mt Madden (a) 33° 14', 119° 50' 26.vi.1990 30 30x10 6.3 74 4. Mt Madden (b) // // " 10 12x2.5 6.0 96 5. Mt Madden (c) // " 15 5x2.5 6.1 38 6. Mt Madden (d) " // " 20 7x5 6.0 61 7. Cable Beach (a) 35° 07', 117° 54' 27.vi.1990 4 1x0.3 7.1 1480 8. Cable Beach (b) // // " 2 1x0.3 7.0 1490 9. Frenchman Bay Rd (a) 35° 06', 117° 57' 27.vi.1990 2 3x2 7.3 380 10. Frenchman Bav Rd (b) // // " 5 0.8 x 0.2 6.4 457 11. Muirillup Rock (a) 34° 39’, 116°15' 2.vii.l990 12 3x1.5 4.6 93 12. Muirillup Rock (b) 34°39',116°15' " 6 2x1.5 5.5 101 13. Cape Leeuwin (a) 34°22', 115°08' 6.vii.l990 2 ’ 2.5 x 1.2 6.7 450 14. Cape Leeuwin (b) // // " 4 5x1.5 6.5 8040 15. Bilya Rock 29°00', 115°51' 12.vii.1990 11 3x2.5 6.1 90 16. War Rock (a) 29°05', 116°00' 13.vii.1990 22 25x9 5.2 55 17. War Rock (b) " // " >100 6x2 7.9 798 18. War Rock (c) // // " 12 4x2 6.1 38 19. Bunjil Rock(a) 29°39',116°21' 13.vii.1990 100 10x6 6.8 60 20. Bunjil Rock (b) " // " 9 8x5 6.1 34 21. Petrudor Rocks (a) 30°25’,116°58' 14.viii.1990 80 11x11 6.5 116 22. Petrudor Rocks (b) " " " 6 3x1.5 6.0 186 23, Petrudor Rocks (c) " « " 19 6x3 6.2 190 24. Elachbutting (a) 30°36', 118°37' 15.viii.1990 10 6x2 5.1 164 25. Elachbutting (b) « // " 19 6x5 5.7 162 26. Sanford Rock (a) 31°14', 118°46' 15.viii.1990 26 8x3 6.3 102 27. Sanford Rock (b) // // " 12 2x2 6.7 88 28. Jilbadgie Rocks (a) 31°29’,119°13’ 16.viii.1990 18 6x4 6.2 34 29. Jilbadgie Rocks (b) " " " 20 13x6 6.3 42 30. Mt Hampton (a) 31°45',119°04' 16.viii.1990 33 10x9 6.3 144 31. Mt Hampton (b) // // " 13 4x3 6.4 64 32. Yorkrakine Rock (a) 31°26', 117°31' 26.viii.1990 9 7x5 6.4 113 33. Yorkrakine Rock (b) // // " 12 4x3 6.2 146 34. King Rocks (a) 32°19',119°09' 28.viii.1990 20 14x10 6.6 105 35. King Rocks (b) " " " 30 16x8 6.4 91 36. Wave Rock 32°27',118°54' 28.viii.1990 12 9x3 6.3 113 168 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 2 List of taxa and their occurrences. Taxa Pool Total number of (see Table 1) occurrences CRUSTACEA: ANOSTRACA Branchinella longirostris Wolf 5,24,25,26,27,31 6 B. sp 34 1 CRUSTACEA: CONCHOSTRACA Cyzicus sp 4,5626 4 Limnadia sp or spp 1533/273330, 313336 10 Lynceus sp 1733 2 CRUSTACEA: CLADOCERA Chydoridae Aloua cf setuloides 3233 2 A. nsp 33 1 Alonella - excisa group 11 1 Biapertura - macropa group spl 15 1 sp2 1135 2 sp3 5,1234353639303136 9 B. rigidicaudis Smirnov 63738 3 B. nsp 9,11 2 Ephemeroporus-barroisi group 63435 4 sp 2 11 1 Monospilus diporus Smirnov & Timms 43336 3 Planicirculus alticarinatus Frey 2631 2 Plurispina chauliodus Frey 12 1 P. multituberculata Frey 2,63533 4 Rak stagnensis Frey 1,43373934 6 Leberis aenigmatosa Smirnov 27393134 4 Other families Daphnia jollyi Petkovski 23363931 4 Neothrix arrnata Gurney 13612,15,1831, 23343630343536 14 Macro thrix breviseta Smirnov 3035 2 M. hardingi Petkovski 2437 2 M. indistincta Smirnov 135 2 M. longiseta Smirnov 2 1 Ceriodaphnia sp 1363730343536 7 Moina sp 13,15,18 4 CRUSTACEA: COPEPODA Calanoida Boeckella opaqua Fairbridge 133,4333435363738, 2930313233343536 19 B. triarticulata (Thomson) 17,193133 4 Calamoecia ampulla (Searle) 21 1 Cyclopoida Mesocyclops australiensis (Sars) 21 1 Metacyclops sp 334 2 Microcyclops varicans (Sars) 131 2 Harpacticoida 11 1 CRUSTACEA: OSTRACODA Alboa n sp 16 1 Bennelongia barangaroo De Deckker 13333539 5 B. sp 331 2 Candonocypris novaezelandiac (Baird) 16 1 Cypretta baylyi McKenzie 3,4,6,11,12,16,18,20,22,23, 25,26,27,28,32,33,36 17 C. sp 2,10,15,22,25,28,31, 32,33,34,35 11 Taxa Pool Total number of (see Table 1) occurrences Cypricercus sp or spp 6,20,22,26,33,34 6 C. n sp 1 2,3 2 C. n sp 2 3,4,5,8,30 5 C. n sp 3 3,34 2 C. n sp 4 6,9,13,14,15,16 6 C. n sp 5 30 1 C. n sp 6 36 1 Heterocypris incongruens (Rahmdor) 2 1 Ilyodromus amplicolis De Deckker 2,10,18,20,23,29,31,32,33 9 1. candonites De Deckker 9,12,16,18 4 I. sp 4,15,21,24,27,36 6 Limnocy there mowbrayensis Chapman 2,3,7,34,35,36 6 L. sp 8 1 L nsp? 13 1 L. nsp 32 1 Sarscypridopsis aculaeata (Costa) 17,21,23 3 ACARI: ORIBATEI Scapheremaeus sp m 2 Trimalaconothrus sp 7 8 2 INSECTA: COLEOPTERA Dytiscidae Allodessus bistrigatus (Clark) 16,21,25 3 Eretes australis (Erichson) 21 1 Lancetes lanceolatus (Clark) 6 1 Megaporus howitti (Clark) 3,16 2 Necterosoma darwim (Babington) 11 1 Paros ter niger Watts 73 2 Sternopriscus multimaculatus (Clark) 16,25 2 Hydrophilidae Berosus approximus Fairmaire 16 1 Limnoxenus zelandicus (Broun) 6 1 INSECTA: DIPTERA Ceratopogonidae Dasyhelea sp 2,29,25 3 Chironomidae Ablabesmyia sp 28 1 Allostrissocladius sp 28,30 2 Chironomus tepperi Skuse 23 2 Dicrotendipes sp 11 1 Paraborniella sp 16 1 ?Paratendipes sp 33 1 Orthocladiinae 25,28,30 3 INSECTA: HEMIPTERA Corixidae Agraptocorixa parvipunctata (Hale) 361633 5 Diaprepocoris personata Hale 11 1 Micronecta annae Kirkaldy 163 2 M gracilis Hale 21 1 M. robusta Hale 4625/26/36 5 Sigara mullaka Lansbury 26 1 Notonectidae Anisops baylii Lansbury 5616 3 A. gratus Hale 21 1 A. hyperion Kirkaldy 3,6,173 4 A. stali Kirkaldy 173 2 A. thienemanni Lundblad 3616,1733330 8 169 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 1. Plot of species richness against the logarithm (base 10) of the estimated volume (in litres) for all 36 gnammas. There is a highly significant positive correlation between these two variables (Pearson r = 0.595, n = 36, P<0.001). The outlier at two species and log volume 4.78 is pool number 19 in Table 1. correlation is considerably improved (r = 0.696, n = 35, PcO.OOl). There is also a highly significant positive correlation between species richness (Pearson r = 0.577, n = 36, PcO.OOl) and the logarithm of approximated pool area (Pearson r = 0.641, n = 35, PcO.OOl). Discussion With seven exceptions, the conductivity (K^) of the pools (Table 1) was less than 200 pS cm1 - a finding broadly consistent with that of Bayly (1982) for a series of 19 granite rock pools in Western Australia. Six of the seven exceptions were pools with a maritime location (Cable Beach, Frenchman Bay and Cape Leeuwin) and whose high conductivities were clearly caused by the accession of ions from marine salt spray. The remaining exception was a deep pit-gnamma located on War Rock; this gnamma probably rarely loses accumulated ions by means of overflow. The pH of almost all of the inland pools is acidic, as expected for a granite substratum. The exceptionally high pH recorded for Coragina Rock (a; Table 1) was discussed by Bayly (1992) who attributed it to the use of cement mortar between the exfoliated slabs of granite used to raise the level of the pool. Although no quantitative data are available, it is noteworthy that granite rock-pools lack the high turbidity that is so characteristic of pools lying on the regolith, especially in arid regions. The total number of taxa listed in Table 2 is 88. This, however, is an underestimate of the total number of metazoan invertebrate taxa because black planarians ( Bothromesostoma ?) and nematodes were present in several collections but were not identified and are omitted from Table 2. The total of 88 taxa greatly exceeds the 18 invertebrate taxa listed by Bishop (1974) as occurring in a series of shallow pools located on sandstone at Kanangra Walls in the Blue Mountains, New South Wales. The 88 taxa from all 36 pools and the 17 taxa from the richest pool (number 21) may also be compared with the 121 invertebrate taxa reported by Lake et al. (1989) from a single temporary pond (not located on a granitic substratum) in western Victoria that was repeatedly sampled over 7 months. The animals listed in Table 2 may be classified into four groups according to their adaptations for tolerating or avoiding the dry phase following the system of Wiggins et at. (1980) as modified by Williams (1985). • Group A (Group 1 of Wiggins et al. 1980) consists of permanent residents of the pools that are capable of only passive dispersal. These organisms are dormant during the dry season and typically avoid desiccation as a resistant stage (shelled egg or embryo). All of the Crustacea listed in Table 2 should be assigned to this group, a salient feature of which is the high proportion of species endemic to Western Australia. Such endemics include Branchinella longirostris, Daphnia jollyi, Macrothrix hardingi, Boeckella opaqua, Ilyodromus amplicolis , I. candonites, Leberis aenigmatosa, Plurispina 170 Journal of the Royal Society of Western Australia, 80(3), September 1997 chaidiodns and P. multituberciilata. This list does not include the new species of Cypricercus some of which are likely to prove endemic. A distinctive negative feature of this group is the absence of notostracans. Despite their possession of a suite of adaptations (including resistant eggs) for occupancy of small temporary waters, an unknown factor excluded notostracans from granite rock pools studied here. A possible reason is that the concentration of calcium in waters on granite is so low that it is physiologically impossible for Lepidurus or Triops to fabricate their relatively large carapace. The carapaces of Ilyodromus and other ostracod taxa living in granite rock-pools contain practically no calcium; instead their carapaces are strengthened by non-calcareous ridges to compensate for this deficiency P De Deckker (pers comm). • Group B (Group II of Wiggins et al. 1980) comprises animals (typically insects) that are capable of active dispersal but which are typically still present during the dry season in a dormant form (quiescent larvae). Allotrissocladius, Paraborniella and Dasyhelea are assigned to this group. The ability of the larvae of these taxa to resist desiccation in gnammas was investigated by Jones (1971, 1975). • Group C (Group IV of Wiggins et al. 1980) consists of animals (typically insects) capable of active dispersal and which have a discontinuous presence in the pools; they avoid the dry period altogether by emigrating to permanent waters before the pools dry up. To this group are assigned all the Dystiscidae, Hydrophilidae, Corixidae and Notonectidae listed in Table 2, some 20 species in all. • Group D [Williams' (1985) replacement for Group III of Wiggins et al.] comprises animals which lack a resistant stage in their life cycle but possess exceptionally good dispersal ability thus allowing a discontinuous presence in the confines of a temporary water basin. A single species, Chironomus tepperi, may be assigned to this group. The ability of this species to rapidly colonize temporary waters has been demonstrated by Maher & Carpenter (1984). The biology of the oribatid mites, Clmdalupia meridionalis (see Bayly 1992), Scapheremaeus sp and Trimalaconothrus sp is insufficiently known to allow their assignment within the Wiggins-Williams system. Wiggins et al. (1980) mentioned the habit of burrowing into bottom sediments as a behavioural adaptation for avoiding desiccation in temporary pools. While this pattern of behaviour may provide an escape for some animals in most pools, there is little scope for it in most pan-gnarnmas because of minimal amounts of sediment and the impossibility of burrowing into granite. Deep pit-gnammas left undisturbed for long periods may accumulate a significant amount of sediment and provide an exception. Fryer (1985), on the basis of studies in England, concluded that chydorid cladocerans "show an unambiguous preference for large water bodies". However, this conclusion is of doubtful validity in the Australian context (Frey 1998). Although sampled once only, some gnammas contained four species of Chydoridae and there is presently no evidence that this number is significantly exceeded in large Australian lakes. Timms (1981), for example, sampled the littoral region of Lake Purrumbete, a large (552 ha) freshwater lake in Western Victoria, monthly for a year, but recorded only six chydorid species. The finding of a significant positive correlation between species richness and the volume of the pools is consistent with earlier investigations. Ranta (1982) reported a significant correlation between the number of water beetle species and the logarithm of pool volume in a series of 20 coastal (marine supralittoral) rock pools located in the Baltic region. A similar finding for a greater diversity of taxa was made by March & Bass (1995) for a series of six temporary pools located in Oklahoma. Acknoivledgements: I am very grateful to the following for providing taxonomic identifications; the late Professor D G Frey (Chydoridae), P De Deckker (Ostracoda}, 1 Lansbury (Corixidae and Notonectidae), M C C.eddes (Anostraca and Conchostraca), D W Morton (Cyclopoida), M Colloff (Oribatei), D H D F.dward (Chironomidae and Ceratopogonidae), J F Lawrence (Hydrophilidae) and CHS Watts (Dytiscidae). Because of the importance of ostracods in gnammas, P De Deckker's task was a particularly onerous one. 1 thank the Royal Society of Western Australia for inviting me to contribute to the Symposium and for paying my travelling expenses. I am indebted to L Dembinski for the word processing. References Bayly I A E 1982 Invertebrate fauna and ecology of temporary pools on granite outcrops in southern Western Australia. Australian Journal of Marine and Freshwater Research 33:599-606. Bayly I A E 1984 Phytotelmata: terrestrial plants as hosts for aquatic insect communities [Book review! Australian Journal of Ecology 9:164-165 Bayly I A E 1992 Freshwater havens. Landscope 7(4):49-53. Benzie J A H & Bayly I A E 1996 Male and epihippial female Daphnia jollyi Petkovski, 1973 discovered in Western Australia and the parthenogenetic female redescribed. Hydrobiologia 331:171-181. Bishop J A 1974 The fauna of temporary rain pools in eastern New South Wales. Hydrobiologia 44:319-323. Carnegie D W 1898 Spinifex and Sand. Arthur Pearson, London. De Deckker P 1981 Ostracoda from Australian inland waters - notes on taxonomy and ecology. Proceedings of the Royal Society of Victoria 93:43-85. Fairbridge W S 1945 West Australian freshwater calanoids (Copepoda). I Three new species of Boeckella, with a description of the developmental stages of B. opaqua n sp and a key to the genus. Journal of the Royal Society of Western Australia 29:25-65. Frey D G 1991 The species of Pleuroxus and of three related genera (Anomopoda, Chydoridae) in southern Australia and New Zealand. Records of the Australian Museum 43:291-372. Frey D G 1998 Expanded description of Leberis aenigmatosa Smirnov (Anomopoda: Chydoridae): further indication of the biological isolation between western and eastern Australia. Hydrobiologia (in press). Fryer G 1985 Crustacean diversity in relation to the size of water bodies: some facts and problems. Freshwater Biology 15:347-361. Jones R E 1971 The ecology of some species of Diptera on granite outcrops. PhD Thesis. University of Western Australia, Perth. Jones R E 1974 The effects of size-selective predation and environmental variation on the distribution and abundance of 171 Journal of the Royal Society of Western Australia, 80(3), September 1997 a chironomid, Paraborniella tonnoiri Freeman. Australian Journal of Zoology 22:71-89. Jutson J T 1934 The physiography (geomorphology) of Western Australia. Geological Survey of Western Australia, Perth. Bulletin 95. Lake P S, Bayly I A E & Morton D W 1989 The phenology of a temporary pond in western Victoria, Australia, with special reference to invertebrate succession. Archiv fur Hvdrobiologie 115:171-202. Lansbury I 1995 Notes on the Corixidae and Notonectidae (Hemiptera: Heteroptera) of southern Western Australia. Records of the Western Australian Museum 17:181-189. Maher M & Carpenter SM 1984 Benthic studies of waterfowl breeding habitat in south-western New South Wales. II Chironomid populations. Australian Journal of Marine and Freshwater Research 35:97-110. March F & Bass D 1995 Application of island biogeography theory to temporary pools. Tournal of Freshwater Ecology 10:83-85. Petkovski T K 1973 Zur Cladoceran-fauna Australiens II Sididae und Macrothricidae. Acta Musei Macedonici Scientiarum Naturalium 13:161-192. Ranta E 1982. Animal communities in rock pools. Annales Zoologici Fennici 19:337-347. Sheldon A L 1984 Colonization dynamics of aquatic insects. In: The Ecology of Aquatic Insects (eds V H Resh & D M Rosenberg). Praeger, New York, 401-429. Smirnov N N & Bayly I A E 1995 New records and further description of Macrothrix hardingi Petkovski (Cladocera) from granite pools in Western Australia. Journal of the Royal Society of Western Australia 78:13-14. Smith L L 1941 Weather pits in granite of the southern Piedmont. Journal of Geomorphologv 4:117-127. Timms B V 1981 Animal communities in three Victorian lakes of differing salinity. Fiydrobiologia 81:181-193. Twidale C R & Corbin E M 1963 Gnammas. Revue de Geomorphologie dynamique 14:1-20. Wallwork J A 1981 A new aquatic oribatid mite from Western Australia (Acari: Cryptostigmata: Ameronothridae) Acarologia 22:333-339. Wiggins G B, Mackay R J & Smith I M 1980 Evolutionary and ecological strategies of animals in annual temporary pools. Archiv fur Hydrobiologie, Supplement 58:97-206. Williams W D 1985 Biotic adaptations in temporary lentic waters, with special reference to those in semi-arid and and regions. Hydrobiologia 125:85-110. Wolf E 1911 Phyllopoda. In: Die Fauna Sudwest-Australiense Ergebnisse der Hamburgen siid west-australischen Forschungsreise 1905 (eds W Michaelsen & R Hartmeyer). Vol. 3. G Fischer, Jena, 251-276. 172 Journal of the Royal Society of Western Australia, 80:173-179, 1997 Aboriginal people and granite domes P R Bindon Department of Anthropology, W A Museum, Francis Street, Perth WA 6000 present address: 38 Mont Street, Yass NSW 2582 Abstract Granite domes provided Aboriginal people living on the surrounding plains with a variety of economic products. Granite domes also acted as focal points for the activities of ancestral heroes who journeyed throughout the landscape. Aboriginal religious practice includes ritual dramas which replicate the activities of these ancestral heroes at such sites. Surface geology therefore determines both the economic practices and religious activities undertaken by Aboriginal people within their territories. Introduction Granite domes are prominent landscape features common to a large part of the southern central region of Western Australia, although they do occur in other places throughout the State. We can reasonably assume that granite domes have played important roles in the settlement and continued occupation of Western Australia since humans first arrived in the area, possibly some 125,000 years ago (Science, Oct 4, 274: 33-34; West Australian, Nov 16, 1996, 34). Not least in importance was the role that granite domes played in providing water in the arid interior, a function which continues to be crucial for many Western Australian towns (Simpson 1926). Use of waters collected on rock exposures allowed grazing in Western Australian shrublands which could not be reasonably exploited until subterranean water could be used (Dimer 1989). Much of the exploration of the State was accomplished by survey teams locating waters or persuading or coercing Aboriginal people to reveal water sources found on or adjacent to granite domes. For Aboriginal people, inselbergs provided or facilitated access to a wide range of resources other than water, but water was and remains crucial to human occupation of much of Western Australia. Water Supplies Physical composition and shape of granite domes determine their water-yielding capacity. Weathered surfaces with water-holding depressions of one shape or another are common to most granite exposures. One form of natural reservoir called a 'gnamma hole' contributes its Aboriginal name to Australian English. The common addition of the English 'hole' to this phrase is redundant as explained below. Gnammas are commonly found in granites, but these kinds of holes also form in lateritic and quartz arenite mesas. Kavanagh (1984) limits his definition of gnammas to holes formed in granites and given the name ' gnamma ' © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 by desert dwelling Aboriginal people. The word 'gnamma' originates further to the west than Kavanagh's usage suggests. According to Wilkes (1978:227) who compiled a dictionary of Australian colloquialisms, the first published usage of the term is by George Fletcher Moore. In his 'Descriptive Vocabulary' of Aboriginal words from the South-west of Western Australia, Moore (1842) lists ' arnar ' and 'gnamar' as meaning "hole or pool of water in a rock." Although we are not told which dialect this word is from, we can assume from other entries in Moore's lists that the word was applied in the south-west where Nyoongar people live and also perhaps a little further eastwards in the Goldfields (Bindon & Chadwick 1992). Although the origin and development of gnammas is not known with absolute certainty, they are believed to be formed primarily by chemical weathering of less well consolidated portions of the rock (Campbell & Twidale 1995). Talbot, a geologist employed by the Western Australian Government in the early part of this century, separated gnammas into two types of holes capable of holding water (Talbot 1912). Elongated ones, which he thought probably developed along cracks and sheet joints formed his first group, and rounded ones, where globular feldspathic crystalline masses were eroded by carbonic acid produced by decomposition of vegetable matter trapped in an initial depression, made up his second group. Talbot believed that the activity of animals scratching for moisture as well as human excavation contributed to the development of gnammas. He based his conclusion on having observed soft decayed granite two centimetres thick lining a hole which he cleaned out at Day's Rock. This soft material could be removed from the surface with a shovel, but beneath this soft layer the rock was quite solid (Talbot 1912:39). We are given no estimation of how long it may have taken for this weakened material to form! The geologist Woodward thought gnammas formed through the rapid disintegration of certain coarsely crystalline pegmatite bunches in the granite, which had segregated out in the original processes of cooling of the molten mass. He says "..for it is in this class of rock [granites] that the "gnamma" holes occur, upon which in the past the aborigines [sic] and also many white explorers have had to rely for their water supply" (Woodward 1912:16). Gnammas can vary in depth from a few centimetres 173 Journal of the Royal Society of Western Australia, 80(3), September 1997 to 10 metres with a diameter up to two metres (Serventy 1973). Maintenance of these holes to maximise water retention and control quality was an important Aboriginal activity wherever the small reservoirs occurred. Mountford observed that "The Aborigines often cover these small but invaluable supplies, nama, with slabs of stone to minimise evaporation and prevent contamination by the creatures" (Mountford, 1976:42). Gnammas were considered important enough by European Australians to be marked on cadastral and geological maps until very recently During the Elder expedition through then unmapped south-eastern parts of the state, Helms made a number of observations concerning Aboriginal use of gnammas. He says "The rock-holes seem to be a special characteristic of this portion of Australia, and without them it would be impossible for the natives to exist. They are mostly found in granite, a softer mass or nodule having weathered away, thus forming natural cisterns of various shapes and dimensions. Some of them will hold many thousand gallons when filled, and as the water cannot escape by percolation the supply will last for a long time. To prevent animals getting at the water, most of the rock-holes are partly or entirely filled with loose- lying sticks, which practice, necessary as it may be to save the water, deteriorates its quality considerably by making it often look quite black and giving it a fetid smell and taste" (Helms 1892:253). Helms' observation was repeated or perhaps copied by the ethnographer Daisy Bates who says "Rock holes are found in granite hills, either at the foot of hills or on a slope, or even on the top of some of the hills. To prevent animals, birds etc., from getting at these holes, the natives sometimes fill them with sticks or branches, which not infrequently spoil the water, or give it a fetid taste, and smell" (Bates 1985: 263). Aboriginal people have indicated to me that the sticks allow animals to reach the water, drink and climb out of the hole without being stranded and dying by drowning. The sticks thus prevent contamination by animal carcasses. Kavanagh examined water supplies in arid Australia from the perspective of their use in defence (Kavanagh 1984). His assessment of gnammas was that they are generally ephemeral, lasting for just a few months depending on the rate of evaporation and exploitation. He did not think that they were useful for defence purposes. His attitude demonstrates the difference between the understanding of arid land exploitation patterns held by desert dwellers and those of itinerant desert visitors. Aboriginal people generally first used ephemeral water resources which disappear most rapidly. Claypans and other playa lakes that have great surface areas and little depth diminish quickly through evaporation. Following observation of storms in particular areas, people moved to the recently watered area. There they exploited whatever food resources they could while awaiting game attracted by new growth, and the flowers and fruits that follow a month or more after the storm passes. During their wait, they used the water in the claypans for their daily needs. As these resources diminished, the group would move back to more reliable water sources, perhaps well- shaded deep rock pools in narrow rocky valleys. Being well acquainted with the probabilities of the climate and knowing intimately all the water storages of their region, their actions and movement to new water sources were always carefully thought out with all likely possibilities considered. When the group did decide to move, their course would often involve travelling between a series of granite domes, which then became not only resource bases, but also navigational markers. Daisy Bates emphasised the importance of knowing the exact location of these life-sustaining landscape features. The necessity for having great familiarity with the resources of the land was a lesson only learnt following great hardship by many European explorers of Australia. She said "What are called 'ngamma holes' are circular hollows in a sandstone formation found principally in the Eucla division, and east of the Coolgardie goldfields district. Every ngamma hole in his district is known to the native. Many of these holes contain hundreds of gallons of water, the quantity varying with the size of the hole" (Bates 1985: 264). Bates goes on to describe how water holes occur at certain intervals across the Great Australian Bight. These waters would have allowed Eyre to traverse this region much more comfortably if he, and his Aboriginal guides, had the knowledge of their location held by those who were familiar with the countryside. A few European explorers who learned the hard way did not survive to pass on their wisdom (Maclaren 1912). Austin, who did survive, explored the State's Ashburton and Gascoyne regions and recognised the importance of gnammas. He recorded in his journal the following observations; [country] "....watered by holes in a granite rock, .... we depended on the precarious supply of rainwater accumulated in the hollows of the rocks, .... finding plenty of water here in the hollows of the rocks" (Austin 1856:236-239). His journey was possible because he and his party located these small water-holes. Until artesian sources were tapped, these uncertain resources provided the only surface water besides that flowing in the intermittent rivers following unpredictable rains. At one stage, Austin was warned by Aboriginal people not to enter the upper Murchison-Gascoyne districts except in the winter, because there was no water available except at that time. Hunting and Gathering Sites The run-off that provided surface water in gnammas, also permitted other forms of life to flourish on, but mainly around the base of, the inselbergs. This is not to deny the significance of the various plants and animals colonising the rock surface itself, but few of these were important for Aboriginal people except through the contribution they made to the life of the higher plants and larger animals usually hunted as game. Various plant species favoured the rim of rocky outcrops, exploiting the zone where run-off from the all too rare rainfall was concentrated. Two very important trees to arid land dwellers, Kurrajongs ( Brachychiton gregorii F Muell) and Quandongs ( Santalum acuminatum (R Br) D C) are commonly found around granite outcrops. They provide fruit, wood and sometimes medicinal products for Aboriginal people, but also attract emus and other bird-life. The 174 Journal of the Royal Society of Western Australia, 80(3), September 1997 medicinally important Rock Isotome, ( Isotoma petraca F Muell) and the Adjikoh or Warrain ( Dioscorca hastifolia Endl in Lehm), a staple yam species, also favour granite outcrops. If for no other reason, Aboriginal people visited granite domes to exploit these resources. The occurrence of food plants and water also attracted animals such as macropods and reptiles, many of which also contributed to Aboriginal diet. Austin observed "In many places about the country, and particularly near some of the rocks, brushwood fences are found that serve, or have served, the purpose of trapping game. These fences are about two feet high, and simply made of broken-down shrubs and branches of trees, mainly mulga, and converge to an angle after extending for a long distance over the ground" (Austin 1856:256). At the end of the fence or at the convergence of two of these, holes were dug into which fell any animals that followed the fences to a gap. In other cases nets were suspended to ensnare animals which traversed the fences to the narrowing funnel. Austin goes on to say, "Near the rocks 1 have seen them constructed in a zig-zag shape, with the self-acting trap at the apex of the angles furthest away from the rocks" (Austin 1856:256). On a number of granite outcrops in the south-west, features called 'lizard traps' can be found. These take the form of a rock plate or slab up to about a metre in diameter that is propped up along one edge by a number of other rocks so that it lies at a slant. As there is no possibility of the top rock falling and holding the lizard, we can assume that these were not true traps. However, they may be purposefully built especially to encourage sustained lizard populations on selected rock exposures by providing protective habitats. One presumes that establishing environments like this ensured the visiting hunter of a supply of animals on recurrent visits. It has been observed that when disturbed away from cover on these rock exposures, and given an opportunity, lizards or any small game run directly to the dark shelter of these slanted rocks. Regrettably, there is no evidence from ethnography confirming the function of these rock structures. However, their existence provides more than a suggestion that an early form of animal husbandry may have been in operation on these granites. During a trip by car between Perth and Albany, a now deceased Aboriginal man from the Great Southern region observed that the areas around some of the granite exposures we passed needed burning to 'clean them up'. He said that traditionally it was permissible to burn around granites quite regularly because the exposed rocks provided a refuge for animals living nearby that fled to the vegetation free area during the burn. He also observed that there was always a piece of adjoining bushland that did not burn because of the topography of the granites, so homeless animals could easily re-establish themselves. There is no easy way of verifying for how long such beliefs about the management of the environment surrounding granite domes have been held by Aboriginal people. Clearly, however, my passenger's understanding of the processes involved with managing the resources available in these environments was quite extensive. Places Frequented by Heroic Ancestral Figures Since all the members of any Aboriginal linguistic group claim to be a descendant of one or another of the ancestral beings, and since the people are living in the landscape created by these ancestors, it follows that every person is linked by their lineage to the landforms, to other living things in the same environment, and to the associated mythology. These various links dominate and to a great extent determine the actions of any individual. Mountford points out in relation to art "that, as the Aborigines' love for their country is so deep, and the myths that tell of its creation so strongly determine their lives and behaviour, these beliefs should be reflected in their art, about which little was known" (1976:55). Mountford 's argument can be applied with equal weight to many other activities including songs, stories, ritual dramas and the like. Having recognised that Aboriginal beliefs about the countryside are related to activities undertaken within that landscape by heroic ancestral figures, we can briefly examine the implications of these tenets. By re-enacting the activities of their ancestors during commemorative ceremonies, Aboriginal people re-affirm and reinforce their religious beliefs. Amongst the activities which ancestors first performed, and which modern Aboriginal groups often maintain, is the creative formative journey first taken by the ancestor figure during the establishment of the present landscape. These ancestral journeys began so long ago that they now possess the qualities of dreams. In affirming their veracity. Aboriginal people use mime, song and dance which bring the totemic ancestors to life before their human worshippers (Strehlow 1971:349). Thus, the activities of ancestral beings around granite domes which occurred during the tjukurrpa (Dreaming) are mirrored by the actions of the most recent Aboriginal groups. Numbers of granite domes were used as ceremonial areas by Aboriginal people. This is partly due to the significance these places receive from being associated with ancestral figures, but there are many other reasons why particular sites were chosen as a focus of ceremonial activity. Stone arrangements often mark these ritual places. The constructions, formed from slabs and other weathering products from the inselbergs, take the form of a 'W', are erected as a sinuous line or may be piled into a series of scattered mounds. Although the particular ceremonies carried out at these places cannot be detailed, it can be assumed that these features represent aspects of landscape and are connected with initiation procedures. As a mark of respect to their Aboriginal custodians, such places should be avoided if they are encountered, and care should be taken that they are not disturbed. Their location should be reported to the appropriate authority (AAD, Western Australian Government, 1982). Art Sites Numbers of granite domes scattered around Western Australia and in other parts of the continent contain extensive galleries of Aboriginal art. The motifs may be recognisable but the themes are often arcane (Mountford 1976). Some regions, such as the Pilbara, have art that 175 Journal of the Royal Society of Western Australia, 80(3), September 1997 seemingly has no relationship in motif or theme to any other Australian Aboriginal artistic province (Wright 1968). In this part of the state, painted art forms are comparatively rare. The various motifs are usually produced by hammering, battering or pecking away the dark patinated surface from the rock to expose the lighter coloured fresh inner core. The anthropomorphic figures produced by this method, are enigmatic in form, lyrical and dynamic in execution. They are evocative of an intense and complex motivation underlying their execution. Despite many years of study, their ultimate meaning remains elusive. However, it is possible that these human-like forms represent the larger-than-life capacities of ancestral heroes. Just south of the Kimberley region, at the northern extent of the bloc of Aboriginal cultures belonging to the western desert, a very different art genre exists. Here, circular forms composed of a number of concentric rings are joined into larger compositions using one or more parallel straight or sinuous lines. These forms seem related to the well-known art style of central Australia that depicts features in the landscape as a series of concentric circles with paths between them depicted as lines. It is difficult to avoid the idea that these motifs sometimes represent or at least include the very location at which they are found, linking that place with other localities visited by some ancestor or another. Unlike the Pilbara art, which might represent the ancestors themselves, this art represents their journeys through the landscape. Such an interpretation was provided to me by a group of Aboriginal men who were explaining motifs painted on the waiting room walls at Wirrimanu airport (Balgo Hills). It is not surprising that ancestral figures are believed to have visited the very localities that their modern heirs visit. The human aspect of ancestral behaviour means that these ancients too must hunt to eat, must drink, and in fact perform all the acts that are necessary for life. Dominating the otherwise level plains of much of inland Australia, granite domes not only form prominent navigational pointers and ritual centres, they also supply many of the daily needs of humans just as they are believed to have done for ancestral heroes. Quarries Very few portable Aboriginal stone artefacts have been found that are made of granite. The crystalline nature of this rock type makes it an unsuitable material from which to shape tools by flaking. However, some ground objects made from granite, including a few fist¬ sized pebbles used as hammers or millstones are in the archaeology collections of the Western Australian Museum. Large grinding plates of granite are also found, especially where seed-grinding contributed significantly to Aboriginal diet. Invariably, these granite grinding bases are made on an exfoliated plate¬ like piece of suitable size. Fissuring and thermoclastic weathering of the surface of granite batholiths results in the scalar detachment of successive layers of roughly circular rock plates that slide down the convex face and stack or heap as a talus at the foot of the slope. Aboriginal people most likely utilised these found objects because there is no physical evidence on batholiths or on the grinding plates that they were detached from the parent body using artificial means. Around the base of granite domes, on the gently sloping aprons, areas used for seed grinding can often be seen. These consist of a polished or smoothed surface, about 40 x by 20 cm, with a central depression one or two centimetres deep. A variety of seeds collected on the surrounding plain from grasses, shrubs and trees was ground to flour or gruel in these depressions using a fist-sized millstone. There may well have been other uses of granite domes more ephemeral than this, but evidence is lacking for these. Although structures interpreted as hunting hides or perhaps the walls of semi-permanent shelters can be found on the surfaces or in the surrounding scree slopes of granite domes in the north of Western Australia, these constructions cannot be considered as typical of Aboriginal activities on granite domes. Using loose tabular pieces from weathering processes, windbreaks can be made fairly quickly, particularly if some brushwood is incorporated into the structure. Lack of archaeological remains other than the walls in these structures hinders their exact interpretation, and the interpretations provided here reflect modern Aboriginal people's comments on the structures. Temporal Perspectives on Aboriginal Use of Granite Domes Aboriginal use of granite domes probably extends much farther back in time than archaeological investigations suggest. The concentration of resources available near these impressive landscape features was clearly of great importance to Aboriginal people, who may or may not have left physical evidence of their visits to the sites. As wre have seen, one obvious indication of Aboriginal use is the occurrence of seed grinding bases that are specially common around the aprons of granite domes. Although certain mineralised depositions may be found in these grinding bases, there is still no satisfactory method for obtaining their age. Portable seed grinding bases made of granite slabs are sometimes found in dateable contexts within archaeological excavations, but their occurrence does little more than indicate that seed grinding is an ancient activity, and does not really elucidate usage of granite domes. Despite this general lack of information, some archaeological excavations have revealed a short chronology of Aboriginal people's interest in granite domes. Walga Rock The inselberg known as Walganna or Walga Rock, located about 60 km east of Cue, is some 1.5 km long and 500 m wide. It emerges from a very flat semi-arid landscape clothed with dispersed Mulga ( Acacia aneura) woodland. Situated adjacent to a temporary water hole, a shallow wrest-facing shelter runs for more than a hundred metres on the south-west side. This shelter developed along sheet joints; the highest and deepest part evolving through haloclasticism as well as thermoclastically. The rear wall of the rock shelter is decorated with paintings in red, yellow and white pigments (Bindon et al., unpublished). 176 Journal of the Royal Society of Western Australia, 80(3), September 1997 A stratigraphic sequence established in the excavation of six square metres which reached about 3 metres in depth revealed three distinct sedimentary units. The upper unit is strongly evident of human activity and disturbed by numerous burrows. Two dates obtained from the lower part of this unit show that it covers the last millennium bp; Ly 2098 is 1 040 ± 180 bp and Ly 2087 is 790 ± 160 bp. Below this, and in-filled between the lower series and the back wall of the shelter lies the middle unit. One date relates to a median layer of this complex; Ly 2099 is 3 820 ± 200 bp. The lowest unit, the upper part of which trends toward the shelter, producing the internal segment of a complex cone of debris has two dates for a central zone; Ly 1847 is 9 950 ± 750 bp and Ly 1846 is 7 010 ± 350 bp. Sediment analysis revealed that firstly a mass movement of sediments occurred, which, in conjunction with mobilised granitic sand, covered the flanking embayments and superficial slab of the batholith by around 10 000 bp. The summit of this depositional event is marked by a discrete zone of small thermoclastically produced platelets. This is evidence for a wet phase followed by a drier period. Between about 7 000 and 4 000 bp, a series of hydrological events took place that caused in-cutting and furrowing into the back portion of the sloping bank formed by the first deposits. Erosional unconformities confirmed that periods of climatic variation occurred, during which deposition and subsequent erosion of a number of separate sediments alternated. Around 3 820 bp, the deposition of the upper friable unit began. Its sub-horizontal arrangement, and feeble development indicate a marked decrease in detrital deposition and relative climatic homogeneity that continued to about 690 bp. Evidence of human use of the shelter was found throughout the whole of the excavation sequence, giving us indications of human activity in the vicinity for the last 10 000 years. Occupation was intermittent and more or less in the same temporal pattern as delineated by other authors writing about arid inland Australia (Gould 1977; Smith 1988; Veth 1989). Periods of sparse use begin the sequence, followed by a gradual increase in visitation that culminates in an intensive occupation over the last few thousand years. At around 4 000 years ago, small delicately flaked stone tools begin to appear here just as they do around this time in many other Australian archaeological sites. A similar temporal sequence was discovered in another shelter in a granite dome close to the south coast. Cheetup Smith (1993) excavated a shelter, Cheetup, about five kilometres from the present shoreline in the Cape Le Grand National Park. This north-east facing shelter is situated on the top and northern end of a granite dome, with a commanding view over the surrounding plain. Protection is provided from cold winds and storms originating from the south and west. There is easy access to a mixture of vegetation zones that include heathlands dominated by the Proteaceae, thickets of Myrtaceae and several swamps. Excavations in the shelter revealed a complex but shallow stratigraphy about 60 cm in depth. Ten radiometric dates bracket various sedimentary events which extend beyond 13 245 ± 315 bp (GX 6605). Before this date, a pit was dug in the shelter floor. It was lined with Xanthorrhoea leaf bases and woody parts and filled with fruits of Macrozamia reidlii (Gaud) CA Gard. This particularly interesting discovery confirms ethno- historic descriptions of a food preparation technique made by early European settlers in the district. Toxins in Macrozamia fruits must be removed by leaching or fermenting and cooking before the fruits are rendered edible. Smith's (1993) discovery of this fermentation pit demonstrates that Aboriginal usage of this plant extends back almost 14 000 years into prehistoric times. Faunal and botanical remains recovered from the excavation point to a more or less stable ecosystem on the granite regardless of what occurred on the surrounding heathlands. Recherche Archipelago At the height of the last global glacial maximum around 18 000 bp, islands now forming the Recherche Archipelago were granite peaks in an extended coastal plain. This plain was some 60 km wide if we consider the coast to be located near the edge of the continental shelf at that time. Like mainland granite exposures, the islands have areas of shallow sands, soil and granite debris. Dortch & Morse (1984) located ten open sites and 30 isolated artefacts on five of the Recherche Archipelago Islands. No shelters that could be occupied with comfort were found by them. During two visits to the Recherche Group, I also was unable to locate any shelters that showed evidence of extended human occupation. Seven of Dortch and Morse's sites are on Middle Island, one on Gulch Island, and two on Stanley Island. Basing their estimations on present-day water depths surrounding the islands, they considered that Middle Island was formed between about 11 000 and 9 000 years ago, with the smaller Stanley and Gulch Islands separating from the mainland 1 000 to 2 000 years later. The ten island sites consist of scatters of between 14 and 99 stone artefacts. Apart from an infilled rockhole on Flinders Peak, Middle Island, from which 14 artefacts were retrieved, no stratified prehistoric archaeological features were discovered. Smith (1993) argued that the stone objects were discarded when the localities were the peaks of granite domes whose bases have subsequently been inundated. However, information about prehistoric aspect, vegetation association, access to resource zones, and distance to freshwater, are almost impossible to determine for these sites. Although most island sites are within 200 m of the present shoreline, this distribution pattern be partly a function of the elevation of the batholith above present sea levels and area of the exposure of the outcrop as well as factors relating to ground visibility. Obviously, with the coastline so distant for most of the time that the peaks seem to have been used, no suggestion is being made that these sites indicate some type of exploitation of the littoral. Although there is not enough evidence to decide with any certainty what criteria were used by Aboriginal people to determine the location of these sites they offer evidence of a low intensity of usage during the time when the surrounding plains were occupied by Aboriginal hunters and gatherers. However, Smith (1993) argues that the sites on the islands represent a land use system that continued on post-transgressive mainland sites. She 177 Journal of the Royal Society of Western Australia, 80(3), September 1997 bases this argument on the occurrence of scattered sites around granite domes on the mainland that contain the same kinds of stone artefacts as those found on the islands. There is also some indication of Aboriginal presence on the islands during the post-contact period. Some assemblages contain Aboriginal artefacts made of European materials, for example china. These are assumed to be "linked with the presence of European and American sealers, who seem to have had with them Tasmanian women and other Aboriginal people" (Dortch &l Morse 1984:34). One of the Stanley Island assemblages includes a tula adze of coastal chert. Tula adzes, hafted in the end of spearthrowers, continued in use until the mid twentieth century in areas to the north east of the study area, and along the coast towards South Australia. Local Aboriginal people claim to have visited these islands during the first half of the twentieth century specifically to exploit nesting birds, mainly mutton-birds and shearwaters that have rookeries on the islands. It is not unlikely that this adze may have been discarded either during the whaling and sealing period of the early nineteenth century or even more recently during raids on bird rookeries. The occurrence of Aboriginal objects manufactured from European materials indicates that, whether willingly or not, Aboriginal people continued their association with these off-shore sites. Following inundation of the surrounding plains, a hiatus imposed by lack of watercraft precluded access until relatively recently. An Earlier Occupation? Following the identification of tools manufactured from bones derived from extinct mega fauna found in sediments cleared early this century from depressions in granite domes in the Balladonia district, I proposed in conjunction with several other authors based in the Western Australian Museum that Aboriginals had inhabited the region perhaps fifty thousand years ago (Western Australian Museum Palaeontology Dept registration numbers; 65.2.80, 79.11.1, 79.11.5, 79.11.6 and 79.11.10). This proposition was based on the presumed date of the final disappearance of certain species of megafauna. Bone tools are difficult objects for archaeologists to identify because the taphonomy of the archaeological material in a site is not always unequivocal. Lacking conclusive dating we have not pressed this claim. However, detractors have not been able to completely demolish our case. What our study did help to emphasise was the importance of granite domes to both Aboriginal people and European settlers in the region. The latter group enhanced the meagre supplies these natural depressions provided by removing calcretised infills with explosives to increase water holding capacity. While their activities produced the sample of fossil bone material, it destroyed all stratigraphic context. Conclusions Aboriginal people used granite domes and their surrounds for a wide variety of purposes. Material evidence that dates the purely economic aspect of this exploitation system is lacking beyond about 15 000 years ago. Perhaps the most important use was in the realm of the ceremonial uses about which no detailed discussion is possible because of on-going ritual connotations. This situation has occasionally led to misunderstandings arising between Aboriginal people and those who wish to mine, quarry or cause other major disturbance to granite domes and their surrounds. When clear scientific evidence for use and value cannot be demonstrated, it is difficult to sustain arguments for non-disturbance. It is to be hoped that, dialogue can occur between the many groups within society who have interests in granite domes and their place in the landscape. With a broad understanding of community concerns, an understanding of the complexity of opinions about granite domes can be formulated and appropriate uses determined. Acknowledgments: I thank M Smith, for permission to use unpublished material from her research on Esperance, R Chadwick, C Dortch and M Lofgren who commented on drafts of this paper. Opinions, errors and omissions are my own. References Austin R 1856 Report by Assist.-Surveyor Robert Austin, of an Expedition to Explore the Interior of Western Australia. Colonial Office, Perth. Bates D 1985 Native Tribes of Western Australia (ed I White). National Library of Australia, Canberra. Bindon P & R Chadwick 1992 A Nyoongar Wordlist from the South-West of Western Australia. Western Australian Museum, Perth. Campbell EM&CR Twidale 1995 The various origins of minor granite Landforms. Caderno Laboratory Xeoloxico de Laxe Coruna 20:281-306. Dimer K 1989 Elsewhere Fine. Published by the author, Kalgoorlie. Dortch C E & K Morse 1984 Prehistoric stone artefacts on some offshore islands in Western Australia. Australian Archaeology 19[Dec]:31-47. Gould R A 1977 Puntutjarpa rockshelter and the Australian desert culture. Anthropological Papers of the American Museum of Natural History 54:1-187. Helms R 1895 Anthropology. Transactions of the Royal Society of South Australia xvi-2:237-332. Kavanagh B L 1984 Survival water in Australia's arid lands. Strategic and Defence Studies Centre, Research School of Pacific Studies, ANU, Canberra. Maclaren M 1912 Notes on desert water in Western Australia, Gnamma Holes and Night Wells. Geological Magazine 9:301- 304. Moore G F 1842 [ 1 978] Diary of Ten years of an early settler in Western Australia. University of Western Australia Press, Perth. [Facsimile edition in which the "Descriptive Vocabulary" is published in its most accessible form. The original appeared as a pamphlet]. Mountford C P 1976 Nomads of the Australian Desert. Rigby, Adelaide. Serventy V 1973 Desert Walkabout. Collins, Sydney. Simpson E S 1926 Problems of water supply in Western Australia. Australian Association for the Advancement of Science XVI 1 1:634-674. Smith M A 1988 The Patterning and timing of prehistoric settlement in central Australia. PhD Thesis. University of New England, Armidale. 178 Journal of the Royal Society of Western Australia, 80(3), September 1997 Smith M V 1993 Recherche a l'Esperance: A prehistory of the Esperance region of South-western Australia. PhD Thesis. University of Western Australia, Perth. Strehlow T G H 1971 Songs of Central Australia. Angus & Robertson, Sydney Talbot H W B 1912 Geological investigations in part of the North Coolgardie and East Murchison goldfields. Geological Survey of Western Australia, Perth. Bulletin 45. Veth P M 1989 The prehistory of the sandy deserts: spatial and temporal variation in settlement and subsistence behaviour within the arid zone of Australia. PhD Thesis. University of Western Australia, Perth. Woodward H P 1912 A general description. ..Yilgarn Goldfield and.. .North Coolgardie goldfield. Geological Survey of Western Australia, Perth. Bulletin 46. Wright B J 1968 Rock Art of the Pilbara Region, North-West Australia. Occasional Papers in Aboriginal Studies, 11. AIATSIS, Canberra 179 Journal of the Royal Society of Western Australia, 80:181-184, 1997 Water harvesting from granite outcrops in Western Australia I A F Laing 1 & E J Hauck 2 1 Office of Water Regulation PO Box 6740, Hay Street, East Perth WA 6892 2 Agriculture Western Australia 3 Baron-Hay Ct South Perth WA 6151 Abstract The value of granite rock outcrops for water supply was noted by Lefroy in 1863, on his expedition to the area now known as the Goldfields. Since about 1890, public water supplies have been developed using runoff from rock catchments at approximately 200 locations throughout the Western Australian wheatbelt Storage reservoirs are formed by concrete walls, reinforced concrete tanks or excavated earth tanks, which range from 500 kilolitres to 168 000 kilolitres. Runoff from rock outcrops is diverted to storage reservoirs using grouted rock or masonry walls on a slight gradient. Some measurements of runoff, and water storage and use characteristics are presented. Bores, wells and soaks have been developed in land adjacent to many granite outcrops. A review of the current situation gives the existing numbers of granite rock outcrops used for public water supply, along with the expected frequency and amount of use. The future role and value of rock catchments for wheatbelt water supply are described in the context of the Farm Water Plan. Introduction The unique hydrologic characteristics of inland south Western Australia result in a dearth of good quality surface runoff and groundwater. It is therefore not surprising that granite outcrops have been widely used for water supply development. The value of granite outcrops as sources of water was recognised by early explorers. The Government, the Water Corporation, and the farming community currently use rock catchments for water supply and this use is likely to continue. However, some rock catchment water supply facilities may be surplus to current requirements, especially in the piped water scheme area, and a rationalisation process is being implemented as part of the Western Australia Farm Water Plan. Early History Erickson ct al. (1973) refer to Lefroy, on his expedition in 1863 to the area now known as the Goldfields, having noted the value of the "bald hills" of granite for water supplies. A later example is given by Batchelor (1965) who described Holland's trek from Broomehill to Coolgardie in 1903. Holland dug a shallow well adjacent to a large granite rock on the east side of Lake Carmody, and obtained water from a gnamma at another site. The discovery of gold at Coolgardie in 1892, and at Kalgoorlie in 1893 caused the population in Western Australia to increase from 49 782 in 1891, to 184 124 in 1901, and to 282 114 in 1911. The resultant increase in demand for water, grain and hay created an immediate need for water supplies for people and livestock. Davis (1977) reported that hundreds of shallow wells were constructed as public water supplies in many of the © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 remote Goldfields areas before 1900. Whittington (1988) quoted an 1894 Mines Department report showing illustrations of a gnamma hole and a shallow well associated with a granite outcrop. Whittington (1988) also described an excavated earth dam of 6 000 kilolitre capacity which was constructed before 1896 at Woolgangie, 130 kilometres east of Southern Cross. This Government dam collected runoff from a 60 ha rock catchment. The dam was 5 m deep. At that time, the daily demand for water by travellers and the railway construction contractor was 73 kilolitres. The Goldfields water supply scheme was constructed from 1896 to 1902. Burvill (1979) reported that from 1905 the wheatbelt expanded into areas where springs and soaks of potable water occurred near granite rocks. Davis (1977) observed that hundreds of small excavated tanks were constructed in the cereal/sheep areas as agriculture began and developed. At least some of these supplies would have been adjacent to granite outcrops to take advantage of the regular runoff. Sutton (1925) described rock catchments as the best natural catchments for water harvesting, and included photos of a rock catchment in use on a farm at Koolberin, near Kulin. Types of Water Supply The simplest types of water supply associated with granite outcrops are gnammas, which are naturally- occurring reservoirs that collect runoff from the surrounding outcrop. Soaks and shallow wells are often located adjacent to granite outcrops where soil and rock formations are conducive to sub-surface water collection. Gnammas, soaks and shallow wells generally have a limited water yield and are therefore seldom used for public water supplies today. However, these sources of water were an important part of traditional Aboriginal culture. Supplies with larger water yields are more likely to involve the construction of walls, weirs or tanks. Engineered water supplies consist of a storage reservoir 181 Journal of the Royal Society of Western Australia, 80(3), September 1997 formed by constructing a concrete wall or weir in a rock valley, or by excavating an earth tank in clay subsoil adjacent to a granite outcrop. In such cases, runoff from the rock catchment is invariably collected using grouted stone or masonry walls surveyed on a slight gradient. A gradient of 1 in 200 was stated by Fernie (1930). Collecting drains on the early structures were formed from slabs of exfoliated granite rock, which were roughly shaped and cement-grouted to form walls up to 0.6 metre high. There has been an increasing use of prefabricated concrete slabs for collecting-drain construction since about the mid-1960s. Major Developments, 1921 to 1971 Fernie (1930) provided very detailed descriptions of water supply design and construction by the Government in the present central wheatbelt area. He referred in particular to supplies constructed during the 1920s at the following locations; Kondinin Rock, Egg Rock, Geelakin Rock, Karlgarin A, Glenelg Hills, Wilgoyne, Narembeen, Knungagin, Barbalin and Waddouring. A reservoir behaviour study was reported for a proposed reservoir at Gramphorne Rocks. The Public Works Department (Anon 1946) described the No. 1 District Water Supply, which consisted of 600 kilometres of pipeline, three major rock catchment sources (Waddouring, Barbalin and Knungagin), which was designed to supply water to 200 000 ha of farmland, four towns, seven railway sidings and railway requirements at two locations, one on the Mt Marshall rail loop, the other on the Dowerin to Merredin rail loop. It also described the No. 2 and No. 3 District Water Supplies which each utilised rock catchment runoff. The No. 2 District Supply distributed water through 129 kilometres of pipeline to Narembeen, and to 48 000 hectares of farmland. The No. 3 District Supply distributed water through 74 kilometres of pipeline to Kondinin, and to 17 000 hectares of farmland. The water source information for the No. 1, No. 2 and No. 3 District Supplies are summarised in Table 1. The Public Works Department (Anon 1946) also presented data relevant to water supply of some wheatbelt towns. Those with water supplies derived from rock catchments (1946 population in brackets) included Bruce Rock (600), Wagin (1200), Dangin and Quairading Table 1 Water source information for the No. 1, No. 2 and No. 3 District Supplies. WATER SOURCE GRANITE OUTCROP CATCHMENT AREA (ha) RESERVOIR STORAGE CAPACITY (kL) Barbalin 110 168 000 Waddouring 35 36 000 Knungagin 65 100 000 Narembeen 48 69 000 Kondinin 28 43 000 TOTALS 286 416 000 (500). Fernie (1930) stated that drains, channels and intake works were designed on a rainfall intensity of 25 mm per ha. Although these rates were exceeded several times in a few years, the nature of the construction has ensured that no damage has occurred to structures and surrounding land. The Public Works Department (Anon 1946) reported that following construction of the District Water Supplies in the 1920s, it was State Government policy to provide excavated earth tanks for public water supply. It was stated that 554 excavated tanks were constructed to provide for the horse traffic of the day. It was also stated that where possible, local small rock catchments were exploited to provide runoff to these excavated tanks, although it was also observed that these rock catchments were "relatively few". Davis (1977) makes reference to public water supply construction that he was directly involved in at Dingo Rock and Hyden Rock in the period 1948 to 1951. He also referred to the following supplies which were constructed in the period 1963 to 1971; Mt Madden, Karlgarin B, Gramphorne, North Cleary, Dulyalbin and Roe Dam. Davis (1977) also referred to plans having been prepared for water supply construction on another nine granite rock outcrops. Lack of funding had delayed these projects, although one of those listed, Mt Hampton was constructed in 1994/95. Value of Granite Outcrops as Sources of Water Supply Measurements of runoff have been made from the 42 ha Puntapin Rock at Wagin, and from the 26.3 ha Hyden Rock at Hyden. The runoff measurements indicate an annual rainfall-runoff relationship of 45 to 47 per cent, although from individual storms, 80 per cent runoff may occur. Fernie (1930) gave detailed design information for water yield from three classes of rock catchment. The design information was expressed as initial daily storm loss, and percentage runoff from additional daily rainfall. A daily rainfall threshold value of 1 to 5 mm is indicated by Fernie (1930), depending on rock surface character, steepness, the occurrence of rock breaks, total catchment area and time of year. Fernie (1930) stated that for daily rainfall greater than the threshold rainfall, between 60 and 90 per cent runoff could be expected, depending on the granite outcrop characteristics. The general increase in water salinity in wheatbelt farmlands tends to increase the value of the water supplies developed from granite outcrops. The water salinity of rock runoff is less than 50 mg L1 total dissolved solids. Bettenay & Hingston (1963) reported that by far the best quality (less than 1500 mg NaCl L ') and most reliable sources of groundwater are to be found near the large granite bosses in the wheatbelt. Water supplies based on granite outcrops are very dependable and the amount of stored water does not vary from year to year as much as in farm dam water harvesting systems. Many farmers therefore value these supplies as sources from which to cart water. Granite rock runoff has very low turbidity, and is therefore quite suitable for pesticide spraying operations on farms. In those areas 182 Journal of the Royal Society of Western Australia, 80(3), September 1997 where farm dam water is generally turbid, rock catchment runoff has additional value. Current Utilisation of Rock Catchments Several rock catchments are linked into the Water Corporation's piped water scheme, and the rock catchment runoff supplements the total scheme supply. The Water Corporation has estimated (Chamberlain, pers. comm., 1996) that more than 625 Agricultural Area dams and tanks exist, and that approximately 200 of these directly utilise runoff from rock catchments. The more significant supplies with strategic value to the farming community are documented by Hauck et al (1996). The water supply facilities at each site vary, along with the intensity of use. Following development of a site as a water supply. there is relatively little regular disturbance of the site by water users. In most instances, the water reservoir is located such that gravity inflow occurs from the catchment, and gravity outflow to a loading point (standpipe) occurs. This allows regular truck access to the standpipe to occur some distance from the rock outcrop, thus resulting in minimal environmental impact on the immediate surrounds of the granite outcrop. Surveys of water carting practice have shown that many farmers within close proximity of rock catchment water supplies use these supplies on a regular basis rather than as emergency water sources. The Farm Water Plan is encouraging these farmers to reduce the amount of water carted, and to be more self-sufficient on their own properties. However, the need for a well-distributed network of emergency water sources will remain a vital part of Government response plans during times of severe water deficiency. 183 Journal of the Royal Society of Western Australia, 80(3), September 1997 Farm Water Plan A Western Australian Farm Water Plan has been developed to cater for farm water supply shortages in dryland farming areas (Fig 1). The Farm Water Plan consists of six major integrated elements, one of which is the rationalisation of Agricultural Area dams and tanks (AA dams and tanks). These water supply facilities include the rock catchment water supplies. Many existing rock catchment water supplies are no longer needed for water supply purposes due to the surrounding farms now being served by a Water Corporation piped water scheme. Government agencies are currently reviewing the network of AA dams and are identifying those dams and tanks for which there is an ongoing role. Disposal of surplus AA dams and tanks will be managed jointly by the Office of Water Regulation, the Water Corporation and the Water and Rivers Commission. Facilities to be disposed of will be offered to the Department of Conservation and Land Management, Local Governments, community groups or individual farmers. The disposal arrangements will, wherever possible, preserve any existing water supply value of the facilities. The network of AA dams and tanks in southern agricultural districts from Albany to Esperance is incomplete, and there is a clear need for more emergency farm water supplies to be provided in those areas. The Office of Water Regulation is actively liaising with Local Governement in these areas to establish permanent emergency farm water supply facilities. References Anon 1946 Comprehensive Agricultural Areas and Goldfields Water Supply Scheme - A request for aid from the Commonwealth Government. Public Works Department. Government Printer, Perth. Batchelor J 1965 Hollands Track. Higher Certificate Thesis. Graylands Teachers College, Perth. Bettenay E & Hingston F J 1963 The quality of groundwaters in the central wheatbelt of Western Australia. Journal of Agriculture Western Australia 4:216-219. Burvill G H 1979 Agriculture in Western Australia, 1829 - 1979. University of Western Australia Press, Perth. Davis J E 1977 Public Works Department supplementary public water supply schemes. Journal of Agriculture Western Australia 18:73-76. Erickson R, George A S, Marchant N G & Morcombe M K 1973 Flowers and Plants of Western Australia. Reid, Sydney. Fernie N 1930 Water supplies from rock catchments in the Western Australian wheatbelt. Journal of Institution Engineers Australia 2:198-208. Hauck E J, Coles N A & Skuthorp R F 1996 Western Australian Farm Water Plan Zone Maps. Office of Water Regulation, Perth. Technical Report 1. Lefroy C E C 1934 Memoir of Henry Maxwell Lefroy, 1818- 1879. Billing & Sons, UK. Sutton G L 1925 The farm water supply. Journal of Agriculture Western Australia 2nd Series 2:264-279. Whittington V 1988 Two fevers, gold and typhoid. A social history of Western Australia, 1891-1900. University of Western Australia Press, Perth. 184 Journal of the Royal Society of Western Australia, 80:185-188, 1997 Management of granite rocks A R Main Department of Zoology, University of Western Australia, Nedlands WA 6907 Abstract This paper discusses the possible goals of management of granite rocks in terms of the values associated with them and the sensitivity of different values to use. It is suggested that they may be viewed as rock islands. Many of the contrasting values are shared between the individual rocks and the surrounds e.g. picnic sites, but such contrasting habitats are not equally tolerant of disturbance. Problems associated with management are diverse and the task of managing rocks cannot be reduced to simple recipes. This paper sets out the difficult management problems which have to be resolved because of conflicts arising from the very disparate values attributed to rocks by different groups, and concludes that adequate management will depend on public cooperation based on an informed awareness of the diverse values attributed to rocks. Introduction Granite Rocks as a title would not pass with image makers or advertising copy writers who are skilled at attracting attention. As a term widely used to identify igneous (granite) or metamorphic (gneissic) rocks, to the public, the term evokes no special image except of mere rocks, usually with a rounded form which often stands above the surrounding countryside. Thus while the term is perfectly adequate as a descriptor it does not inspire any belief among the public that granite rocks have important attributes, especially any that would warrant efforts for their conservation or management. Management implies striving for a goal, to achieve something worthwhile, such as conservation. Such a suggestion leads to expressions of disbelief and the general response is to ask what is special or worthwhile about granite rocks? The pervasiveness of this attitude is manifest in the frequent laments about the paucity of information on the terrestrial animals of granite rocks. Clearly a team with a more evocative connotation would be helpful in stimulating interest and research. Other widely used terms are available, such as monadnock or inselberg. These have quite specific technical meanings as landforms but for conservation purposes where development of a conservation ethic and public support is required for success, a less forbidding term which will indicate to the public, the potential of rocks as sites for conservation, is needed. Island hills (Jutson 1934:350), the anglicised version of inselberg, suggests that the rocks, rising as they do from the surrounding countryside, may be viewed as the analogue of islands in the sea. Looked at in this way they may be seen as having potential as sites of biological evolution or be places where relicts of former plant and animal distributions may be found. Should this be so, then granite rocks have a significance in terms of the State's conservation strategy, one of its goals being to maintain biodiversity, as living resource conservation (Anon 1987). But, there are others in the community who are aware of other values possessed by © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 these rocky island hills. Because of these different values there is the potential for conflicts which lead to the need for a method of resolving them. A standard approach to such matters is to establish management goals and plans to achieve them. Whether such an approach is feasible for such isolated entities is the main test of this paper. Management Only a small proportion of the large number of granite rocks available are designated as, or are within, conservation reserves or State Forests. The Porongorups are a very large system of rocks and tors designated as National Park and therefore have a specific management plan. However, when included within larger reservations, granite outcrops are usually too small to be mapped at the scale commonly employed when devising management plans. Similarly, those granite outcrops within State Forests are not usually large; Christensen (1992:99) when writing of the karri forest makes the point under the heading of granitic monadnocks that "Many of the more prominent occur within parks and reserves; most of the remainder are in State Forests where they are secure because of their inaccessibility and absence of commercial value". Forest working plans may exclude granite areas from hazard reduction burns. Those occurring in the developed areas of the State are subject to a variety of uses and tenures; some are designated as water supply, others vested in local authorities, or are on private land or pastoral leases. They may be adjacent to or distant from settlement. A large number are unvested on vacant crown land. There is thus no hope of implementing a management regime run by a local or resident staff of rangers controlled by a central authority such as the Department of Conservation and Land Management. Moreover they are accessible. A more diffuse plan based on education, understanding of the problems and opportunities offered and involving co¬ operation between all those involved would seem to be a more sensible goal. Ideally before management for conservation can be implemented it is necessary to know what is there, its rareness, abundance, whether it is replicated elsewhere 185 Journal of the Royal Society of Western Australia, 80(3), September 1997 and its significance. Because flowering plants are easier to observe than animals their distribution and abundance is better known but in particular the question "Why and to whom are the features significant?" needs to be addressed because this will determine the nature of the conflicts which might arise when developing management plans. Also to be considered are the implications of this knowledge for historical interpretation as well as for management. Ultimately, the foregoing information will allow a statement to be made of what is to be managed, why it is to be managed e.g. significance in terms of uniqueness of landforms, rarity of biotic elements or their manifestation in an evolutionary or distributional pattern. Finally/ it will be possible to assess how these attributes might be managed so that the elements of value for conservation are retained. Progress in achieving the above is possible once the basis of individual preference and perception of values can be established. But, granite rocks are usually so small and accessible that any management based on a system of zoning is impracticable. Perceived Values Values are established by people familiar with, and conscious of, the uses to which rock sites may be put. These include local residents, visitors, tourists and tour operators, botanists and zoologists, geologists (land forms, local water supply, quarry sites for road material and aggregate). In a general sense the usages can be classified as aesthetic and recreational (picnic sites, vantage points for viewing the countryside), cultural, (natural history e.g. displays of wildflowers in spring), heritage (historical sites, both Aboriginal and European) and utilitarian (quarry sites, water supply, slabs for facing culverts and channels for diverting run-off water into holding tanks, linings for wells in the apron soils adjacent to the rocks) and conservation. Additionally retention of aesthetic, recreational and conservation values is dependent on the maintenance of surrounding bushland from which lateritic gravel may be removed for road building material. Clearly these values are not all compatible. Consequences of Use Visitors, tourists and picnickers This category of users do not see themselves as having much impact on the attractions of granite rocks. Nevertheless their compaction of soil at picnic sites at the rock base, initiation of erosion, trampling of lichens and displacement of slabs from rock surfaces lead to slow and irreversible degradation of the site. When such activities are accompanied by the use off-road vehicles which have an immense capacity to destroy the lichens of the rock surface, furrow the small rock bound meadows and disturb and pollute rock pools, the consequences are nothing short of catastrophic. Moreover, pristine rock pools frequently had slabs or rocks on their floor; when ponds dried these became sites beneath which frogs could over-summer (Crinia, Pseudopttryne) or lay eggs before the pools filled in winter ( Pseudophryne ). Unfortunately, it seems to be common practice to gather rocks and slabs into piles or to throw them down rock slopes; either activity destroys essential frog and lizard habitat and is the anthesis of good conservation practice. However, for those who enjoy the exhilaration of climbing to the top of the hill and viewing the country to the distant horizon these changes mean nothing. Cultural Wildflower displays in spring are the principal attraction. However, people are frequently attracted to mass displays of ephemerals such as everlastings ( Rhodanthe and other composites) though some are especially attracted to orchids and other less conspicuous flowers. These flowering species vary in the intensity and size of their display depending on the season so most visitors understand the causes of the observed variability. Such is not the case with shrubs such as Kunzea, Thryptomene and other Myrtaceae which show a decline in vigour and failure to recruit. These changes are usually not remarked upon until the population approaches extinction. Even then awareness of the seriousness of the situation is only possible if the folk memory of the local population can recall what it was like in earlier times. Less conspicuous plants e.g. Isoetes or Phyllogtossum, less spectacular flowering plants and invertebrates can disappear, without their loss being noticed, as a result of mere visitor pressure arising from activities mentioned in the previous paragraph. Heritage Few are aware of the significance of rocks as Aboriginal Heritage. Sites are known with stone arrangements and stenciled hands but the myths and legends relating to rocks and gnamma holes are little known. Information regarding the significance and distribution of localities needs to be assembled before appropriate management is possible. European heritage essentially begins with the early explorers who used, and were often dependent on. Aboriginal watering points. These water holes were essential for the maintenance of populations of birds and mammals. With the discovery of the Coolgardie and later the Kalgoorlie goldfields, heavily used trails were developed to them from York and Albany. These trails are characterised by the rocks, gnamma holes and watering points along the route. Later, many of these were used by early settlers and carefully constructed stone lined circular wells attest to their importance to both travellers and settlers. There is thus much of heritage value associated with rocks. Unfortunately it is easily destroyed and readily lost. Utilitarian Depending on mineral composition and rock fabric, granite rocks may be suitable as sites for quarries providing a source of aggregate for concrete making or road building material. Rock slabs resulting from exfoliation of rock surfaces were often cemented edge-on to form shallow spiral channels around rocks so that water normally shed to the apron soils was diverted into large holding tanks and thus provided a more dependable water supply. Such constructions destroyed much cover and many shelter sites for lizards which formerly frequented them. Furthermore, the diversion of water deprived the vegetation growing on the apron and surrounding soils of soil recharge with a consequent increase in aridity of these habitats. 186 Journal of the Royal Society of Western Australia, 80(3), September 1997 Under natural conditions, water shed from rock surfaces contributed to ground water recharge, and much of this was subsequently transpired by the natural vegetation. However, in settled areas where natural vegetation is removed or reduced to mere fragments around rocks, transpiration is reduced and run-off from rock surface contributes significantly to ground water recharge and hence to rising saline water tables beneath soils of nature reserves and farms lower on the landscape. It is clearly an advantage to maintain native vegetation around granite rocks. Conservation The broad goals of conservation are to maintain biodiversity. This clearly has less tangible value than the matters discussed above. Maintenance of biodiversity can be seen as public good, the fulfilment of an ethical and moral obligation to cherish and leave an environment for the next generations that is as rich and enjoyable as the one experienced by the current generation. The values and usages enumerated in the preceding paragraphs are often incompatible with the goals of conservation. Moreover, it is often said that since conservation has no value in the market place, then it is not justifiable to use conservation value as an argument against useful developments. It is true that what nature provides as a free service is never appreciated in monetary terms until its loss has to be halted or the former service restored. In these terms one can look at the loss of picnic amenities, tourist dollars, or correction of rising saline ground waters as an approximation of the true monetary value of the so-called free services of nature. Management should address all of these issues while being aware that not only humans but feral animals such as cats, foxes and rabbits make a significant contribution to the loss of conservation values and biodiversity. Discussion Viewing granite rocks as islands suggests that the emergent rocks of the Western Australian plateau can be seen as an extensive archipelago stretching from the high rainfall South west to the arid interior. This interpretation offers special opportunities for management. In part this is because of the range of climates spanned by their distribution, but recent geological history, when arid areas and their biotas expanded and contracted (Hopper 1979) has meant that many rocks in the region of fluctuating environment now have unusual combinations of biotic elements. Such rocks are of special significance as a record of the way the biota responded to past climatic changes. Additionally the array of rock islands offers an opportunity to establish the minimum viable population size of many taxa under an array of climatic and other conditions. This is an important practical and theoretical conservation problem so an opportunity to obtain an answer relevant to local conditions should be a major consideration for management. The above discussion suggests that management will only be possible within a framework of social values, usage, cultural, geological and biogeographic history and conservation values. The complexity of the problems and conflicts along with possible management approaches are presented in Table 1. To the authoritarian mind, the complex set of interactions revealed in Table 1 can only be solved by having an enforceable set of rules accompanied by suitably punitive penalties. But such an approach implies a management presence which will be expensive. Equally it will tend to ignore special local circumstances and interests. Such an outcome leaves a disgruntled group who, more often than not, ignore the regulations. A more community centred programme based on discussions leading to the development of an understanding of the views and values held by others might be more rewarding. Conclusion Management of granite rocks will be complex and difficult. It will be impossible without a sympathetic and understanding public. Thus, before adequate management can develop, the public needs to become aware and appreciative of the values inherent in granite rocks. It is upon such an understanding that a satisfactory and enduring management can be developed based on a view of rocks as repositories of valuable information relating to Australia's biota as well as being venues for aesthetic, cultural, recreational and occasionally utilitarian values. Table 1 Values, threats, conflicts and management of granite rocks Conservation Recreation Heritage Utility Values All biota Climbing, Aboriginal, Watersupply viewing. Historical. aggregate, road picnicking. material Threats Use, trampling, Overuse of Lack of Sites have rock removal. picnic sites, awareness, other values picnic fires. soil erosion failure to recognise. Conflicts Recreational Use destroys Other uses. With all other use is harmful other values values Management Education to A caring Education, Regard for appreciate responsible awareness other values values. attitude. 187 Journal of the Royal Society of Western Australia, 80(3), September 1997 References Anon 1987 A State Conservation Strategy for Western Australia: a sense of direction. Department of Conservation and Environment, Perth, Western Australia. Bulletin 270. Christensen P 1992 The Karri Forest. Department of Conservation and Land Management, Perth, Western Australia . Hopper S D 1979 Biogeographical aspects of speciation in the south-western Australia flora. Annual Review of Ecology and Systematics 10:399-422. Jutson JT 1934 The Physiography (Geomorphology) of Western Australia. Geological Survey of Western Australia, Perth. Bulletin 95. 188 Journal of the Royal Society of Western Australia, 80:189-191, 1997 Geography, environment and flora of Mt Mulanje, Central Africa J S Beard 6 Fraser Road, Applecross W A 6153 Abstract Mt Mulanje in Malawi is the highest mountain in Central Africa, rising to 3002 m, and is one of the world's largest granite inselbergs. The mountain is formed of a series of coalescing domes of syenite, rising 2000 m above the surrounding plain. It is completely isolated and measures over 20 km at the base along each of its four sides. Almost bare slabs of rock sweep up from the base, covered with little but lichens and scattered Xerophyta plants. At about 1800 m, summits of some of the granite domes begin to provide plateau surfaces covered where the soil is deepest by the famous forests of Mulanje cedar Widdringtonia nodiflora , and where it is shallow and rocky by scrub. The latter community contains scattered individuals of a dwarf form of Widdringtonia with Philippia benguelensis and other species including Protea nyasae and P. welwitschii . Towards the mountain summits the composition of the scrub changes to dominance by Erica whyteam and £. johnstoniana with Blaeria kiwuensis. Lichens are dominant on exposed summits. The rainfall on Mulanje is said to be very high, between 2000 and 3000 mm annually, with a high incidence of cloud. The mountain is one of the most important surviving habitats for Afro-montane forest and fynbos scrub. Introduction Mt Mulanje is one of the world's most impressive granite inselbergs, a huge isolated mass of rock rising to 3002 m, which makes it the highest mountain in Central Africa. It is formed of a series of coalescing domes of syenite which stand up at least 2000 m above the surrounding plain. The mountain is completely isolated and covers an area of 640 km2 (Eastwood 1988). When one visits Ayers Rock one is told that it is the "world's largest monolith", whatever that is supposed to mean, presumably an outcrop free from stratification, joints or fissures. Mulanje also fits that description and is many times larger than Ayers Rock in length, breadth and volume. Geography and Geology The mountain is situated at 16 ° S 36 ° E, well within the tropics. It owes its existence to the tectonic movements along the great African Rift Valley, in which it is situated some 200 km south of Lake Malawi (formerly Lake Nyassa) which lies in a particularly deep and well-marked section of the Rift. The lake, 600 km long, stands at 474 m above sea level and is 695 m deep at its deepest point. The lake is drained southward by the Shire River which flows to the Zambezi. The Rift Valley south of Lake Malawi contains a few smaller lakes, e.g. Lake Malombe and Lake Chilwa, but gradually ceases to have surface expression. The horst forming Mt Mulanje has been upthrust at the southern end of the Rift. Like the famous Mt Kinabalu (4100 m) in north Borneo, it is a mass of granite thrust upward by underground pressures and probably of no very great geological age. © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 Vegetation Owing to its topographic elevation the mountain attracts a very high rainfall of between 2000 mm and 3000 mm annually, and is frequently covered by cloud. On the other hand, soils are generally shallow and patchy with much bare rock exposed and it is this factor, together with the fires that regularly sweep the mountain in the dry season, which principally determines the nature of the vegetation. Rock surfaces are never absolutely bare but are covered with lichens ( Peltula and Usnea spp) and a sedge (Coleochloa setifera ), forming grass¬ like tufts. Larger, woody plants grow in crevices. Where more soil is available these grow taller and more widespread, and eventually one particular shrub Widdringtonia nodiflora may assume tree stature and form extensive forests. The botany of Mulanje was explored by Alexander Whyte in 1891, leading to the description of a number of new and endemic species. More detail is obtainable from the recent work of Chapman (1962), Chapman & White (1970) and Porembski (1996). When I visited the mountain some years ago, an expedition was kindly laid on for me by the Forests Department of Nyasaland, who provided four Africans as guides and carriers. The inaccessibility of Mulanje is such that it is one of the few parts of the world that have not been conquered, even now, by the 4WD vehicle, and we had to make the ascent on foot. The prospect at first was daunting. From gentle lower forested slopes the mountain suddenly sweeps up in towering slabs of bare rock covered only with lichens, Coleochloa and a few woody plants mainly Xerophyta splendens (Velloziaceae) growing in crevices. The Xerophyta community has many unusual features which have been described by Porembski (1996). Fortunately a path for the ascent was available, through less abrupt, bush-covered slopes. At about the 1800 m level we reached the top of the first ascent and gained an uneven plateau affording sweeping views of the clouds covering 189 Journal of the Royal Society of Western Australia, 80(3), September 1997 the lower country around. From here too a view is available of the off-lying Chambe Peak of the main massif, embracing a plateau devoted to forestry where natural cedar forests are being worked and augmented by plantations. Timber is (or was at that time) extracted by aerial ropeway. Eventually we reached the welcome shelter of the forest resthouse, Lichenya cottage, built entirely of local cedar timber with cedar shingles. The resthouse is surrounded by cedar forest formed by the cypress-like tree Widdringtonia nodiflora which is very similar to any Australian cypress pine ( Callitris ). Just as the latter usually represent a relict vegetation growing in fire- protected places, so Widdringtonia trees have evidently been more abundant in the past in more pluvial times. As wTith Callitris, e.g. C.roei in Western Australian kwongan, there has been a speciation into shrub-sized species which have become components of the South African fynbos. The trees on Mulanje are heavily lichened as a result of the damp cloudy atmosphere. None-the-less, fires may often occur in dry periods in spite of the high rainfall, and the forest can be seen to be very patchy. The summit of the mountain was in plain view from the resthouse, but having taken most of the day to walk up there, enough was enough for the present; I spent the evening in front of a cheerful log fire of cedar wood. Next morning, unfortunately and to my consternation, the clouds were down and we could see nothing. It was useless to try to find one's way to the summit under these conditions, so my African attendants and I padded around in the fog and the drizzle doing a little botanising. The following day wTas still cloudy but much better, and so it was decided to strike for the summit. Shallow rocky places above the forest are covered by a scrub similar to the fynbos of the Cape mountains containing a dwarf form of the Widdringtonia with Philippia benguelensis and other species of Ericaceae, and an endemic protea, P. nyasae, which I was particularly looking for at the time. Higher up the composition of the scrub changes to dominance by Erica spp, chiefly E. ivhyteana and E. johnstoniana , with Blaeria kiwuensis. These communities with local endemic species are typical of the so-called ericaceous zone of high East African mountains (Hedberg 1951). Some ericas are mat-forming species, while others are substantial shrubs. Keeping the top of the mountain in sight, we luckily discovered a little track that had been beaconed, as I discovered later, by the Mountain Club of Nyasaland, and which led to the top. I was glad of this, as I was nervous that the clouds might come down again and make it difficult for us to find our way back. Luckily this did not happen, but one can understand that in a topography of granite domes to strike off straight down hill may lead one into a situation of even increasing steepness and can be dangerous. We returned safely and in clear weather. The summit has very little vegetation except lichens. Biogeography From the point of view of biogeography, the flora of East African mountains is interesting for its close relationship to that of the Cape mountains in South Africa which are well known to feature a shrubland formation known locally as fynbos (Cowling 1992) in which woody species belonging to the families Ericaceae, Fabaceae and Proteaceae are dominant, with many herbaceous perennials in the ground layer, especially Restionaceae. Fynbos is in many ways similar to the kwongan of south-western Australia (Pate & Beard 1984), the principal difference floristically being that Ericaceae are replaced by Myrtaceae in Australia. As with kwongan, fynbos typifies a particular phytogeographic region known at one time as the Cape Floral Kingdom, nowadays preferably as the Cape Floristic Region. Climatically this is subject to a winter rainfall maximum in the west, merging into a non- seasonal regime in the east. Beyond the eastern end of the Cape Floristic Region the rainfall pattern changes to summer maximum which encourages the growth of grass, so that annual fires become a feature of the environment and grasslands are predominant. Within the fynbos region where grasses are insignificant fires occur less frequently and the shrubland can regenerate free from grass competition. In the summer rainfall area we have evidence from relictual patches that fynbos is properly the climax vegetation in the mountains but has been largely eliminated by burning (Beard 1993). Such relics continue to occur all along the mountain chains and escarpments of eastern Africa at increasing altitude right up to the Equator and beyond into Ethiopia. Another feature of this situation is that small scattered populations of Widdringtonia are also found, the best known being in the Cedarberg in the western Cape mountains. These follow the fynbos relics as far north as Mulanje, beyond which they are replaced by Junipenis procera as the ecological equivalent. The fossil pollen record shows widespread abundance of Widdringtonia in the past and it is reasonable to assume that it represents a prior climax vegetation which preceded fynbos on poor acid soils in these mountain situations. Widdringtonia is very sensitive to fire and it can only have been dominant when fynbos fires were less frequent and destructive, perhaps w'hen rainfall generally was much higher, and also when humans - particularly cattle-keeping humans - were less numerous. A parallel situation exists in Western Australia where small relict populations of Callitris exist in fire-protected places - on coastal islands which were not reached by aborigines, or on cliffs and rocky places in the interior where there is insufficient grass to carry fire (Beard 1990). In this context Mt Mulanje is the most important refuge for the survival of Widdringtonia populations. References Beard J S 1990 Plant Life of Western Australia. Kangaroo Press, Sydney. Beard J S 1993 The Proteas of Tropical Africa. Kangaroo Press, Sydney. Brass L J 1963 Vegetation of Nyasaland. Report on the Vernay Nyasaland Expedition of 1946. Memoirs of the New York Botanical Garden 8:161-190. Chapman J D 1962 The Vegetation of the Mulanje mountains, Nyasaland. Government Printer, Zomba. 190 Journal of the Royal Society of Western Australia, 80(3), September 1997 Chapman J D & White F 1970 The Evergreen Forests of Malawi. Commonwealth Forestry Institute, Oxford. Cowling R M 1992 The Ecology of Fynbos - Nutrients, Fire and Diversity. Oxford University Press, Cape Town. Eastwood F 1988 Guide to the Mulanje Massif. Lorton Communications, Johannesburg. Hedberg O 1951 Vegetation belts of the East African mountains. Svensk BotaniskTidskrift 45:140-202. Pate J S & Beard J S 1984 Kwongan, Plant Life of the Sandplain. University of Western Australia Press, Perth. Porembski S 1996 Notes on the vegetation of inselbergs in Malawi. Flora 191:1-8. 191 Journal of the Royal Society of Western Australia, 80:193-199, 1997 Inselberg vegetation and the biodiversity of granite outcrops S Porembski, R Seine & W Barthlott Botanisches Institut der Universitat, Meckenheimer Allee 170, D-53115 Bonn, Germany email: unbll2@oni-bonn.de Abstract Granite inselbergs occur as mostly dome-shaped rock outcrops in all climatic and vegetational zones of the tropics. Consisting of Precambrian rocks, they form ancient and stable landscape elements. Due to harsh edaphic and microclimatic conditions, the vegetation of inselbergs differs markedly from those of the surroundings. Well defined inselberg habitats (e.g. cryptogamic crusts, rock pools, monocotyledonous mats, ephemeral flush vegetation) can be distinguished based on physiognomy. Plant diversity of inselbergs is influenced by both deterministic processes and stochastic environmental disturbances. The latter promote higher species richness due to the prevention of competitive exclusion. Considerable regional differences in floristic composition, life forms and species diversity exist concerning both the vegetation of whole inselbergs as well as those of individual habitats. Introduction Introduced by the German geologist Bornhardt (1900), the term "inselberg" has achieved general acceptance in international literature. As solitary, usually monolithic mountains or groups of mountains, they rise abruptly from the surrounding plains (Fig 1). Consisting of Precambrian granites and gneisses, inselbergs are old landscape elements that may possess an age of more than 50 million years. Bremer & Jennings (1978) and Thomas (1994) provide detailed surveys of their geomorphology. Inselbergs are widely distributed on the old crystalline shields and occur particularly in tropical and subtropical regions but can also be found in temperate zones (e.g. southeastern USA, southwestern Australia). Due to harsh edaphic (i.e. more or less devoid of soil cover) and microclimatic (i.e. high degree of insolation and evaporation rates) conditions, the vegetation of inselbergs differs markedly from that of the surroundings. In contrast to their temperate counterparts (for surveys of literature, see: Australia; Ornduff 1987; Hopper 1992: USA; Quarterman el al. 1993), tropical inselbergs have not yet attracted many biologists. In North America and Australia the interest in rock outcrop vegetation is well established and consequently there has been an impressive number of ecological studies, including reproductive ecology (e.g. Hopper 1981; Wyatt 1983), competition (e.g. Sharitz & McCormick 1973; Ware 1991) and speciation (Hopper & Burgman 1983; Moran & Hopper 1983). Apart from regional, rather descriptive studies for tropical inselbergs (e.g. Adjanohoun 1964; Bonardi 1966; Fleischmann et al. 1996; Granville 1978; Hambler 1964; Ibisch et al. 1995; Porembski 1995; Porembski el al. 1994, 1996a; Reitsma et al. 1992; Richards 1957; Sarthou 1992; Villiers 1981), comparative analyses of their vegetation are rarely available. Since 1991, the vegetation of African and South American inselbergs has been the focal point of research at the Botanical Institute, University of Bonn. Within the © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 frame of this project, primary emphasis has been both upon a comparative floristic analysis of tropical inselbergs and on the identification of ecological factors responsible for regulating the species richness of inselberg plant communities. The objective of the present paper is to provide a general overview of plant communities on tropical inselbergs in combination with some comments on the relation between regional floristic richness and species numbers on inselbergs. Flora and Vegetation of Inselbergs Floristically, inselbergs in different geographical regions are clearly distinct. Apart from families that are of certain importance in regard to species number on inselbergs throughout the tropics (e.g. Poaceae, Cyperaceae, Rubiaceae), there are also region-specific families. For details on temperate inselbergs see Ornduff (1987), Hopper (1992) and Quarterman et al. (1993). • African inselbergs: Fabaceae, Scrophulariaceae and Lentibulariaceae belong to the most species- rich families. The percentage of endemics is comparatively low. Therophytes are the predominant life form. Floristic differences are low on a local scale (low p-diversity). • South American inselbergs: Melastomataceae, Orchidaceae, Cactaceae and Bromeliaceae are the most species-rich and characteristic families. The percentage of endemics is high. Phanerophytes are the predominant life form, whereas therophytes are far less important. Local floristic differences vary greatly (high (3-diversity). • Tropical Asian inselbergs: These are particularly well represented on the Indian subcontinent (Krebs 1942). However, information about the floristic composition of their vegetation is very sparse (Bharucha & Ansari 1962; Willis 1906). From consideration of these works and of several local floras (e.g. Matthews 1991), it can be assumed that the vegetation of Indian and Ceylonese inselbergs is close to their African counterparts at the family and genus level. The vegetation of granite outcrops harbours a high percentage of highly adapted functional plant groups. 193 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 1. Dome-shaped granite inselberg in Brazil (Minas Gerais, near Camponario). Above all, poikilohydric vascular plants (i.e. "resurrection plants", providing many examples of convergently developed taxa) are exceptionally well represented. Additional quite typical plant groups are succulents (Barthlott & Porembski 1996) and carnivorous plants (Seine et al 1995). Inselbergs form isolated insular ecosystems that host a vegetation consisting of physiognomically defined and well delimited habitats. Most characteristic are cryptogamic crusts, seasonally water-filled rock pools, monocotyledonous mats, ephemeral flush vegetation and wet flush vegetation. In the following, short characteristics of these habitats are given. A more detailed description is in preparation. Cryptogamic crusts Exposed rock surfaces are almost completely covered by either specialized cyanobacterial lichens (typically Peltula spp) or cyanobacteria (frequently Stigonema spp and Scytoncma spp; Budel et al. 1994), which are responsible for the characteristic brownish or greyish colour of inselbergs. In seasonally dry, savanna regions, cyanobacterial lichens dominate on granite outcrops, whereas under rain forest climates the rocky slopes are commonly covered by cyanobacteria. Particularly where seepage water is available, moss cushions (in West Africa, frequently Bryum arachnoideum) may establish (Frahm & Porembski 1994). Seasonally water-filled rock pools These pools, water-filled after rainfall, soon dry out if not replenished by subsequent rain, and form temporal habitats. Short-lived herbs predominate. Besides species which are otherwise widespread on marshy ground (e.g. Cypems spp, Ludwigia spp), there are specialists which are restricted to this habitat both on tropical and extra- tropical rock outcrops. Prominent examples are richly represented within the Scrophulariaceae, such as the Namibian Chamaegigas intrepid us, Amphianthus pusillus (southeastern USA) and species belonging to the genera Lindernia (e.g. L. nwnroi and L. conferta, Zimbabwe), Dopatrium (e.g. D. longidens, West Africa; Fig 2), Glossostigma (e.g. G. drummondii, Australia). Interestingly, there is a high percentage of poikilohydric species amongst these. Widespread on East African inselbergs are geophytic waterplants of the genus Aponogeton (e.g. A. stuhlmannii, Zimbabwe). Characteristic but frequently overlooked are geophytic Isoetes species (e.g. 1. nigritiana, West Africa) that have also been recorded from extra- tropical inselbergs, like I. melanospora (Georgia, USA) and I. australis (Australia). These terrestrial species can be considered to be Gondwanan elements which have only 194 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 2. The Scrophulariaceae Dopatrium longidens is a characteristic colonizer of seasonally water-filled rock pools on West African granite outcrops. Photograph by Nadja Biedinger. little or no long-range dispersal ability (Taylor & Hickey 1992). Monocotyledonous mats Soil cover is usually absent. However, a substrate mainly consisting of decaying plant material is present. Carpet-like mats formed by Bromeliaceae, Cyperaceae and Velloziaceae cover even steep rocky slopes (Fig 3). In tropical Africa and Madagascar, Cyperaceae dominate ( Afrotrilepis in West Africa; Coleochloa in East Africa and Madagascar) and Velloziaceae (Xerophyta) sometimes attain the status of co-dominants. On neotropical inselbergs, Bromeliaceae (e.g. Pitcairnia spp, Dyckia spp, Vricsca spp) and Velloziaceae dominate, whereas Cyperaceae (Trilepis spp) are of minor importance. On granite outcrops in southwestern Australia, several poikilohydric species of the genus Borya (e.g. B. constricta, B. sphaerocephala) form dense mat-like stands on exposed slopes. Mat-forming Cyperaceae and Velloziaceae possess convergently developed morphological (treelet- like habit), anatomical (roots possessing a velamen radicum; Porembski & Barthlott 1995), and physiological (poikilohydry) adaptations in order to withstand the harsh ecological conditions on inselbergs. Remarkably, most mat-forming species host a highly specific set of epiphytic orchids (e.g. Polystachya microbambusa on Afrotrilepis pilosa in West Africa). Apart from monocotyledons, only a few other groups of vascular plants occur as mat formers on inselbergs, for example the poikilohydrous shrub Myrothamnus ( M . flabellifolia in East Africa; M. moschata in Madagascar) and a number of likewise poikilohydric species of the fern genus Selaginella (e.g. S. nicwmiamensiSj S. dregei in East Africa; S. convohita , S. sellowii in Brazil). Ephemeral flush vegetation This term was introduced by Richards (1957) and denotes a vegetation type developing at the base of steep slopes over thin soil where water continuously seeps during the rainy season. Poaceae and Cyperaceae make up the largest part of the phytomass. Most striking are tiny ephemerals with members of Eriocaulaceae, Xyridaceae, Burmanniaceae and carnivorous plants (Fig 4; Droseraceae and Lentibulariaceae; Utricularia, Gmlisea; Seine et al. 1995). In tropical Africa, this community is especially well developed (i.e. most rich in species) on inselbergs situated in savanna zones (Dorrstock et al. 1996). Wet flush vegetation This occurs on inclined, bare rocky slopes where water flows continuously during the rainy season. Typical are small-sized annuals, in particular Xyris spp and Utricularia spp, which are attached to cyanobacterial 195 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 3. Mat-forming monocotyledons (here a bromeliad Encholirium sp) are a typical element of the vegetation of tropical inselbergs. Photograph by Nadja Biedinger. crusts. Occasionally, small patches of mosses can be found which provide establishment sites for other vascular plants. This community develops best under humid tropical climates. Despite fundamental floristic differences, the physiognomy of plant communities on inselbergs remains largely unchanged throughout the tropics. Inselbergs located in temperate regions are very similar to their tropical counterparts in regard to habitat composition. However, a major difference is evident in the widespread occurrence of monocotyledonous mats on tropical outcrops and their near absence from temperate inselbergs. The reason for this distinction is not clear yet. Presumably, climatic conditions (i.e. low temperatures) are responsible for the absence of slightly succulent (e.g Bromeliaceae) or poikilohydric monocotyledons. Species Diversity of Plant Communities on Inselbergs Traditionally over the past, islands have played a crucial role in ecological studies designed to achieve a better understanding of those factors influencing the species richness of habitat fragments. Though the bulk of scientific interest was directed to oceanic islands, naturally occurring continental island biotas may provide even better opportunities (e.g. because of less anthropogenic interference) for such studies. Due to their worldwide distribution, granitic outcrops offer excellent venues for comparative phytogeographical studies as well as for research on the controlling factors of species richness in isolated plant communities. Within the frame of this paper, consequences of abiotic influences and implications of biotic interactions for the diversity of the vegetation of inselbergs will be examined (for more details see Porembski et al. 1995, Porembski et al. 1996b). Seasonality promotes higher species diversity Observations on inselbergs in the Ivory Coast have revealed considerable habitat-specific differences in both alpha (i.e. the number of species) and (3-diversity (i.e. the degree of change in species diversity along a transect or between habitats; Magurran 1988) between Afrotrilepis mats extending over large areas of rock and ephemeral flush communities which usually cover only a few square meters. In contrast to the almost monospecific monocot mats, which are dominated by the highly competitive Cyperaceae Afrotrilepis pilosa (with stems more than 1 m high, attaining an age of several hundred years), ephemeral flush communities may harbour several 196 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 4. Tiny annuals and carnivorous species ( Drosera indica) are the characteristic components of ephemeral flush communities on tropical inselbergs (Benin, West Africa). Photograph by Nadja Biedinger. dozens of tiny ephemerals. Each year, with the onset of the first rains in the rainy season, the component species of the latter community must establish themselves anew, leaving much room for stochastic events. It is this abiotically driven dynamic between dry and rainy season that may prevent competitive exclusion, and therefore guarantees the persistence of a highly diverse plant community. The bulk of ephemeral flush species is very small (i.e. less than 15 cm in height), and most species are generally considered as weak competitors ( e.g . the carnivorous plants). In contrast, the species-poor Afrotrilepis mats are in a state of equilibrium with Afrotrilepis pilosa which is the clearly dominating element, leaving few opportunities for the establishment of additional less competitive species. Species diversity not only varies between different inselberg habitats, but there is also a considerable degree of within-habitat variation in species diversity along ecological gradients. For most habitats on Ivorian inselbergs, a decline in species diversity along a gradient from the seasonally dry savanna region towards the rainforest zone was observed that is in marked contrast with the diversity of the surrounding vegetation. Again, it is possible to conclude that climatic seasonality favors the maintenance of species-rich plant communities on inselbergs by preventing competitive exclusion. The declining diversity of inselberg vegetation in the Ivory Coast from savanna towards rainforest is probably enforced by increasing isolation of rock outcrops in the latter region. Granite outcrops are less frequent in the rainforest zone which, also lacks further azonal habitats such as ferricretes. In the savanna zone, the higher number of inselbergs and ecologically similar sites may serve to reduce extinction rates (via metapopulation dynamics) and thus promotes more species-rich communities. Regional differences in plant species-richness on inselbergs Studies on inselberg vegetation in different parts of the tropics showed large regional variations, both floristically and in regard to plant species-richness. The exact reasons for the pronounced differences in local plant species diversity on inselbergs situated, for example, in the Upper Guinea region of West Africa (low diversity) and in the Brazilian Atlantic rainforest (high diversity) have not been analyzed in detail yet. However, we assume that this difference is a consequence of the higher regional species diversity of the latter region (one of the global centers of biodiversity), resulting in a greater number of potential colonizers of inselberg habitats. This is illustrated by the considerably higher number of mat- 197 Journal of the Royal Society of Western Australia, 80(3), September 1997 forming species on Brazilian inselbergs (more than a dozen species of Bromeliaceae, Velloziaceae, Cvperaceae; unpublished data) compared to the more or less monospecific Afrotrilepis mats on West African inselbergs. On the local scale the species richness of the vegetation of inselbergs is probably also influenced by processes, such as source-sink effects and metapopulation dynamics which both affect extinction rates on rock outcrops (Porembski et al. 1996a). It has to be the aim of future comparative studies to explain existing floristic differentiations and the variations in plant species diversity between inselbergs located in geographically distinct areas. Because they are relatively uniform in geology, granitic and gneissic inselbergs offer a unique opportunity for the search for general determinants of species diversity in plant communities. Acknoivlcdgements: Financial support by the Deutsche Forschungsgemeinschaft (DFG-SPP "Mechanismen der Aufrcchterhaltung tropischer Diversitat") is gratefully acknowledged. The authors would also like to thank our Australian colleagues and, in particular, SD Hopper (Perth, Kings Park and Botanic Gardens) for support during fieldwork in Australia. References Adjanohoun E 1964 Vegetation des savanes et des rochers decouverts en Cote d'Ivoire centrale. Memoires ORSTOM 7:1-178, Barthlott W & Porembski S 1996 Biodiversity of arid islands in tropical Africa: the vegetation of inselbergs. In: The Biodiversity of African Plants (eds L J G van der Maesen, X M van der Burgt & J M van Medenbach de Roov). Proc. XIVth AETFAT Congress, Kluwer Academic Publishers, Dordrecht, 49-57. Bharucha F R & Ansari M Y 1962 Studies on the plant associations of slopes and screes of the Western Ghats, India. Vegetatio 11:141-154. Bonardi D 1966 Contribution a 1 'etude botanique des inselbergs de Cote d'Ivoire forestiere. Diplome d etudes superieures de sciences biologiques. University d 'Abidjan, Ivory Coast. Bornhardt W 1900 Zur Oberflachengestaltung und Geologie Deutsch-Ostafrikas, Reimer, Berlin. Bremer H & Jennings J (eds) 1978 Inselbergs/Inselberge. Zeitschrift fur Geomorphologie, N F, Supplement 31. Budel B, Luttge U, Stelzer R, Huber O & Medina E 1994 Cyanobacteria of rocks and soils of the Orinoco Lowlands and the Guayana Uplands, Venezuela. Botanica Acta 107:422-431. Ddrrstock S, Porembski S & Barthlott W 1996 Ephemeral flush vegetation on inselbergs in the Ivory Coast (West Africa). Candollea 51:407-419. Fleischmann K, Porembski S, Biedinger N & Barthlott W 1996 Inselbergs in the sea: vegetation of granite outcrops of Mahe and Silhouette, Seychelles. Bulletin of the Geobotanical Institute ETH 62:61-74. Frahm ] P & Porembski S 1994 Moose von Inselbergen aus Westatrika. Tropical Bryology 9:59-68. Granville de J J 1978 Recherches sur la flore et la vegetation guvanaises. These de Doctorat d'Etat, USTL, Montpellier. Hambler D J 1964 The vegetation of granitic outcrops in Western Nigeria. Journal of Ecology 52:573-594. Hopper S D 1981 Honeyeaters and their winter food plants on granite rocks in the central wheatbelt of Western Australia. Australian Wildlife Research 8:187-197. Hopper S D 1992 Patterns of plant diversity at the population and species level in south-west Australian mediterranean ecosystems. In: Biodiversity of Mediterranean Ecosystems in Australia (ed R J Hobbs). Surrey Beatty & Sons, 27-46. Hopper S D & Burgman M A 1983 Cladistic and phenetic analyses of phylogenetic relationships among populations of Eucalyptus cacsia. Australian Journal of Botany 31:35-49. Ibisch P L, Rauer G, Rudolph D & Barthlott W 1995 Floristic, biogeographical, and vegetational aspects of Pre-cambrian rock outcrops (inselbergs) in eastern Bolivia. Flora 190:299- 314. Krebs N 1942 Uber Wesen und Verbreitung der tropischen Inselberge. Abh. PreuR Akad. Wiss., Math.-nat.wiss. Kl. Nr. 6. Magurran A E 1988 Ecological Diversity and its Measurement. Chapman & Hall, London. Matthew K M 1991 An Excursion Flora of Central Tamilnadu, India. Oxford & IBH Publishing, New Delhi. Moran G F & Hopper S D 1983 Genetic diversity and insular population structure of the rare granite rock species Eucalyptus cacsia Benth. Australian Journal of Botany 31:161- 172. ‘ Ornduff R 1987 Islands on Islands: Plant Life on the Granite Outcrops of Western Australia. H L Lyon Arboretum Lecture 15. University Press of Hawaii, Honolulu. Porembski, S & Barthlott W 1993 Okogeographische differenzierung und Diversitat der Vegetation von Inselbergen in der Elfenbeinkiiste. In: Animal-plant Interactions in Tropical Environments (eds W Barthlott, C M Naumann, K Schmidt-Loske & K L Schuchmann). Annual Meeting of the German Society for Tropical Ecology, Bonn 1992, 149-158. Porembski S, Barthlott W, Ddrrstock S & Biedinger N 1994 Vegetation of rock outcrops in Guinea: granite inselbergs, sandstone table mountains, and ferricretes - remarks on species numbers and endemism. Flora 189:315-326. Porembski S & Brown G 1995 The vegetation of inselbergs in the Comoe-National Park (Ivory Coast). Candollea 50:351- 365. Porembski S & Barthlott W 1995 On the occurrence of a velamen radicum in tree-like Cvperaceae and Velloziaceae. Nordic Journal of Botany 15:625-629. Porembski S, Brown G & Barthlott W 1995 An inverted latitudinal gradient of plant diversity in shallow depressions on Ivorian inselbergs. Vegetatio 117:151-163. Porembski S 1996 Notes on the vegetation of inselbergs in Malawi. Flora 191:1-8. Porembski S, Szarzynski J, Mund J P & Barthlott W 1996a Biodiversity and vegetation of small-sized inselbergs in a West African rain forest (Tai, Ivory Coast). Journal of Biogeography 23: 47-55. Porembski S, Brown G & Barthlott W 1996b A species-poor tropical sedge community: Afrotrilepis pilosa mats on inselbergs in West Africa. Nordic Journal of Botany 16:239- 245. Quarterm an E, Burbanck, M P & Shure, D J 1993 Rock outcrop communities: Limestone, Sandstone, and Granite. In: Biodiversity of the Southeastern United States (eds W H Martins, S C. Boyce & A C Echternach). John Wiley & Sons, New York, 35-85. Reitsma J M, Louis A M & Floret J J 1992 Flore et vegetation des inselbergs et dalles rocheuses: premiere etude au Gabon. Bulletin Museum National d'Histoire Naturelle B, Adansonia 14:73-97. Richards P W 1957 Ecological notes on West African Vegetation 1. The plant communities of the Idanre Hills, Nigeria. Journal of Ecology 45:563-5 77. Sarthou C 1992 Dynamique de la vegetation pionniere sur un inselberg en Guyane Fran^aise. These de Doctorat d'Etat, Universite Paris 6, Paris. 198 Journal of the Royal Society of Western Australia, 80(3), September 1997 Seine R, Porembski S & Barthlott W 1995 A neglected habitat of carnivorous plants: inselbergs. Feddes Repertorium 106:555- 562. Sharitz R R & McCormick J F 1973 Population dynamics of two competing annual plant species. Ecology 54:723-740. Taylor W C & Hickey R J 1992 Habitat, evolution, and speciation in Isoetes. Annals of the Missouri Botanical Gardens 79:613-622 Thomas M F 1994 Geomorphology in the tropics. A Study of Weathering and Denudation in Low Latitudes. John Wiley & Sons, New York. Villiers J F 1981 Formations climaciques et relictuelles d'un inselberg inclus dans la foret dense camerounaise. These de Doctorat d'Etat, Universite Paris 6, Paris. Ware S 1991 Influence of interspecific competition, light and moisture levels on growth of rock outcrop Talinum (Portulacaceae). Bulletin of the Torrey Botanical Colub 118:1-5. Willis J C 1906 The Flora of Ritigala, an isolated mountain in the North-Central Province of Ceylon; a study in Endemism. Annals of the Royal Botanic Gardens, Peradeniya 8:271-302. Wyatt, R 1983 Reproductive biology of the granite outcrop endemic Sedum pusillum (Crassulaceae). Systematic Botany 8:24-28. 199 Journal of the Royal Society of Western Australia, 80:201-208, 1997 Remnant vegetation, priority flora and weed invasions at Yilliminning Rock, Narrogin, Western Australia J P Pigott & L VV Sage Department of Conservation and Land Management, Science and Information Division, WA Herbarium, Locked Bag 104, Bentley Delivery Centre WA 6983 email: patrickp@calm.wa.gov.au Abstract Remnant vegetation on and around Yilliminning Rock, a granite inselberg near Narrogin, Western Australia, was surveyed several times between 1992 and 1997. Six plant communities comprising 238 vascular species from 54 families are reported. These include 2 undescribed species and 5 priority species; many of these populations are reported for the first time. In addition to the diverse vascular flora, 33 species of lichens from 11 families, including 2 priority species, are recorded for Yilliminning Rock. Of 19 recorded introduced vascular species, several are serious environmental weeds. The weed threat to populations of priority species and other management issues of concern are discussed. Introduction Yilliminning Rock is located in the Shire of Narrogin, about 230 km south-east of Perth, on the western edge of the Western Australian wheatbelt (Fig 1). The rock is a granite inselberg, rising to approximately 48 m, with commanding views across the farmland and remnant vegetation of the surrounding agroecosystem. Yilliminning Rock is a prominent feature in an undulating transitional landscape; less dissected and with fewer large granite outcrops than the Pingelly district to the north (Beard 1980). This contrasts the flat landscape to the southeast that includes the extensive wetlands of the Northern Arthur River System. Beard (1980) describes the vegetation type around Yilliminning Rock as York gum ( Eucalyptus loxophleba) and wandoo (E. wandoo) woodlands with some heath vegetation dominated by Dryandra species. However, this is more typified by remnant vegetation at Birdwhistle Rock about 20 km to the north-east. Here there is also rock sheoak ( Allocasnarina huegeliana) forest, but with less prominent granite outcropping and a more visible disturbance history (eg. stock grazing) and weed invasion than is apparent at the Yilliminning Rock reserve. The remnant woodlands and heaths around Yilliminning Rock have more in common with the vegetation types described for the Dryandra Forest to the northwest (see Coates 1993), although there are no large granite outcrops there. In fact, most of the granite outcrops mapped in the area by Beard (1980) have been cleared, with the exception of a small remnant about 10 k to the south near Nomans Lake (unpublished observations). Yilliminning Rock occurs on a reserve of 86 ha, vested in the Shire of Narrogin as an A Class Reserve (11016) in 1960. The reserve was originally set aside for water and recreation in 1906, when the area was first surveyed for agriculture. Yilliminning Rock is surrounded by about 60 ha of remnant heath and woodlands to the north and west and bordered by farmland on its east side (Fig 2). The Harrismith Road, which runs along the southern edge of the rock, provides access for the steady number of visitors from the town of Narrogin, 17.5 km to the west. Spectacular views in all directions are found at the top of Yilliminning Rock, which takes about 10 minutes to climb. With abundant wildflowers in winter and spring, the bushland at the base of the rock provides an attractive setting for picnics and walks. This paper primarily describes the vegetation and flora of Yilliminning Rock and adjacent remnant vegetation. Secondly, it discusses the current threats to the nature conservation values of the remnant vegetation, as a basis for future management. Figure 1. Location of Yilliminning Rock near Narrogin in the south-west of Western Australia. © Royal Society of Western Australia 1997 Granite Outcrops Symposium, ^996 201 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 2. Aerial photograph of Yilliminning Rock Reserve, near Narrogin, Western Australia. 0 200 400 600 800 meters LHhk Complex Thicket Dense low forest Low forest Low Woodand Woodand Dense heath Oeekline Figure 3. Vegetation map for Yilliminning Rock Reserve with vegetation associations from Table 1 shaded as Lithic Complex (1), Thicket (2), Dense Low Forest (3a). Low Forest (3b), Low Woodland (4), Woodland (5) and Dense Heath (6). The creekline running across the southwest of the reserve is unshaded. 202 Journal of the Royal Society of Western Australia, 80(3), September 1997 Materials and Methods Vegetation at Yilliminning Rock was mapped from a 1972 aerial photograph (Fig. 2) and 1996 remotely sensed image data (unpublished data). Associations and ecotone boundaries were checked during field surveys in 1996 and 1997. A list of species for Yilliminning Rock was compiled from records of the WA Herbarium (WAHERB) and collections made in spring 1992 (Anon 1993) and 1996 and autumn 1997. The survey technique known as 'randomized stratified walk' (Hopper el al. 1997) was used to sample each of the observable vegetation associations. This included a detailed circumnavigation of apron vegetation around Yilliminning Rock and several systematic traverses over the top of the rock. Vegetation classification of associations follows Muir (1977). The authors identified specimens with assistance from staff of the WA Herbarium and S Hopper, Director of the Kings Park and Botanic Gardens. Nomenclature follows WACENSUS (WA Herbarium census of Western Australian vascular plants) and conservation status of species assigned from Atkins (1996) according to the following definitions (Table 1). Table 1 Categories of Priority Flora according to the degree of perceived threat. Category Definition Priority One - Poorly Known Taxa Taxa which are known from one or a few (generally <5) populations which are under threat Priority Two - Poorly Known Taxa Taxa which are known from one or a few (generally <5) populations, at least some of which are not believed to be under immediate threat i.e. not currently endangered. Priority Three - Poorly Known Taxa Taxa which are known from several populations, at least some of which are not believed to be under immediate threat i.e. not currently endangered. Priority Four - Rare Taxa Taxa which are considered to have been adequately surveyed and which, while being rare (in Australia), are not currently threatened by any identifiable factors. Results and Discussion Vegetation associations Six vegetation associations, including canopy cover variations for 1 of these, were mapped for the Yilliminning Rock reserve (Fig 3). A seventh category, described as a degraded creekline, was also mapped. A description of each association and variants were made and important structural species and generalised understorey descriptions noted (Table 2). Table 2 Vegetation associations at Yilliminning Rock, including variations and important species. 1: Lithic Complex (granite rock) Kunzea pulchella, Pimelea graniticola, Borya sphaerocephala, Cheilanthes spp and many herb species. 2: Thicket (granite apron) Acacia lasiocalyx, Calytrix leschenaultii, Dodonaea spp, Dryandra spp and other shrub species. 3: a, Dense low forest Allocasuarina huegeliana with scattered Acacia and b, Low forest acuminata , A. saligna and Eucalyptus wandoo, with (granite perimeter) shrub and herbaceous understorey species. 4: Low Woodland Eucalyptus wandoo with scattered Acacia acuminata and sparse, open shrub understorey. 5: Woodland Eucalyptus salmonophloia with sub-shrub and herbaceous understorey species 6: Dense heath Dryandra spp, Melaleuca spp, Isopogon teretifolius with other shrub and herbaceous species. 7: Creek line Degraded; dominated by Ehrharta calycina Vegetation on Yilliminning Rock is not uniformly distributed; the shape and slope of the rock are different on all sides. Vegetation usually only associated with granite outcrops occurs on the west side (Association 2; Table 2) dominated by a diversity of flowering shrubs (e.g. Dryandra and Dodonaea spp) and sedges (e.g. Lepidosperma spp). Crevice and meadow vegetation, also endemic to granite outcrops in the southwest of Western Australia, is scattered across the rock. Steeper parts of the north and western slopes are the exception to this. Distinctive endemics, such as Kunzea pulchella and Baeckea camphorosmae and Cheilanthes ferns, grow in rock crevices (Association 1). Small meadows are also found on Yilliminning Rock comprising Borya sphaeocephala and herbaceous endemics such as Thelmytra antennifera, Leporella fimbria ta, Crassula spp and Levenhookia spp (see also Appendix 1). Allocasuarina huegeliana forest (Association 3; Table 2) clothes the southern face of Yilliminning Rock to the road to the top (Fig 2). The understorey is primarily herbaceous due to the copious needle litter produced by this tree species. Most notable in this association are 14 species of orchids and numerous species of everlastings. This dense forest association is also found on the northeast side of Yilliminning Rock (Fig 3) where it forms an impenetrable forest with Acacia saligna and Xanthorrhoea brunonis. A small but distinctive pocket (about 10 ha) of salmon gum woodland (Association 5; Table 2) lies between the wandoo woodland (Association 4) and rock casuarina forest (Association 3a & 3b; Fig 3). Salmon gum woodland is poorly represented in conservation reserves in the area and this remnant contributes greatly to habitat values of the reserve. Heath (association 6; Table 2) occupies the northern part of the reserve (Fig 3). It is an important connection to heath on the other side of Birdwhistle Road (Reserve 13343) (Fig 2) and Nature Reserve 17115, approximately 2 km to the northeast. 203 Journal of the Royal Society of Western Australia, 80(3), September 1997 The creekline (category 7) runs across the southwest of the reserve. It seems not to be degraded native creek vegetation as at other remnants east of the reserve ( unpublished data) but a drainage channel. Ehrharta calycina (veld grass) and Chenopodinm album (fat hen) dominate the creekline. Regardless of this, seasonal inundation of stormwater from adjacent farmland, including high levels of nutrients, favour agricultural weeds (e.g. veld grass) which largely replace native species in these low-lying areas. Flora Species richness. Granite outcrops in south-west Western Australia are local centres of plant diversity, with high endemism (Hopper et al 1997). Yilliminning Rock fits this model with a high number of species recorded and several known threatened taxa. Appendix 1 lists 238 vascular plant species, including 219 native species, compiled from all vegetation associations (Table 2). Fifty-four families are represented in the list. Appendix 2 lists 33 species of lichens representing 11 families. This diversity of species and communities is high when compared to other woodland reserves in the district ( unpublished data). Highest numbers of species were recorded from the granite rock crevice, meadow and fringing plant communities, although woodland species are well represented in the species list (Appendix 1). The seven best represented families at the Yilliminning Rock Reserve (see Appendix 1) were Proteaceae (21), Orchidaceae (19), Asteraceae (17) Myrtaceae (17), Poaceae (16), Papilionaceae (13) and the lichen family Parmeliaceae (16). Genera with the highest number of species were Acacia (8), Hakea (8), Dryandra (6) and Caladenia (6) and the lichen genera Xanlhoparmelia (7). Priority species. Many populations of priority taxa (Table 1) in the southwest of Western Australia are found outside conservation reserves (Coates & Atkins 1997), and Yilliminning Rock highlights this. Five priority taxa of vascular plants and 2 priority lichens are known to occur in the reserve, as well as species of Boronia (B. aff subsessilis) and Melaleuca (aff scabra) that are undescribed (see Appendix 1). • Caladenia Integra. Priority 4. C. integra is known from scattered populations between Tenterden in the south and Clackline in the north, with a disjunct occurrence near Kalbarri (Hoffman & Brown 1992). The southern populations seem to be confined to dense rock sheoak woodlands on and around granite outcrops. C. integra was collected (LWS 738) from low sheoak forest (Fig 3) on the reserve in 1996 where it was found to be common. • Dryandra fasciculata. Priority 3. D. fasciculata is only known from a few collections in the southern wheatbelt shire of Kulin and further south at Nyabing. A collection was made from a small population in 1996 (LWS 846) in thicket at the base of Yilliminning Rock (Fig 3). The occurrence of this species at Yilliminning Rock extends its known distribution westward by approximately 60 km. • Dryandra meganotia. Priority 3. D. meganotia (syn D. serratuloides subsp meganotia) was recorded from heath on the reserve in 1981 by D Hart (then Department of Fisheries and Wildlife). Since then it has been collected by G Durell in 1995 (GD 98) and by the authors (LWS sn) This species was also found in heath on the adjacent Reserve 13343 (Fig 2). These populations extend the range of D. meganotia west, previously only reported from between Kulin and Nyabing (George 1996). • Pimelea graniticola. Priority 2. P. graniticola occurs only on granite outcrops, in soil pockets over granite sheets (Rye 1988). The authors, with S Hopper, collected from 2 new populations (LWS 719 & 840) on Yilliminning Rock in October 1996, each comprising of over one hundred plants, extending the known range west by about 100 km. • Stylidium tenuicarpum. Priority 4. S. tenuicarpum is known only from Tutanning Nature Reserve east of Pingelly and the immediate area (Carlquist 1969). The authors, with S Hopper, discovered a new population (LWS 842) on Yilliminning Rock in October 1996. Lichens. As with other granite outcrops in the wheatbelt, Yilliminning Rock is an important centre of diversity for lichens. A total of 33 lichens, belonging to 11 families have been recorded from the rock. These include 31 species recorded in the Western Australian Herbarium's database WAHERB from a collection made by N Sammy in October 1981 (Appendix 2). Two additional species, Paraparmelia sammyi and P. sargentii have been recorded by Elix & Johnston (1988) and are listed by the Department of Conservation and Land Management as Priority 2 threatened flora. Introduced weeds. Only 19 species or 8% of the flora at Yilliminning Rock reserve are introduced weeds. This is lower than ratio's for other remnant woodlands in the area (unpublished data). However, many of these are cosmopolitan environmental weeds, some dense in various habitats at Yilliminning Rock, which pose a significant threat to native species and plant communities. The major weeds recorded at Yilliminning Rock are Avena fatua , Briza maxima , Ehrharta longiflora, E. calicyna, Ereesia leichtlinii x alba , Hypochoeris glabra , Romulea rosea var australis and Ursinea anthemoides. The worst of these is Ereesia leichtlinii x alba , first recorded in 1992 by one of the authors (JPP) at the base of Yilliminning Rock and observed the following year in a crevice on top of the rock. It appears to have invaded from the settlement on the east side Yilliminning Rock (Fig 2) and has now spread through all crevice and meadow vegetation there. Of particular concern is the threat to the newly recorded populations of Pimelea graniticola. The latter species appears to be an obligate seeder, adapted to a reduced disturbance habitat with minimal competition for resources. The rapid and overwhelming invasion by Freesia leichtlinii x alba must eventually have a detrimental effect on the recruitment and survival patterns of this and other native species growing in the crevice and meadow communities. Management issues Access. There are many management issues at Yilliminning Rock common to most other wheatbelt granite outcrops (see Main 1997) that need to be 204 Journal of the Royal Society of Western Australia, 80(3), September 1997 addressed. The most significant issues are the control of public access and weeds. Whilst there is no doubt about the importance of public access to the top of the rock, assessment of the current situation is needed to halt the continuing damage to the western face of the rock. Timber has been cut from woodlands in the reserve but some time ago (Anon 1981). At present, the public is allowed access to the western side of Yilliminning Rock along a graded track from Birdwhistle Rock Road (Fig 2). A small area at the end of this track allows for vehicle turnaround and parking at the base of the rock. A wide and well-worn path is visible up the rockface at this point. Damage over a wide area is a consequence of loose rock being removed or smashed over time. Most lichens and mosses are gone on this part of Yilliminning Rock (Fig 2) and the few meadows left are degraded and weedy. There is also some broken glass and litter about the entry to the rockface from the carpark. A wildfire in 1990, presumably deliberately lit because of its origin and spread, has resulted in weed invasion in this area. There is also unwanted access off the main entry track north to another part of the rock some 100 m around from the main carpark. Other undesirable car tracks lead north and south from the main track. Salmon gum and wandoo woodland herbfields and soils are sensitive to this kind of damage and it may take years for vehicle tyre marks to disappear. These tracks needed to be closed off with barriers, augmented by the planting of local trees. Weeds. Weeds invade in response to disturbance and altered environmental conditions. The 1990 wildfire burnt most of the vegetation around the apron on the west side of Yilliminning Rock. Although resprouters dominate these communities, the disturbance allows weeds to establish, particulary in the meadows at the base of the rock and the more open areas of rock casuarina woodland. The common bushland weeds Hypochoeris glabra and Ursinea anthemoides are prominent here. Another weed, the small sedge J uncus bufonius, is common in the granite rock meadows and threatens the stability of this community because of its dense soil seed-bank (unpublished data). The most serious weed, Freesia leichtlinii x alba, has spread rapidly since the early 1990's into much of the low forest and crevice plant communities on the east side of Yilliminning Rock. Immediate measures to control this aggressive weed and monitor its impact on populations of Pimelea graniticola are required. Excess water runoff, which often includes fertilizer from nearby agricultural areas, also provides suitable conditions for weed invasion (Hobbs 1991). One such area, the creekline vegetation in the southwest of the reserve (Fig 3), is infested with grass weeds (e.g. Ehrharta calycina). Likewise, the open area at the foot of the northeastern face of Yilliminning Rock, is also heavily infested with grass weeds (e.g. Arena fatua and Bromus diandrus). This area has been disturbed by unwanted vehicle access also receives high runoff from the steepest part of Yilliminning Rock in winter. Grass weeds at these sites could be controlled with selective herbicides. Conclusions Yilliminning Rock is an important local remnant with a broad diversity of habitats and a relatively high level of species richness in vascular plants and lichens. It therefore has high nature conservation value, providing an important link between larger nature reserves to the northeast and the chain of wetland reserves and farm remnants to the southeast, many of which are degraded. The rock itself is of great significance because of its large size and good condition compared to the many degraded granite outcrops on farms in the area, including the nearby Birdwhistle Rock reserve. A management plan is needed to assist the Narrogin Shire in directing future management Yilliminning Rock (see Anon 1995). Major problems such as access and weed invasion should be addressed. This could be done in as joint consultative project involving the shire, the Department of Conservation and Land Management, the Land Conservation District Committee and other community groups. Part of this process could include public education about ecological issues and the recreation in the Narrogin area, resulting in better management of other remnants. Acknowledgements. The authors thank K Morris and N Marchant (Science and Information Division, CALM) for supporting this project. The assistance of other CALM staff, G Durrell of the Narrogin District and R Cranfield and T Macfarlane of the Western Australian Herbarium, with information on vascular plant and lichen species is also acknowledged- W Roe provided valuable assistance with field work and an inspection report in 1993. S Hopper of the Kings Park and Botanic Gardens is thanked for accompanying the authors on a field trip in October 1996 and sharing his extensive knowledge of granite rock flora. The thorough collection of lichens made at Yilliminning Rock in 1981 by Mr N Sammy are gratefully acknowledged. A Hopkins, N Gibson and S Hopper made valuable comments on an earlier version of the manuscript. References Anon 1981 Reserve Inspection Report, Yilliminning Rock (11016). Department of Conservation and Land Management, Perth. Anon 1993 Recommendation for Nature Reserve Report, Yilliminning Rock (11016). Department of Conservation and Land Management, Perth. Anon 1995 Dryandra Management Plan 1995-2000. Department of Conservation and Land Management, Perth. Atkins K A 1996 Declared Rare Flora and Priority Flora list. Wildlife Branch, Department of Conservation and Land Management, Perth. Beard J S 1980 The vegetation survey of the Corrigin area. Map and explanatory memoir 1:250000 series. Vegmap Publications, Perth. Carlquist S 1969 Studies in Stylidaceae: New taxa, field observations and evolutionary tendancies. Aliso 17:13- 164. Coates A 1993 Vegetation survey of the Dryandra Forest. Report. Department of Conservation and Land Management, Perth. Coates D J & Atkins K A 1997 Threatened flora of Western Australia: A focus for conservation outside reserves. In: Conservation Outside Reserves (eds P T Hale & D Lamb). Centre for Conservation Biology, University of Queensland, in press. Elix J A & Johnston J 1988 New species in the lichen family Parmeliaceae (Ascomycotina) from the southern hemisphere. Mycotaxon 31:491-510. 205 Journal of the Royal Society of Western Australia, 80(3), September 1997 George A S 1996 New taxa and a new infrageneric classification in Dryandra R Br (Proteaceae: Grevilleoideae) Nuytsia 10:313-408. Hobbs R J 1991 Disturbance a precursor to weed invasion in native vegetation. Plant Protection Quarterly 6:99-104. Hoffman I & Brown A 1992 Orchids of the South Western Australia. University of Western Australia Press, Perth. Hopper S D, Brown A & Marchant N G 1997 Plants of Western Australian granite outcrops. Journal of the Royal Society of Wstern Australia 80:141-158. Main A R 1997 Management of granite outcrops. Journal of the Royal Society of Western Australia 80:185-188. Muir B G 1977 Vegetation and habitats of the Bendering Reserve. Biological survey of the Western Australian wheatbelt. Part 2. Records of the Western Australian Museum. Suppl 3. Rye B L 1988 A revision of Western Australian Thymelaeaceae. Nuytsia 6:129-278. Appendix 1 Vascular plant species at Yilliminning Rock Reserve listed by family and with vegetation association from Table 2 and Fig 3; Lithic Complex (1), Thicket (2), Dense Low Forest (3a), Low Forest (3b), Low Woodland (4), Woodland (5) and Dense Heath (6). Introduced species (mostly invasive weeds) are marked * *. Family & species Vegetation Association ADIANTACEAE Cheilanthes austrotenuifolia H Quirk & TC Chambers 1 C. distans (R Br) Mett 1 C. sieberi Kunze subsp sieberi 1 ASPLENIACEAE Pleurosorus rutifolius (R Br) ex Benth 1 JUNCAGINACEAE Triglochin calcitrapa Hook 1 POACEAE *Aira caryophyllea L 1,2-3 Amphipogon strict us R Br 3,4 A. turbinatus R Br 3,4 Austrostipa elegant issima (Labill SWL) Jacobs & J Everett 1, 2-3,4 A. semibarbata Labill 5 *Avena fatua L 1,2,3 *A. sativa L 2-3 *Briza maxima L 1,4 *B. minor L 1,4 *Bromus diandrus Roth 1 * Ehrharta calycina Smith 4 & creekline *E. longiflora Smith 4 *Lolium rigidum Gaudin 2-3 Neurachne alopecuroidea R Br 2-3,6 Notodanthonia cacspitosa (Gaudich)Zotov 5 Spartochloa scirpoidea (F Muell) Maiden & E Betche 1 *Vulpia rnyuros (L) C Gmelin 1,4 CYPERACEAE Caustis dioica R Br 2-3,5 Gahnia australis (Nees) KL Wilson 2-3,5 Lepidosperma gracile R Br 2-3 L. pruinosum Kuek 6 L. resinosum (Nees) Benth 2-3,5 Mestnolaena stygia (R Br) Nees aff subsp stygia 4 Schoenus brevisetis (R Br) Benth 6 RESTIONACEAE Desmocladus flexuosa (R Br)LASJohnson & BG Briggs ms 2-3 Desmocladus sp GPP sn) 5 CENTROLEPIDACEAE Aphelia brizula F Muell 2-3 Centrolepis aristata(R Br) Roemer & Schultes 2-3 Centrolepis pilosa Hieron 2-3 C. polygyna (R Br) Hieron 1 JUNCACEAE *]uncus bufonius L 1 ANTHERICACEAE Arthropodium capillipes Endl 1,2-3 Borya sphaerocephala R Br 1, 2,3 Caesia parviflora R Br 2-3 Chamaescilla corymbosa (R Br) F Muell ex Benth 2-3,4 Laxmannia squarrosa Lindley 2-3 L. grandiflora Lindley 4 Soiverbaea laxiflora Lindley 2-3,4 Thysanotus patersoni R Br 1,2-3 DASYPOGONACEAE Lotnandra effusa (Lindley) Ewart 5 L. micrantha (Endl) Ewart 5 L rupeslris (Endl) Ewart 3b L suaveolens (Endl) Ewart 4 Lomandra sp (LWS 724) 3 XANTHORRHOEACEAE Xanthorrhoea brunonis Endl In Lehm 2-3,6 PHORMIACEAE Dianella revoluta R Br 4 Stypandra glauca R Br 1,2-3 HAEMODORACEAE Anigozanthos humilis Lindley 2-3,6 Conostylis aculeata R Br subsp bromeloides (Endl) JW Green 2-3 C. pusilla Endl 2-3,4 G setiga R Br 2-3 Haemodorum laxum R Br 2-3 H. aff paniculatum QPP sn) 4 IRIDACEAE *Freesia leichtlinii Klatt x alba 1,2,3,5 Orthrosanihus laxus (Endl) Benth 2-3 Patersonia juncea Lindley 2-3 *Romulea rosea (L) Ecklon 1,4 ORCHIDACEAE Burnettia nigricans (R Br) Hopper & Brown 2-3,4 Caladenia chapmanu ms (LWS 740,742,743,744) 2-3, 1,4 C. falcata (Nicholls) MAClem & Hopper 2-3 C.flam R Br 2-3,1, 4 C. hirta Lindley subsp hirta ms (LWS 747) 4 C. integra E Coleman 2-3,1 C. longicauda Lindley subsp cminens ms (LWS 745) 4 Cyanicula deformis (R Br) 2-3 C. gemmata (Lindley) Hopper & Brown 1 Diuris corymbosa Lindley 2-3,4 Diuris sp GPP sn) 2-3 Eriochilus dilatatus Lindl 5 Leporclla fimbriata (Lindl) AS George 5 Pterostylis hamiltonii Nicholls 2-3 P. recurva Benth 2-3 P. sanguinea DL Jones & MA Clem 4 Spiculaea ciliata Lindley 1 Thclymitra nuda R Br 1 T. antennifera (Lindley) JD Hook 1 CASUARINACEAE Allocasuarina huegeliana (MIQ) L Johnson 2-3,4 A. microstachya (Miq) L Johnson 2-3,1 PROTEACEAE Adenanthos cygnorum Diels aff subsp cygnorum 6 Banksia sphaerocarpa R Br var sphaerocarpa 6 Conospermum stoechadis Endl subsp sclerophyllum (Lindl) EM Benn 2-3 Dryandra armata R Br 6 206 Journal of the Royal Society of Western Australia, 80(3), September 1997 D.fasciculata AS George 2-3 D.fraseri R Br var fraseri 5 D. meganotia AS George 6 D. nivea (Labill) R Br subsp nivea 2-3 D. squarrosa R Br subsp squarrosa 2-3 Grevillea pilulifera (Lindley) Druce 2-3 Hakea arnplexicaulis R Br 2-3 H. baxteri R Br 6 H. incrassata R Br 6 H. lissocarpha R Br 5 H. petiolaris Meissner 2-3 H. prostrata R Br 5 H. trifurcata (Smith) R Br 2-3 H. undulata R Br 2-3 Isopogon teretifolius R Br 6 Persoonia trinervis Meissner 2-3 Petrophile squamata R Br 2-3,6 Synaphea petiolaris R Br 2-3 Synaphea sp (Pieroni 14) 4 SANTALACEAE Choretrum glomeratum R Br var glomeratum 4 Exocarpos aphyllus R Br 5 Leptomeria pauciflora R Br 2-3 L. spinosa (MIQ) A DC 2-3 Santalum acuminatum (R Br)A DC 2-3 LORANTHACEAE Amyerna miquelii (Miq) Tiegh 4,5 GYROSTEMONACEAE Gyrostemon subnudus (Nees) Baill 2-3 CHENOPODIACEAE Atriplex exilifolia F Muell 4-5 *Chenopodium album L 4 & creekline PORTULACACEAE Calandrinia calyptrata JD Hook 1 CARYOPHLLACEAE *Petrorhagia velutina (Guss)P Ball 1 LAURACEAE Cassytha aurea JZ Weber var hirta 2-3 DROSERACEAE Drosera glanduligera Lehm 1,2-3 D. macrantha Endl subsp macrantha 1,4,5 D. menziesii R Br subsp menziesii 2,3,4 D. aff pycnoblasta (LWS 717) 5 D. aff giganta (LWS 715) 1,2-3 CRASSULACEAE Crassula closiana (Gay) Reiche 1 C. decumbens Thunb var decumbens 1 *C. natans Thunb var minus 1 PITTOSPORACEAE Sollya heterophylla Lindley 5 MIMOSACEAE Acacia acuminata Benth 2-3,4, 5 A. erinicea Benth 5 A. lasiocarpa var sedifolia (Meisn) Maslin 4 A. lasiocalyx CRP Andrews 2-3 A. microbotra Benth 3,4 A. pulchella R Br 5 A. saligna (Labill) HL Wendl 2-3 A. stenoptera Benth 4,5 Acacia sp (LWS 716) 2-3 Acacia sp (BR Maslin 6754) 4 Acacia sp (BR Maslin 6757) 6 PAPILIONACEAE Bossiaea eriocarpa Benth 4 Daviesia cardiophylla F Muell 2-3,6 D. hakeoides Meisn subsp hakeoides 2 D. preissh Meissner 2-3,5 D. uncinata Crisp 6 Gastrolobium calycinum Benth 4 G. spinosum Benth var spinosum 2-3,4 Gompholobium marginatum R Br 2-3,4 Isotropis cuneifolia (Smith) Benth ex BD Jackson 4 Jackson ia aff sternbergiana Huegel 6 /. eremodendron E Pritz 6 Sphaerolobium medium R Br 2-3 Templet on ia sulcata (Meissner) Benth 5 GERANIACEAE *Erodium botrys (Cav) Bertol 1 RUTACEAE Boronia sp nov LWS 836 (aff B subsessilis ) 2-3 TREMANDRACEAE Tetratheca hirsuta Lindley 2-3,4 POLYGALACEAE Comesperma ciliatum Steetz 2-3 C. scoparium Steetz 2-3 EUPHORB1ACEAE Poranthera microphylla Brongn 2-3 STACKHOUSIACEAE Stackixousia huegelii Endl 2-3 S. pubescens A Rich 2-3 SAPINDACEAE Dodonaea viscosa Jacq subsp spathulata 1,4 D. pinifolia Miq 2-3 RHAMNACEAE Cryptandra nutans Steud 2-3 C. pungens Steud 2-3 Cryptandra sp (LWS 710) 3 STERCULIACEAE Thomasia foliosa Gay 2-3 DILLENIACEAE Hibbertia acerosa (R Br ex DC) Benth 2-3 H. exasperata (Steudel) Briq 4 H. rupicola (S Moore) C Gardner 2-3 STYLIDIACEAE Levenhookia dubia Sonder 1 L. pusilla R Br 1 Stylidium calcaratum R Br 1 S. ecorne (F Muell ex R Erickson & JH Willis) PG Farrell & SH James 1 S. petiolare Sonder 2-3 S. piliferum R Br 1 S. tenuicarpum Carlq 2-3 THYMELIACEAE Pimelea graniticola Rye 1 P. rosea R Br 2-3 MYRTACEAE Baeckea camphorosmae Endl 2-3 B. crispiflora F Muell 5 Baeckea sp Narrogin (R Hnatiuk 780011) 2-3 Calothamnus quadrifidus R Br 2-3 Calytrix leschenaultii (Schauer) Benth 2-3 Eucalyptus salmonophloia F Muell 5 E. rudis Endl 4 (N only) E. wandoo Blakely 3,4,5 Kunzea pulchella (Lindley) AS George 1,2-3 K. recurva Schauer in Lehm 6 Leptospermum erubcscens Schauer in Lehm 2-3 Leptospermum sp (LWS 838) Melaleuca pungens Schauer in Lehm 6 M. scabra R Br 5 M. undulata Benth 6 M. sp nov LWS 987 (M. scabra group) 5 Thryptomene australis Endl 1 Verticordia endlicheriana Schauer 2-3,6 V. grandiflora Endl 2-3,5 V. insignis Endl subsp compta 6 V. plumosa (Desf) Druce 6 207 Journal of the Royal Society of Western Australia, 80(3), September 1997 HALORAGACEAE Glischrocaryon aureum (Lindley) Orch var aureum 1,2-3 Gonocarpus nodulosus Nees 1 APIACEAE Hydrocotyle sp 2-3 Trachymene ornata (Endl) Druce 1 T. pilosa Smith 2-3,1 EPACRIDACEAE Astroloma aff drummondii Sonder 2-3,5 A. pallidum R Br 5 Leucopogon dielsianus E Pritzel 4, 6 Lysinema ciliatum R Br 2-3,4 PRIMULACEAE *Anagallis arvensis L 2-3 LOGANIACEAE Logania sp (JPP sn) 2-3 Mitrasacme paradoxa R Br 2-3 VERBENACEAE Halgania sp (JPP sn) 2-3 SOLANACEAE Nicotiana rotundifolia Lindl 1 SCROPHULAR1ACEAE *Parentucellia latifolia (L) Caruel 2,3 MYOPORACEAE Eremophila aff glabra (LWS 729) 5 RUBIACEAE Opercularia vagimta Labill 2-3 GOODENIACEAE Dampiera alata Lindley 2-3 D. lavandulacea Lindl 5 D. lindleyi Vriese 2-3 Goodenia hebttsii (E Pritz) Carolin 1 G. micfaniha Hemsl ex Carolin 1, 2-3 G. pulchella Benth 1,2 Lechenaultia biloba Lindley 6 ASTERACEAE Brachycome ibidcrifolia Benth 2-3,4, 5 B. perpusilla (Steetz) J Black 1, 2-3 Gnephosis tenuissima Cass 2-3 Helichrysum leucopsideum DC 5 Helipterum laeve (AGray) Benth 5 Hyaospermum demissutn (A Gray) PG Wilson 1 *Hypochaeris glabra L 1 ,2-3,4 Lagenifera huegelii Benth in Endl 2-3 Podolepis canescans Cunn Ex DC 5 P. lessonii (Cass) Benth 2-3,5 Podotheca august ifolia (Labill) Less 2-3 Quinetia uroillei Cass 2-3 Rhodanthe manglesii Lindley 2-3 * Ursinia anthemoides (L) Poiret 2-3, 1,4 Watizia acuminata Steetz 5 W. citrina (Benth) Steetz 2-3 Appendix 2 Lichen species, arranged by family, recorded at Yilliminning Rock near Narrogin, Western Australia Family & species ACAROSPORACEAE Acarospora schleicheri (Ach) Mass (syn A. citrina) CLADIACEAE Cladia aggregata (Sw) Nyl C. ferdinandii (Mull Arg) R Filson DICRANACEAE Campylopus australis Catches & Frahm C. bicolor (C. Muell.) Hook f & Wils HETERODACEAE Heterodea muelleri (Hampe) Nyl LECIDEACEAE Lecidea laeta Stir ton Lecidea sp (N Sammy 840850) Toninia sp (N Sammy 840848) PARMELIACEAE Canoparmclia pruinata (Mull Arg) Elix & Johnston Flavoparmelia rutidota (JD Hook & Taylor) Hale Neo fused m imitatrix (Taylor) Esslinger N. incantata (Esslinger) Esslinger N. pulla (Ach) Esslinger Paraparmelia sammyi Elix & Johnston P. sargentii Elix & Johnston Punctelia subalbicans (Stirton) Galloway & Elix Rhizocarpon sp (N Sammy 840860) Xanthoparmclia concomitans Elix & Johnston X. eilifii Elix & Johnston X. flavescentireagens (Gyelnik) Galloway X. notata (Kurokawa) Hale X. reptans (Kurokawa) Elix & Johnston X. substrigosa (Hale) Hale X. tasmanica (JD Hook & Taylor) Hale PHYSC1ACEAE Buellia sp (N Sammy 840862) P. tribacia (Ach) Nyl Physcia aipolia (Ehrh ex Humb) Furnrohr SIPHULACEAE SiphuUi coriacea Taylor ex Nyl TELEOSCHISTACEAE Caloplaca sp (N Sammy 840861) THELOTREMACEAE Diploschistes auslralasicus Lumbsch & Elix Diploschistes sp (JA Elix 41042) VERRUCARIACEAE Endocarpon sp (N Sammy 840825, 840826, 840827) 208 journal of the Royal Society of Western Australia, 80:209-220, 1997 Why is Santalum spicatum common near granite rocks? JED Fox School of Environmental Biology, Curtin University, GPO Box 1987, Perth WA 6001 email: rfoxjed@cc.curtin.edu.au Abstract Sandford Rocks Nature Reserve is dominated by a large granite outcrop. This reserve is notably well-endowed with trees of the root parasite sandalwood {Santalum spicatum). These are comparatively common in and among granite exposures. Trees attain 4 m in height and 20 cm basal diameter on favourable sites but are small gnarled shrubs in rock fissures. Fruiting ability differs considerably between trees. Despite apparently high densities of rabbits, continuous regeneration appears to have occurred, but only in the vicinity of parent trees. The reserve contains a number of distinct vegetation associations that are soil determined. Although sandalwood is common near exposed granite it is rarely found in association with Eucalyptus stands. It is suggested that the water-shedding properties of the granite exposures are less important to sustaining sandalwood than the presence of preferentially parasitised host species. Introduction Most investigations of parasitism have been concerned with relatively short lived species (Matthies 1995) and although parasitic plants may have major impacts on structure and dynamics of natural vegetation (Pennings & Callaway 1996), little is known of their regeneration. Confirmation that particular host species can enable long-lived parasitic plants to perform better under field conditions is time consuming but a necessary prerequisite if such species are to be cultivated (Fox et al. 1996) . Perennial woody semi-parasites tend to have larger seed (Musselman & Mann 1978) than invasive shrubs (Grice 1996) and longer-lived perennials are likely to have temporally distinct growth phases. In much of the eastern wheatbelt of Western Australia the obligate semi-parasitic small tree sandalwood {Santalum spicatum [R Br] A DC) is now uncommon. This species was the most valuable natural product to settlers and was severely exploited from about the turn of the century after the railway reached Southern Cross in 1894 (Loneragan 1990) and is now absent from some areas (Lange 1960). It remains an important and valuable resource in more arid regions of the State, accounting for some $7 million in royalties paid to the Department of Conservation and Land Management in 1993-94. That represents some 22 % of all royalty revenue, only marginally less than that from wood chips (Anon 1994). A management plan for sandalwood was prepared in 1991 (Kealley 1991). Little is known of the impact of S. spicatum on natural plant communities. Study Area and Methods Sandford Rock Nature Reserve (SRNR) lies 9 km north east of the small town of Westonia (31° 15' S, 118° 45' E). Curtin University has undertaken research into the vegetation of the reserve since 1991. The family © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 Santalaceae is represented by three small tree species: Exocarpos aphyllus R Br (leafless ballart), Santalum acuminatum (R Br) A DC (quandong) and Santalum spicatum (sandalwood). The latter is known to parasitise a number of potential hosts, usually including acacias ( e.g . A. acuminata Benth, A. aneura F Muell ex Benth, and A. tetragonophylla F Muell), Allocasuarina, Melaleuca and some herbaceous species. Eucalypts are not considered to be good hosts (Barrett & Fox 1995). Eight of 17 species recorded as hosts to Santalum spicatum by Loneragan (1990) occur at SRNR. The Reserve contains at least 14 species of Eucalyptus , 4 species of Allocasuarina and 9 species of Melaleuca (Fox et al 1993). Up to 25 species of Acacia occur, often in well-defined habitats. These numbers compare with those of Hopper et al. (1997) of 83 Acacia and 25 Melaleuca for granite outcrops generally. The climatic regime is described by Beard (1981) as extra dry mediterranean. The mean annual rainfall is around 330 mm and some summer rain is usually experienced (Fig 1). Water shed from the massive reserve 60 50 C ■ o4 40 30 c o s 20 G 10 400 cm in height (Table 1). Mean heights by localities at March 1996 are; PA- 258 cm ± se 13 (n= 39); CK3- 244 cm ± se 10 (n= 54); CK3B- 301 cm ± se 17 (n= 34); and R- 265 cm ± se 16 (n= 34). Distribution within 50 cm height classes is illustrated (Fig 3). The whole population has a normal height class distribution but the four localities differ. R trees depart most from normal distribution with two peaks. The Height classes (cm) Figure 3. Plants closest to the rocks are less normally distributed in height class than those further away. small spread of heights at CK3 probably reflects inter¬ plant competition. The high proportion of taller trees at CK3B may indicate best growing conditions. Mean stocking of sandalwood on 5.83 ha sampled (excluding area R where trees are scattered individuals across a large area) is 26.1 trees ha1. The density of sandalwood trees is very variable. For the 2.5 ha of woodland sampled at PA, where no tree is near any exposed granite, and all lie west of the stream, the mean number of trees is 15.6 ha 1 (Table 1). On the rocky sites, where the area taken is that occupied by trees and shrubs, stocking varies from 5.5 ha*1 on 2 ha constituting the east patch of Allocasuarina huegeliana (RE), through 18 ha 1 on 0.33 ha of mainly Leptospermum erubescens thicket (RS), to 40 ha 1 on the 0.2 ha of mixed vegetation at RO. Stocking is highest at the localities adjacent to granite: 180 ha 1 on 0.3 ha at CK3 and 68 ha 1 on 0.5 ha at CK3B. Sandalwood trees tend to be grouped and at each of PA, CK3 and CK3B clumps of 4-30 trees are treated separately as subsets (Table 1). These clumps occupy areas of 40 to 100 m-2 and have local densities of 400- 3 000 trees ha*1. Seedlings (sandalwoods generally <150 cm tall) also tend to be clumped and are not currently found away from parent trees. Hosts Some 48 different perennial plant species from 23 families occur in the vicinity of the measured sandalwood trees taken as 110 individuals or clumps (Table 2). Those listed include most of the important perennials found in vegetation adjacent to sandalwood at the four localities examined. The species Acacia acuminata , Dianella revoluta, Dodonaea inaequi folia and Allocasuarina huegehatia appear most frequently. The localities differ, with more species recorded at R sites (31 species, mean species per sandalwood 3.5) and least at PA and CK3 (14 species; mean species per sandalwood 3 and 2.5 respectively). Locality CK3B, with 24 species has most co-occurences per sandalwood (8). Despite the large number of Acacia species present in SRNR, only A. acuminata appears to be important as a potential host. It is the most frequent species at 3 of the 4 Table 1 Stocking data for designated sample areas at Sandford Rock Nature Reserve. Locality 1 Sample area2 (m2) Number of sandalwoods by height classes (cm) Seedlings3 Trees < 200 200-300 300-400 >400 Number ha- Trees only i All plants PA 25 000 63 9 18 10 2 15.6 40.8 Sub-set a 50 0 2 5 1 1 1800 1800 Sub-set b 40 36 0 5 0 0 1250 10 250 R diffuse 4 56 3 3 3 0 nc4 nc4 RE 20 000 0 4 3 4 0 5.5 5.5 RS 3 333 0 4 0 2 0 18 18 RO 2 000 9 1 0 5 2 40 85 CK3 3 000 41 15 24 15 0 180 316 Sub-set a 100 30 7 15 8 0 3 000 6 000 CK3B 5 000 119 3 16 9 6 68 306 Sub-set a 100 29 0 1 1 2 400 3 300 1 See Figure 2 for localities. 2 The sub-sets represent small areas of concentrated Santalum spicatum plants. 3 Seedlings generally < 150 cm tall. 4 nc= not calculated 211 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 2 Frequency of occurrence of 110 trees or clumps of Santalum spicatum with other perennial species at 4 localities within Sandford Rocks Nature Reserve; number (#) and percentage (%). Possible host species (ranked alphabetically in order of frequency) Family Life form R(n= # 34) % Locality [as in Table 1] CK3 (n=17) CK3B(n=32) # % # % PA (n= # = 27) % All as % (n= 110) Acacia acuminata Benth Mimosaceae Tree 10 29 15 88 25 78 18 67 62 Dianella revoluta R Br Phormiaceae Herbaceous 7 21 2 12 28 88 1 4 35 Dodomea imecjuifolia Turcz Sapindaceae Medium shrub 11 32 1 6 12 38 14 52 35 AUocasuarim huegetiana (Miq) L Johnson Casuarinaceae Tree 9 26 2 12 25 78 - - 33 Dodonaea viscosa Jacq Sapindaceae Medium shrub 14 41 - - 17 53 - - 28 Comesperma volubile Labill Polygalaceae Twiner 2 6 - - 19 59 - - 19 Chamaexeros fimbriata ( F Muell) Benth Dasypogonaceae Herbaceous 8 24 - - 12 38 - - 18 Eucalyptus loxophleba Benth Myrtaceae Mallee 1 3 1 6 3 9 15 56 18 Hibbertia glomerosa (Benth) F Muell Dilleniaceae Low shrub - - - - 20 63 - - 18 Lepidosperma gracile R Br Cyperaceae Sedge 4 12 - - 13 41 1 4 16 Stipa elegantissima Labill Poaceae Grass - - - - 18 56 - - 16 AUocasuarim campestris (Diels) L Johnson Casuarinaceae Medium shrub 2 6 2 12 13 41 - - 15 Alyxia buxtfolia R Br Apocynaceae Tall shrub 1 3 - - 6 19 9 33 15 Calothamnus asper Turcz Myrtaceae Medium shrub 3 9 5 29 5 16 - - 12 Hakea recurva Meissner in DC Proteaceae Medium shrub 3 9 - - - - 8 30 10 Leptospermum embescens Schauer in Lehm Myrtaceae Medium shrub 7 20 4 24 - - - - 10 Melaleuca hamulosa Turcz Myrtaceae Tall shrub - - 3 18 6 19 - - 8 Melaleuca radula Lindley Myrtaceae Medium shrub - - 2 12 7 22 - - 8 Lepidosperma drummondii Benth Cyperaceae Sedge 8 23 - - - - - - 7 Solanum nummularium S Moore Solanaceae Low shrub 2 6 - - 6 19 - - 7 Stypandra imbricata R Br Phormiaceae Herbaceous 1 3 - - 6 19 - - 6 Eucalyptus capillosa Brooker & Hopper Myrtaceae Tree - - 1 6 - - 4 15 5 Greinllea pmuculata Meissner in Lehm Proteaceae Medium shrub 5 15 - - - - - - 5 Hibbertia exasperata (Steudel) Briq Dilleniaceae Low shrub 6 18 - - - - - - 5 Keraudrema integnfolia Steudel in Lehm Sterculiaceae Low shrub - - - - 5 16 - - 5 Rhagodia drummondii Moq in DC Chenopodiaceae Low shrub - - - - 5 16 - - 5 Acacia hemi teles Benth Mimosaceae Medium shrub - - - - - - 4 15 4 Solanum lasiophyllum Dunal ex Poiret Solanaceae Low shrub - - - - 3 9 1 4 4 Spartochloa scirpoidea (Steudel) C E Hubb Poaceae Grass 4 12 - - - - - - 4 Acacia erimcea Benth AUocasuarim acutivalvis (F Muell) Mimosaceae Low shrub 1 3 - - - - 2 7 3 L Johnson Casuarinaceae Medium shrub - - - - 3 9 - - 3 Cassia nemophila Cunn ex Vogel Caesalpimaceae Medium shrub 2 6 1 6 - - - - 3 Daviesia nematophylla F Muell ex Benth Papilionaceae Medium shrub - - - - 3 9 - - 3 Scaevola spinescens R Br Goodeniaceae Low shrub - - - - - - 3 11 3 Solanum orbiculatum Dunal Solanaceae Low shrub - - - - 3 9 - - 3 Acacia colletioides Benth Mimosaceae Tall shrub - - - - - - 2 7 2 Acacia saligna (Labill) Wendl Mimosaceae Tall shrub 2 6 - - - - - - 2 Astroloma serratifolium (DC) Druce Epacridaceae Low shrub 2 6 - - - - - - 2 Melaleuca macronychia Turcz Myrtaceae Medium shrub 1 3 1 6 - - - - 2 Olearia revoluta F Muell ex Benth Asteraceae Low shrub - - 2 12 - - - - 2 Baeckea behrii (Schldl) F Muell Billardtera erubescens (Putterl) E Myrtaceae Tall shrub 1 3 - - - - - - <1 M Bennett Pittosporaceae Twiner 1 3 - - - - - - <1 Eremophila scoparia (R Br) F Muell Myoporaceae Low shrub 1 3 - - - - - - <1 Eucalyptus crucis Maiden Myrtaceae Tree 1 3 - - - - - - <1 Melaleuca uncinata R Br in W T Aiton Myrtaceae Tall shrub - - - - 1 3 - - <1 Persoonia striata R Br Proteaceae Medium shrub 1 3 - - - - - • - <1 Pittosporum phylliraeoides DC Pittosporaceae Tree - - - - - - 1 4 <1 Pityrodia teckiana (F Muell) E Pritzel Chloanthaceae Low shrub 1 3 - - - - - - <1 Botanical nomenclature after Green (1985). Life forms: tree - a single stem; mallee - multi-stemmed Eucalyptus; tall shrub - mature individuals > 2m in height; medium shrub - mature individuals > 1 m tall, but not > 3 m; low shrub - generally < 1 m tall. 212 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 4. Profile (tree 10; ht 377 cm, stand of Santalum spicatum (S.s.) with Eucalyptus loxophleba (E.l.) with A. acuminata (A. a.) and Dodonaea inaequifolia (D.i.), at site RO). Figure 5. Profile (tree 24; ht 330 cm, stand of Santalum spicatum (S.s.) with Allocasuarina huegeliana (A.h.), Chamaexeros fimbriata (C.f.)/ Dianella revoluta (D.r.), Lepidosperma drummondii (L.d.), and Persoonia striata (P.s.) at site RE). 213 Journal of the Royal Society of Western Australia, 80(3), September 1997 localities where more than two thirds of sandalwood have it nearby. The R locality has fewest A. acuminata with none at RS and RE. When A. acuminata is absent then Dodonaea inaequifolia and Alyxia buxifolia are present at PA and Allocasuanna huegeliana at CK3B. At RO, D. inaequifolia is also important (Fig 4); at RE A. huegeliana dominates (Fig 5); at RS and R both Dodonaea viscosa and Leptospermum erubescens are the main shrubs present. Acacia colletioides , A. hemiteles and A. erinacea are only important as associates of S. spicatum at PA, the woodland locality. Not all species listed in Table 2 are likely to be valuable as host plants to sandalwood and the 8 species listed once only may be assumed to be chance co¬ occurences. For example, Eucalyptus crucis is common at SRNR at rock edges but there is only one joint occurence with sandalwood. Similarly 5 other species from locality R, including Persoonia striata (Fig 5), occur once with sandalwood. Some 27 species co-occur with sandalwood at a frequency of 5% or less. Sandalwood is more often found adjacent to many of these species rather than within patches of them. Thus, sandalwood does not occur under trees of Eucalyptus capillosa , that, similar to the related E. zvandoo Blakely, tend to lack a shrub understorey (Lamont 1985). Many of the sandalwood with either Eucalyptus loxophleba or £. capillosa also have Acacia acuminata present. Relatively poor crowns of A. acuminata (eg. CK3B) compared with the same species where S. spicatum is absent, suggest it is damaged by the parasite. Of the thicket species, Leptospermum erubescens is highest up the list Sandalwood does not occur within dense thickets of this species, rather to the edge. The thicket (or broom bush) habit of most Melaleuca species, and of Allocasuarina campestris and Calothamnus asper (important at CFG), may hinder the development of S. spicatum as it is not found directly under shade (Figs 4, 5). Reproductive efficiency Flowering has been observed regularly in March but the intensity of flowering on particular individuals does not seem to be related to the size of the following nut crop. In March 1994, 50 sandalwood trees at 3 localities were examined for freshly fallen nuts from the 1993 season. Thirty six trees (72%) had 174 (± se 36) nuts and 14 had none. Of these 36, 14 of 27 examined in March 1995, ( i.e . 52%), had 83 (± se 36) nuts and 13 had none (9 were not examined). Thirty three of the same 36 trees were examined in March 1996 when 29 (88%) had 80 (± se 18) nuts. In March 1995, a total of 76 trees at 3 localities was examined for nuts from the 1994 season. Thirty seven (49%) had 56 (± se 15) nuts and 39 had none. In March 1996, 139 trees from 4 localities were examined for nuts from the 1995 season. Ninety seven (70%) had 149 (± se 36) nuts and 42 had none. There are significant differences by locality for the 36 trees with fruit from the 1993 season (F = 6.371; P = 0.005). Those at CK3 had more nuts (369 ± se 101; n = 9) than at PA (mean 118 ± se 36; n = 14) and R (100 ± se 36; n = 13). The latter two do not differ significantly. Differences are not significant for those of the 36 that also fruited in 1994; CK3, 82 ± se 19 (n = 2); PA, 24 ± se 12 (n = 4); R, 112 ± se 61 (n = 8); or 1995; CK3, 105 ± se 52 (n = 6); PA, 100 ± se 34 (n = 11); R, 49 ± se 17 (n = 12). Mean nut numbers differ significantly by locality for the 76 examined in 1995 (F = 4.137; P = 0.02). Those at R have most (68 ± se 32; n = 16 of which 13 had nuts and 3 did not). The other 2 sets have significantly fewer nuts: CIO, 22 ± se 7 (n =36), of which 17 had nuts and 19 did not; PA, 9 ± se 4 (n= 24), 7 with and 17 without nuts. Those with nuts (In= 37) have 83 (R); 46 (CK3); and 31 (PA) with ± se of 38, 12 and 8 respectively. These means do not differ significantly ( F = 0.992, P = 0.381 ). Significantly more nuts (F = 3.806; P = 0.012; Xn = 139) are again found for 1996 (from the 1995 fruit season) for R trees (258 ± se 110, n = 29, of which 28 had nuts), when trees are compared for each of the 4 localities. The trees from CK3B (96 ± se 38, n= 34, of which 31 had nuts), PA (88 ± se 21, n= 32, of which 28 had nuts), and CK3 (20 ± se 9, n= 44 of which 10 had nuts) do not differ in nut numbers. Those with nuts (Xn= 97) have means of 267 (R); 106 (CK3B); 100 (PA); and 88 (CK3) with ± se of 113, 41, 23 and 31 respectively. These means do not differ significantly (F = 1.461, P = 0.230). Numbers of nuts below trees from the 1995 crop are highly correlated with each of 1996 measures of mean crown diameter (CD), tree height (Ht) and stem diameters (SD 5, 15 or 130 cm). For example, of the 34 trees at CK3B correlation coefficients are as follows; SD at 130 cm, r = 0.738 (P - 0.001); CD, r - 0.625 (P = 0.001); Ht, r = 0.597 (P = 0.01); SD 5 cm, r = 0.488 (P = 0.01); and SD 15 cm, r = 0.374 (P = 0.05). Division of CK3B trees into 2 sets based on height > or < 300 cm shows that for 15 larger trees (>300 cm tall, mean 383 ± se 21 cm) the nut number per tree is 176 (± se 81). Nuts are correlated with tree dimensions as follows; SD 130 cm, r = 0.780 (P = 0.001); Ht, r = 0.698 (P Table 3 Statistics for numbers of nuts counted on 10 trees of Locality R in each of 1994, 1995 and 1996. SET Tree Tree height Number of nuts counted in (cm) in 1996 1994 1995 1996 R 1 225 191 500 50 R 30 324 480 215 216 RS 16 158 28 3 18 RS 17 178 30 7 45 RS 18 161 8 2 20 RS 19 130 0 0 0 RE 25 273 172 19 33 RE 24 330 160 40 23 RE 22 311 42 111 15 RE 23 167 7 0 10 Mean 226 112 90 43 se 24 47 51 20 F = 5.049 7.624 13.386 4.691 P = 0.044 0.018 0.004 0.051 Means R 275 a 336 a 358 a 133 a RS 157b 17b 3b 21 b RE 270 3 95 h 43 b 20 b superscript letters indicate significant differences within columns, by LSD test. 214 Journal of the Royal Society of Western Australia, 80(3), September 1997 = 0.01); CD, r = 0.674 (P = 0.01); SD 5 cm, r = 0.525 (P = 0.05); but not SD 15 cm, r = 0.284 (NS). For the 19 smaller trees (<300 cm tall, 237 ± se 11 cm) the nut number per tree is 34 (± se 10) and nut numbers are correlated only with CD, r = 0.732 (P = 0.001). Other values are; Ht, r = -0.034 (NS); SD 15 cm, r = 0.398 (NS); SD 5 cm, r = 0.164 (NS); and SD 130 cm, r = 0.029 (NS). At locality R, 1996 nut numbers vary from 1 to 2 500. Despite considerable variation, there is no significant difference in 1996 nut counts between the four sub-sets of R (F = 1.653; P = 0.203). The following summarises 1996 counts; set R, n with nuts = 8 351 ± se 308, range 1- 2 500; set RO, n = 7, 603 ± se 259, range 20-2 000; set RE, n = 9, 28 ± se 9, range 8-77; set RS, n = 5, 40 ± se 11, range 18-80. Nut records are incomplete for the 3 year period but the available yields from 10 trees of locality R reveal significant differences between sets for each year (Table 3). RS trees are of smaller stature, yield fewest nuts, occur very close to rock (tree 19 is in a fissure) and have Dodonaea viscosa or Lepidosperma sedges as closest associates. Trees at R are taller, yield most nuts, occur in a range of distinctive habitats and of the two listed one is associated w'ith Acacia acuminata and Allocasuarina huegeliana , the other with Leptospermum erubescens and Dodonaea viscosa. Trees at RE occur with Allocasuarina huegeliana , Dodonaea viscosa or Leptospermum erubescens. These 10 indicate a decline in nut production over the 3 seasons, although only 4 trees give largest yields for the 1993 fruit season and 3 have highest yields from the most recent season. Consideration of the 1996 nut assessment (n = 97) suggests that quantity of nuts is extremely variable but the larger trees have more. The best fit of nuts with tree dimension is the exponential relationship; nuts = 3.914x10 (o.mMCV) (r2= 0.316; Fig 6). Seedlings Sandalwood seedlings are present at each locality, with a mean stocking per ha of 39.8, excluding the diffuse set R (Table 1). Seedlings are not present with all trees e.g. on the rockier sites (R) no seedlings are present at either RE or RS. Seedlings are currently only found near or directly underneath putative parent trees. At Crown diameter (cm) Figure 7. Relationship between seedling numbers and crown diameter of parent trees for n = 73 at localities PA and CK3B combined. Seedlings = -3.539 + 0.02304 Crown diameter (cm); r2 = 0.180. locality PA, 1996 seedling numbers near measured trees are correlated with the 1996 nut count (n = 32, r = 0.805, P = 0.001) and with soil depth (r = 0.385, P = 0.05). In contrast, at CK3B, seedlings are correlated with sandalwood height (n = 34, r = 0.367, P = 0.05) and SD 15 cm (r = 0.398, P = 0.05). Lumping PA and CK3B observations (n = 73) suggests that crown diameter is the best predictor of seedling numbers (r = 0.424, P = 0.001). Figure 7 illustrates the relationship described by the regression; seedlings = -3.539 + 0.02304 crown diameter (cm) r2= 0.180 (n = 73, P < 0.001). If trees with no seedlings are excluded, then the regression is not significant; seedlings = -0.9566 4- 0.02164 crown diameter (cm) r2= 0.111 (n =31, P is NS). At CK3B, the 119 seedlings represent a mean of 3.5 (± se 1.2) per tree (n = 34). However, 12 trees have no seedlings, those 22 with seedlings have 1-38 present with 5.4 (± se 1.7). Sixteen sandalwoods have 1-5 seedlings each, accounting for 73% of all seedlings. The largest numbers per tree are associated with tree 26 (38 seedlings) and tree 11 (16 seedlings). Seedlings associated with the 15 tallest (> 300 cm) trees (6.7 ± se 2.5 seedlings) are not significantly correlated with any other observations. Those associated with the 19 shorter (< 300 cm) trees are fewer in number (0.9 ± se 0.3) and Figure 6. 1996 nut production and mean crown diameters. 500 IT 400 ;§> 300 jy S 200 J-t H 100 0 0 5 10 15 20 25 30 35 40 45 Mean soil depth (cm) Figure 8. Relationship between tree height and mean soil depth for sandalwood trees. 215 Journal of the Royal Society of Western Australia, 80(3), September 1997 500 400 g 300 u ^ 200 O) 100 < < Ph < Ph < C/1 £4 cn < < < < < < cn P4 F ' < cn s cn p^ < Ph Ph Ph, £h Ph Ph Ph, Ph K t>s t-H in t-H o 300 cm) at CK3B have a soil depth of 28.6 (± se 1.9) compared with the 19 trees < 300 cm with 24.6 (± se 1.5). This difference is not significant. Growth Six of the 10 taller sandalwoods (Table 1) are in CK3B, which also has the tallest individual sandalwood. The upper limit to growth is likely to be determined by a combination of site influences and may require a considerable time period for it to be established. Tree architecture in parasitic tree species appears less robust and formalised than in free living species. In sandalwood, growth increments are difficult to detect as foliage is brittle and easily lost to wind, bird damage and fungal infections. This affects height Figure 10. Mean heights and standard errors for 12 trees consistently measured at locality PA. 216 Figure 11. Relationship between plant height and crown diameter for plants immediately adjacent to granite (1996, n = 36) . Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 12. Relationship between stem diameters at 130 and 5 cm for measured trees at locality CK3 (two large outliers removed). Height classes (m) Figure 13. Sandalwood stocking March 1996 at CK3B, with and without seedlings. and crown estimates. From year to year changes may be subtle (Fig 9) although seedling growth may be more easily recorded. For tree heights at least five years is necessary for reasonably robust estimates of change to be detected (Fig 10). Tree architecture tends to be characteristic with some populations having a distinct main trunk, albeit not very long; others are shrubby with multiple branches from a metre or so. The characteristics of mature populations are revealed by correlations between dimensions. For example there is usually a consistent and reliable relationship between plant height and crown diameter (Fig 11) and between the several measures of stem diameter (e.g. Figure 12). The problem of assessing population dynamics in relation to seedling establishment is important. Figure 13 indicates that seedling populations can be considered separately from that of trees or can be considered together with trees. Thus far, no seedling populations have been encountered away from mature trees. Discussion On granite exposures of the Darling scarp above Perth, the Christmas tree Nuytsia floribunda (Labill) R Br is common and may have a similar role in vegetation (Lange 1960) to that of the sandalwood described here. Sandalwood does not have the same ability to persist through suckering and populations may have greater genetic diversity than in N. floribunda or than Santa him album L which can sucker profusely. The phenomenon of clumping with individuals of seed origin may render populations more robust. However, some autoparasitism must occur amongst the several dense seedling clumps and this may reduce growth. Highest apparent densities are at localities adjacent to granite but areas sampled are comparatively small. Sandalwood clumping involves both large individuals and seedlings occurring in close proximity. Ultimately this results from poor seed dispersal. The large fruits appear not to be currently taken by emu ( Dromaius novaehollandiae Latham) although evidence of its presence is seen frequently with scats heavily loaded with seed of Eremophila ionantha. Continuous regeneration is suggested by seedling densities but is at present confined to clumps. R trees show greatest departure from a normal distribution with two peaks; this may have more to do with the variation in habitats encompassed than to past recruitment periods. Seedling presence near parent trees may reflect current rabbit pressure. It will be interesting to observe whether rabbit decline following the calicivirus virus coincides with sandalwood seedlings establishing further away from parent trees. The distribution of sandalwood at SRNR reveals the circumstances in which it is able to persist. It is absent from pure stands of the larger eucalypt trees e.g. Eucalyptus capillosa, E. salmonophloia F Muell and £. salubris F Muell, presumably as these dominate their edaphic environments and may allow insufficient light for sandalwood. Competition for soil moisture in dry weather is the main factor limiting the stocking of large trees that can be supported in such stands. The taller woodlands are confined to areas of deeper soils with limited stands of salmon gum (£. salmonophloia) forming the tallest woodland stratum. Although sandalwood is absent with this species, Santalum acuminatum is able to occur (Keenan 1993). Sandalwood can persist in the vicinity of York gum mallee; this species is frequently in small patches that border Acacia and Leptospermum erubescens thickets. Plant parasites can attack many hosts and can use multiple hosts simultaneously (Musselman & Mann 1978; Gibson & Watkinson 1989). A single S. spicatum may parasitise variable numbers and species of hosts (Herbert 1925). It is assumed that plants frequently with sandalwood are both good hosts and good survivors. Hosts that cause best parasite performance may suffer more from parasitism and there will be some upper limit to the amount of parasites that can be tolerated by particular host communities. Whereas host preferences of epiphytic parasites can be observed directly, those of root parasitic trees can only be inferred without exploring root connections (Pennings & Callaway 1996). Pot trials with host/ parasite pairs suggest direct negative effects on host 217 Journal of the Royal Society of Western Australia, 80(3), September 1997 growth (Marvier 1996) and that growth is better when associated with some hosts rather than others (Rai 1990; Fox et al 1996; Marvier 1996). Six of those listed are noted by Loneragan (1990) as host species; Acacia acuminata, A. colletioides , A. hemiteles, Cassia nemophila , Eucalyptus loxophleba and E. capillosa. The jam tree Acacia acuminata, the lily Dianella revoluta, the medium shrub Dodonaca inaequifolia and the tree Allocasuarina huegeliana, as the most frequent associates of Santalum spicatum at SRNR, are likely the best hosts. All except D. inaequifolia are present at the successful sandalwood plantation at Dryandra (Struthers et al. 1986) and Loneragan (1990) has indicated another Dodonaea (D. lobulata F Muell) as an important host. A species listed by Loneragan (1990) that does not feature in Table 2 is Eremophila ionantha (Diels) in Diels & E Pritzel. This is common at SRNR as an understorey species in eucalypt stands. Species listed by Loneragan (1990) that are probably not important hosts with S. spicatum near granite areas at SRNR are; Acacia colletioides and A. hemiteles, and the eucalypts E. loxophleba and E. capillosa. These Acacia species, and A. erinacea, are associated with salmon gum (Fox et al 1993) and only occur as associates of S. spicatum at PA, the woodland locality on slightly deeper soils. Other Acacia species of this reserve are confined to sandier soils where S. spicatum is entirely absent (e.g. A. stereophylla, A. neurophylla W Fitzg, A coolgardiensis Maiden). Other evidence (Fox, unpublished) suggests that eucalypts generally are poor hosts and this is generally associated with competition for soil moisture (Lamont 1985). Allocasuarina campestris, Melaleuca hamulosa and M. lateriflora form thickets to < 3 m (Fox et al. 1993) and sandalwood does not occur amongst stands of these or of Calothamnus asper. Some overhead light appears necessary to support sandalwood growth. Loneragan (1990) listed Casuarina cristata Miq and Dodonaea lobulata F Muell, as hosts. These species are not found at SRNR but it is postulated that species in these genera (and Allocasuarina) are likely to be universally good hosts for sandalwood. The tree Allocasuarina huegeliana is of considerable interest as it provides the greatest standing biomass over much of the distribution of S . spicatum and, presumably, also the greatest volume of available roots for haustorial connections; A. huegeliana is often the dominant species of rockier areas. The frequency of occurrence of the perennial lily Dianella revoluta suggests that this species is probably utilised as a host by sandalwood. Earlier circumstantial evidence linking D. revoluta to sandalwood at Dryandra is alluded to by Struthers et al. (1986). Low or sporadic fruit set is common in semi-parasites (Musselman & Mann 1978). Contrary to perceived belief (Loneragan 1990; Kealley 1991), sandalwood trees are regular in flowering. Flowering is more likely a photoperiodic response than one to rainfall, as buds invariably open at SRNR between February and May. The quantity of flowers, duration of flowering, fruit set and maturation are all variable and may depend on seasonal rainfall, prior fruiting history and tree size. Up to 700 flowers may result in only one fruit (Barrett 1987). Nut production is irregular (Davies 1976) and differs between trees and seasons. There is considerable range in yield with 2 000 to 2 500 nuts recorded from some larger trees. The proportion of trees fruiting in 1993 and 1995 is similar, and higher than in 1994. Similarly, the number of nuts per tree on those fruiting is lowest from the 1994 season. Of trees observed to have fruited in 1993 and also examined subsequently, mean nut production from the 1994 and 1995 seasons is about half that from 1993, and the lowest proportion of trees fruited again in 1994 compared with 1995. Significant differences in nut production between trees at different localities tend to be lost when trees barren in that year are not included in the analysis. However, it is of interest that trees at the more northern location produced most fruit per tree in the 1993 season whereas some trees from the rockier localities produced more nuts from both the 1994 and 1995 seasons. Nut production from taller trees (> 300 cm height) is correlated with several dimensions of plant size, all reflecting age and relative maturity, but for smaller trees nut yield is only related significantly with crown diameter. This suggests that crown spread has more influence on nut yield than stem size or tree height. Nut weight and diameter in S. spicatum differs in relation to longitude (Fox & Brand 1993), mainly associated with falling away of mean rainfall inland. Some trees located adjacent to granite may have higher effective soil moisture available in the fruit forming winter months due to redistribution. This may have an influence on the size of the nut crop. The massive loss of flowers (Barrett 1987), large nut size (Musselman & Mann 1978; Fox & Brand 1993), poor germination (Loneragan 1990) and brittle foliage suggests tow resource efficiency is a physiological consequence of parasitism. Regeneration of dominants is uncommon. Release from canopy held seed in eucalypts may compensate for irregular fruiting (Yates et al. 1994) but there is no such effect in sandalwood. The availability of viable seed may limit seedling regeneration and this is a topic for further research. Concern has been expressed that seedling regeneration of sandalwood is retarded by herbivore activity. Until 1996 rabbits and kangaroos were frequently seen at SRNR. A viral disease reduced kangaroo numbers in the summer of 1995-96, and rabbit numbers appear to fluctuate with myxoma virus outbreaks. It is possible that saturation of space under parents by nut material may have deterred rabbit herbivory. With the introduction of the calici virus it is possible that the next few years may see increased sandalwood regeneration, away from parent trees. In the terminology used by other speakers at this symposium, some sites near granite outcrops may be termed as apron or petticoat places. These have shallow or poorly developed soils unable to support perennial species. Other sites lie in long narrow gaps between large rounded sections of granite where deeper or more developed soils do support perennials. Sandalwood is rarely found at such sites occupied by Eucalyptus crucis; these narrow garter sites lie at areas where the steeply shelving rock appears to dive beneath the regolith such that organic matter has accumulated and where soil moisture may be comparatively high. In supporting this long-lived eucalypt the sites may be too dry in summer to support sandalwood. It is possible that sandalwood is restricted to relatively nutrient poor sites (Matthies 1995). Soils are sandy loams. 218 Journal of the Royal Society of Western Australia, 80(3), September 1997 grey and gritty (Muir 1979). Both main tree hosts occur on granitic sandy soils (Lange 1960). However, granitic soils of the region are characteristically high in potassium (McArthur 1991), important in sandalwood nutrition (Struthers et al. 1986). The two important host tree species both fix nirogen; A. acuminata (Roughley 1987), A. huegeliana (Rodriguez-Barrueco 1968). Best occurence is often between or adjacent to different vegetation types. Such locations may provide the widest range of hosts (Kealley 1991) as sandalwood root extension is considerable and allow some access to better soils. Sandalwood is not found in very shallow soils but is also scarce in well-developed eucalypt woodlands where soil depth may be much deeper. Light competition coupled with root saturation may restrict it in such areas. Alternatively, present absence may result from lack of regeneration following exploitation last century. The balance between competitive and host effects is related to growth conditions (Watkinson & Gibson 1988). Do larger parasites directly diminish host growth? Acacia acuminata is in poor condition near a number of sandalwood occurences. Another favoured species, A. tetragonophylla (Loneragan 1990) is present, but now scarce, at location PA and may have declined due to sandalwood parasitism. Do host species differ in tolerance of parasitism? Allocasuarina huegeliana appears to be more robust than the Acacia species, it reaches a large size and probably persists longer. It may not suffer as much from parasitism as A. acuminata. Conclusions Hosts function as sources of water and nutrients for the parasite. Potential hosts may offer competition by reducing light available to the parasite. Santalum spicatum appears better able to occur adjacent to the lighter shade cast by the nitrogen-fixing host trees Acacia acuminata and Allocasuarina huegeliana than under most eucalypt species. Poor crown condition of A. acuminata near S. spicatum indicates that the parasite damages at least this host. It is considered that the particular edaphic environments that permit the development of the recognised host tree species of Acacia acuminata and Allocasuarina huegeliana to flourish are mainly responsible for the prolific occurrence of Santalum spicatum at Sandford Rocks Nature Reserve. It is possible that most species of Dodonaea are likely hosts for sandalwood. The commercial size for harvest of sandalwood is a trunk diameter of > 127 mm measured at 150 mm from ground level (Keally 1991). A number of trees in the populations at SRNR exceed this criterion. Evidence is presented to demonstrate that there is no shortage of seedlings of Santalum spicatum in an environment where rabbits have been abundant. At the seedling stage sandalwood is shade tolerant but no mature trees occur under overhead shade. Four growth stages are postulated from the evidence produced; • 1) initial establishment of small seedlings (1-2 years from seed); • 2) persistence as seedling-sapling to ca 1 m height (2-8 years); • 3) growth to mature size 2. 5-3.5 m (10-30 years); • 4) continued growth in stem diameter and crown spread, little change in height (30-100 years). Acknowledgements : Thanks are expressed to my colleagues D R Barrett, A I Doronila and P M Reeve for assistance. Final year students of Terrestrial Ecology 301 have assisted with ecological studies at Sandford Rocks Nature Reserve. D Barrett, A Dick, D Ginger and S Collins are especially mentioned for measurements of the sandalwood trees in March 1996. The Department of Conservation and Land Management is acknowledged for permission to undertake research at Sandford Rocks Nature Reserve. References Anon 1994 Annual Report 1993-1994. Department of Conservation and Land Management, Western Australia, Perth. Barrett D R 1987 Initial observations on flowering and fruiting in Santalum spicatum (R. Br. ) A. DC. -the Western Australian sandalwood. Mulga Research Centre Journal 9:33-37. Barrett D R 1995 Review of the ecological characteristics of Acacia acuminata Benth. Mulga Research Centre Journal 1 2:31—38. Barrett D R & Fox JED 1995 Geographical distribution of Santalaceae and botanical characteristics of species in the genus Santalum. In: Sandalwood Seed Nursery and Plantation Technology (eds L Gjerum, JED Fox & L Ehrhart). FAO, Suva, Fiji. RAS/92/361. Field Document 8, 3-23. Beard J S 1981 Explanatory Notes to Sheet 7 Swan. Vegetation Survey of Western Australia 1: 1 000 000 Series. University of Western Australia Press, Perth. Davies S J J F 1976 Studies on the flowering season and fruit production of some arid zone shrubs and trees. Journal of Ecology 64:665-687. Fox J E D & Brand J E 1993 Preliminary observations on ecotypic variation in Santalum spicatum. 1. Phenotypic variation. Mulga Research Centre Journal 11:1-12. Fox JED, Doronila A I, Barrett D R & Surata I K 1996 Desmanthus virgatus (L.) Willd. an efficient intermediate host for the parasitic species Santalum album L. in Timor, Indonesia. Journal of Sustainable Forestry 3(4):13-23. Fox JED, Keenan C & Shepherd D P 1993 Some stand characteristics of remnant vegetation in the eastern wheatbelt of Western Australia. Research Bulletin of the Forest Research Institute, Kupang, Timor. Santalum 14:35-47. Gibson C C & Watkinson A R 1989 The host range and selectivity of a parasitic plant: Rhinanthus minor L. Oecologia 78:401-^406. Green J W 1985 Census of the Vascular Plants of Western Australia. Western Australian Herbarium. Department of Agriculture, Perth. Grice A C 1996 Seed production, dispersal and germination in Cryptostegia grandiflora and Ziziphus mauntiana, two invasive shrubs in tropical woodlands of northern Australia. Australian Journal of Ecology 21:324-331. Herbert D A 1925 The root parasitism of Western Australian Santalaceae. Journal and Proceedings of the Royal Society of Western Australia 11:127-149. Hopper S J, Brown A P & Marchant N G 1997 Plants of Western Australian granite outcrops. Journal of the Royal Society of Western Australia 80:141-158. Kealley I G 1991 Management of sandalwood. Western Australian Wildlife Management Program 8. Department of Conservation and Land Management, Perth, Western Australia. Keenan C 1993 A description of relatively undisturbed salmon gum woodland in the eastern wheatbelt. Honours Thesis. School of Environmental Biology, Curtin University, Perth. Lamont B B 1985 Gradient and zonal analysis of understorey suppression by Eucalyptus wandoo. Vegetatio 63:46-66. Lange R T 1960 Rainfall and soil control of tree species 219 Journal of the Royal Society of Western Australia, 80(3), September 1997 distribution around Narrogin, Western Australia. Journal of the Royal Society of Western Australia 43:104-110. Loneragan O W 1990 Histoncal review of sandalwood ( Santalum spicatum ) research in Western Australia. Department of Conservation and Land Management, Perth. Research Bulletin 4. Marvier M A 1996 Parasitic plant-host interactions: plant performance and indirect effects on parasite-feeding herbivores. Ecology 77:1398-1409. Matthies D 1995 Parasitic and competitive interactions between the hemi-parasites Rhinanthus serotinus and Odontites rubra and their host Medicago sativa. Journal of Ecology 83:245- 251. McArthur W M 1991 Reference Soils of South-western Australia. Department of Agriculture, Perth, Western Australia for the Australian Society of Soil Science. Muir B G 1979 Some nature reserves of the Western Australian wheatbelt. Part 14. Department of Fisheries and Wildlife, Perth, Western Australia. Musselman L J & Mann W F 1978 Root Parasites of Southern Forests. General Technical Report SO - 20. USDA Forest Service, New Orleans. Pennings S C & Callaway R M 1996 Impact of a parasitic plant on the structure and dynamics of salt marsh vegetation. Ecology 77:1410-1419. Rai S N 1990 Status and cultivation of sandalwood in India. In: Proceedings of the Symposium on Sandalwood in the Pacific (eds L Hamilton. & CE Conrad). General Technical Report PSW - 122. USDA Forest Service, Berkeley, 66-71. Rodriguez-Barrueco C 1968 The occurence of nitrogen-fixing root nodules on non-leguminous plants. Botanical Journal of the Linnean Society 62:77-84. Roughley R J 1987 Acacias and their root-nodule bacteria. In: Acacias in Developing Countries (ed J W Turnbull). Australian Centre for International Agricultural Research. Proceedings 16:45-49. Struthers R, Lamont B B, Fox JED, Wijesuriya S and Crossland T 1986 Mineral nutrition of sandalwood ( Santalum spicatum). Journal of Experimental Botany 37:1274-1284. Watkinson A R & Gibson G C 1988 Plant parasitism: the population dynamics of parasitic plants and their effect upon plant community structure. In: Plant Population Ecology (eds. A J Davy, M J Hutchings & A R Watkinson). Blackwell, Oxford, 393-411. Yates C J, Hobbs R J & Bell R W 1994 Factors limiting the recruitment of Eucalyptus salmonophloia in remnant woodlands. 1. Pattern of flowering, seed production and seed fall. Australian Journal of Botany 42:531-542. 220 Journal of the Royal Society of Western Australia, 80:221-229, 1997 Rocks as museums of evolutionary processes J D Bussell & S H James Department of Botany, The University of Western Australia, Nedlands WA 6907 email: jbussell@cyllene.uwa.edu.au Abstract The isolated granitic outcrops of Western Australia may harbour relics of the past such as Isoetes and Stylidium merrallii. They may also preserve a record of evolutionary change both within single populations and across population systems. Thus, the initial stages in the evolution of the complex hybrid genetic system are preserved amongst the extant lineages of Isotoma petraea on Pigeon Rock in the southwest of Western Australia. The elaboration of this bizarre genetic system is preserved across neighbouring outcrops. Study of /. petraea focuses attention on the evolutionary consequences of the controversial association of high levels of inbreeding with polymorphism for lethal genes within populations. In addition, it has generated novel and important insights into the genetic structure of Western Australian plant species and of evolutionary processes generally. Pigeon Rock and its unique living evolutionary museum is worthy of World Heritage listing, and should be preserved by removing the present stockyards and excluding the rock from the surrounding pastoral lease. Introduction Granite rock outcrops may be regarded as evolutionary museums on a number of fronts. They may house relict flora, such as the primitive Stylidium merrallii (Kenneally & Lowrie 1994), Isoetes, an ancient gymnosperm, and relict fauna. Alternatively, they may provide windows on evolutionary processes. The species supported by the granite outcrop habitat necessarily exist in an archipelago of small populations. These populations have a genetic structure which is a consequence of the levels of communication (or gene flow) between them, and of the devices they have evolved to cope. This paper deals with the sequential consequences of isolation and inbreeding imposed on Isotoma petraea (Lobeliaceae) on these granite rocks by the species' habitat preference and evolved adaptations. It traces the evolution of the genetic system, from a conventional outbreeding state with normal chromosome behaviour and high levels of seed production, to complex hybridity with self fertilisation, highly restricted chromosome behaviour and stringent seed selection systems. Classically, diploid sexual reproduction is thought to promote evolutionary capability by enhancing the level of genotypic diversity among gametes through recombination at meiosis and the distribution of that diversity between lineages by biparental reproduction (Fisher 1930; Muller 1932). In contrast, sex, comprising meiosis and fertilisation, is now widely recognised as having a more primitive role. Firstly, corruptions of the DNA molecule which escape biochemical correction through cellular DNA repair mechanisms may be corrected by recombinational processes, even in somatic cells (Bernstein el al. 1988). Secondly, mutations arising from errors in the repair processes or from mobile genetic element activity, and which have become incorporated into the genome, may be removed from © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 lineages by outcrossing and segregation (Bell 1988). While mutation is the ultimate source of the variation which is exploited in evolution, most mutations are deleterious, ranging from selectively neutral to lethal. Modern interpretations also suggest that diploidy has evolved as a redundancy (fail-safe) mechanism for protecting the soma against mutational corruption (Maynard Smith 1988). Providing it operates within an allelically diverse gene pool, sexual reproduction can also raise evolutionary capability to new levels relative to those achievable by mutation alone, and diploid gene pools are more capable than those associated with haploidy. Evolutionary capability is thus a consequence of unencumbered sexual genetic repair. If the sexual system is encumbered with devices which diminish its efficiency, both the evolutionary capability of the lineage and its ability to remove deleterious mutations are impaired. Classical population genetic theory predicts that gene flow between populations comprising a species' metapopulation will bind it into a single evolutionary entity. Breakdown of this communication, by geographic isolation or by inbreeding, will lead to divergence of the non-communicating populations and, perhaps, speciation. Classical theory predicts that population divergence will be accelerated under conditions of inbreeding, leading to the fixation of alleles, including deleterious recessives, through genetic drift. It also indicates that genetic diversity can be maintained in diploid gene pools only when the heterozygote is the most fit genotype. This condition is readily achieved by associative overdominance in which loci may be rendered stably polymorphic by linkage to deleterious recessives (Zouros 1994), even in inbreeding populations. Thus, the occurrence of deleterious recessives, which may occur in large numbers and are often measurable in terms of lethal equivalents, may provide diploid sexuality, including inbreeding populations, with evolutionary advantage. However, these lethal equivalents result in a genetic load. The expression of the genetic load in a diploid sexual 221 Journal of the Royal Society of Western Australia, 80(3), September 1997 population will depend on the breeding system. In a normally outbreeding population, deleterious mutations are expected to be effectively invisible, being masked by the uncorrupted DNA of the homologous chromosome (diploid redundancy). Outbreeding populations are thus predicted to accumulate elevated levels of heterozygosity for deleterious mutations. If the population is forced to inbreed, this genetic load will be expressed as inbreeding depression due to the formation of elevated levels of deleterious recessive homozygotes. In a normally inbreeding population, on the other hand, mutations should be regularly exposed as homozygotes and purged out of the gene pool if they are sufficiently deleterious (Lande & Schemske 1985; Charlesworth et al. 1990). Thus inbreeders may be expected to be relatively free of seriously deleterious recessives, and exhibit little inbreeding depression and a reduced genetic load. Empirical evidence (Husband & Schemske 1996) suggests that in outbreeding plants, higher levels of inbreeding depression are exhibited at earlier stages in the life cycle, especially at seed formation, whereas with inbreeders, inbreeding depression is more likely to be expressed at later stages in the life cycle, if at all. Genetic system evolution in 1. petraea contradicts many classical expectations, particularly in its maintenance of high genetic loads in self-fertilising lineages. This paper reviews the genetic phenomena in I. petraea which led to this situation and documents the evolution, in response, of complex hybridity, a genetic system which effectively maintains genetic hybridity in highly inbred lineages. In addition, alterations to phenotypic characters, their patterns of variance within the population system, and their association with genetic system adaptations are discussed. Finally, the relevance of the observations made on l. petraea to patterns of variation in other native plant species is briefly reviewed. Inbreeding in I. petraea 1. petraea occurs in discrete isolated populations on granitic and other rocky outcrops of arid Australia. Due to this habitat preference, population size is often limited and subject to numerical fluctuation and genetic bottlenecks. These features alone would impose a measure of inbreeding upon the populations. In addition, populations may be characterised by different breeding systems. The primitive breeding system in /. petraea, and the Lobeliaceae generally, is one promoting outbreeding. Outcrossing may result when the stigma protrudes through the anther tube, dispensing the flower's pollen on the way, and then becomes receptive to pollen from other flowers. Self pollination, or autogamy, results when the style fails to elongate and the stigma remains enclosed within the anther tube, receiving pollen from the same flower Games 1965). The frequency of stigma protrusion may be taken as a measure of outcrossing potential. Because of population size, habitat preference and the possibility of pollination between flowers on the same plant, populations exhibiting high frequencies of stigma protrusion may still be relatively inbred. In populations where little or no stigma protrusion occurs, inbreeding approaches 100 per cent. Thus, all I. petraea populations are inbreeding to some extent, but some are much more inbreeding than others. The adoption of autogamy may be regarded as a means of reproductive assurance, and would be selected for where ovules would otherwise remain unfertilised (Jarne & Charlesworth 1993). In accordance with classical theory, inbreeding in I. petraea promotes genetic uniformity within and differentiation between populations. I. petraea Population Differentiation I. petraea populations differ with respect to their stigma protrusion frequencies (Table 1). While stigma protrusion frequencies show some variation according to environmental conditions, the population differences are largely retained whether the plants are grown in the glasshouse, garden or in their natural habitat. This demonstrates that the stigma protrusion frequency is genetically determined, to a large extent, and that the populations are genetically differentiated with respect to this character. Population divergence is also evidenced in flower colour and a number of metric characters such as leaf and flower dimensions, and the number of ovules per flower. Flower colour may vary from white, through varying shades of patterned pink to deep purple and blue. There is considerable variation between populations, and little within, so that populations are generally characterised by a single flower colour. Partitioning of variance in leaf, pedicel and flower dimensions into within plant, within population and between population components was performed dames 1978; SH James, unpublished data). The characters measured were not closely related to aspects of plant fitness. The analysis (Fig 1; only structural homozygotes considered at this stage) showed that for these characters, while comparable variance components were found at all levels as would be predicted by classical genetics for Table 1 Proportion of stigma emergence of plants from various populations of Isotoma petraea. Field measurements are given where possible. For plants from the same population, stigma emergence is usually higher in the more mesic garden and glasshouse environments than in the field. Measurements of stigma emergence other than those from the field will generally be an overestimate. Population Genetic System Stigma emergence Where Measured Rainy Rocks 711 1% Field Pigeon Rock 711/06 2% Field Iron Knob 711 6% Field Victoria Rock 711 32% Garden Merredin Peak ©14 40% Garden Boorabbin 711 42% Glasshouse 3 Mile Rock 711/010 56% Garden Moorine Rock ©12 79% Garden Bullabulling 711 91% Garden Gnarlbine Rock 711 100% Garden Yackyackine Rock 71 1 100% Field Disaster Rock 711 100% Field 222 Journal of the Royal Society of Western Australia, 80(3), September 1997 y/^ Within plants 7 II 7 II o Metric Character Variance Ovule Number Variance Figure 1. Partitioning of components of total variance for structural homozygote (711) and complex hybrid (©) populations of lsotoma petraea. The distribution of variance reflects genetic system and selective importance of the characters (see text). interbreeding populations, the greatest variance component was within populations. However, partitioning of variance for the average number of ovules per ovary (Kiew 1969) contrasted with that for the metric characters in that there was little variance within plants and by far the largest component was found at the between population level (Fig 1). This may be explained in that the number ovules produced may be expected to be a significant component of fitness and that it should be rigorously selected within populations. These observations generally conform with the expectation of differentiation between isolated populations. Inbreeding, the Distribution of Deleterious Alleles and the Accumulation of Lethals Half of the gametes produced by an organism heterozygous for a deleterious mutation will be free of the defective allele. On selfing, one quarter of the progeny will be homozygous for the mutation, one half heterozygous, and one quarter free of the mutation. Thus, sexual reproduction allows a self-fertilising heterozygote to remove deleterious alleles. However, providing the deleterious homozygote is only slightly inferior to the heterozygote and the other homozygote, it may persist in the population (Charlesworth et al. 1990). With stochastic effects associated with small population size and bottlenecks, the deleterious allele may actually become fixed in the population. It then cannot be removed without genetic communication from another population. Further deleterious mutations may similarly become fixed leading to the accumulation of debilitating mutations by a process analogous to Muller's ratchet (Muller 1964; Lynch & Gabriel 1990). In addition, if both of a pair of homologous chromosomes carry deleterious alleles, mutation-free gametes can only be produced by a recombination event occurring between the two heterozygous loci at meiosis. In I. petraea, recombination processes, which result in chiasma formation at meiosis, are strongly localised to the ends of the chromosomes. This means that recessive deleterious alleles occurring in pairs in the body of each chromosome cannot be removed in a strictly self- fertilising lineage and will accumulate (James et al. 1990; James 1992). Normally in /. petraea there are 7 pairs of homologous chromosomes which form seven bivalents (711) at meiosis. Each pair of homologous chromosomes is thus effectively a single supergenic locus with the two homologues being the two alleles (James et al. 1990; James 1992), each loaded with deleterious, even recessive lethal, mutations. For each supergenic locus, 50% of the selfed products will be heterozygous and 50% homozygous, so for 7 such loci only 1/27 of the progeny will display parental levels of heterozygosity. The genetic load, that is, the proportion of inviable or inferior progeny of varying degrees of homozygosity for deleterious mutations, will approach 127/128 as the deleterious mutation content of each supergenic allele approaches lethality. This load strongly diminishes the reproductive potential of inbreeding populations, a problem which generates a focus for natural selection and adaptive evolution. On the other hand, outbreeding populations handle their deleterious mutations much more effectively. Outcrossing facilitates the removal of deleterious alleles, even with chromosomes that show chiasma localisation. The lower frequency of homozygotes in outcrossing lineages, especially for newly mutated alleles, impedes any ratchet driven accumulation of deleterious alleles. The potential for genetic debilitation in populations with significant levels of cross pollination is therefore much lower or negligible, compared to that of strict inbreeders. 223 Journal of the Royal Society of Western Australia, 80(3), September 1997 Competitive Ability of 711 Material of Contrasting Levels of Hybridity The debilitating effects of inbreeding in I. petraea have been demonstrated in several unpublished dissertations (Cohen 1982; Playford 1987; SH James & N Cohen unpublished data). Cohen (1982) compared the performance in competition experiments of selfed and outcrossed progenies from populations characterised by different levels of inbreeding. Pairwise competition trials were established by sowing, intermixed into pots, two progeny arrays of different origins (Table 2). Progeny arrays were derived from 711 seifs, 711 intrapopulational crosses, 711 interpop ulational crosses and selfed complex hybrids (see below). Within pots there were generally a few clearly superior plants and a number of suppressed weaklings, and replicate pots were consistent. Almost all comparisons were highly determinate in their outcomes, with the superior genotype clearly apparent from both aerial biomass (Table 2) and number of plants. Where indeterminate results occurred, they were most likely in competitions involving progenies of the same hybridity class, and especially where interpopulational hybrids were involved. The mean competitive index calculable from the results may be viewed as a measure of the relative vigour of each progeny type. The most effective competitors were the progeny of crosses within 711 populations, and by far the poorest Table 2 Mean competitive index of progenies of differing hybridity classes, listed in decreasing order of competitive ability (see footnote). Mean Hybridity Class Progeny Array Competitive Index 711 Intrapop Cross 711 Gnarlbine Rock 0.91 711 Self 711 Gnarlbine Rock 0.77 711 Intrapop Cross 711 Victoria Rock 0.82 711 Intrapop Cross 711 Boorabbin Rock 0.76 Complex Hybrid Self The Humps 0.55 Complex Hybrid Self Mt. Stirling 0.56 Complex Hybrid Self Keokanie Rock 0.54 711 Interpop Cross Victoria Rock x Boorabbin Rock 0.46 711 Interpop Cross Yellowdine x Wargangering Rock 0.35 711 Interpop Cross Victoria Rock x Mt. Caudan 0.11 711 Self Boorabbin Rock 0.09 711 Self Victoria Rock 0.01 Twenty five seeds of each of 2 progeny arrays of contrasting hybridity levels were sown intermixed and directly into a 125 mm pot containing a standard potting mix. Seeds had previously been tested to ensure uniform germination rates. All plants greater than 20 mm in height were then harvested at the soil surface after 6-8 months of free competition, counted, weighed and identified using GOT, LAP, PGM or IDH assays (James et al. 1983) This process was performed in triplicate for each pair of contesting progenies for 3 progeny arrays at each of 4 hybridity levels in all pairwise comparisons (198 pots). At harvest the relative aerial biomass of each competing type in each pot was calculated as a decimal fraction and averaged for each group of three replicates. The mean competitive index for each progeny type was calculated as the average of the aerial biomass decimals- for all competitions involving the progeny array, but excluding within hybrid level comparisons. A similar index was calculated for relative numbers of plants; results closely mirrored those for aerial biomass and are not presented here. were selfed progeny of the 711 plants (Table 2). An exception was the selfed 711 Gnarlbine Rock progeny, which was rated third best of all the progenies. These findings may be related to the degree of outbreeding in the source populations as indicated by their stigma protrusion frequencies. Both the Victoria Rock and Boorabbin populations have a reduced incidence of stigma protrusion (Table 1). These populations are composed of highly inbred lineages, each carrying their own suite of accumulated deleterious recessives, many of them in homozygous condition. Intrapopulational crosses will have many of these deleterious recessives masked in heterozygosity and thus display considerable heterosis over seifs. Gnarlbine Rock plants are much more outcrossed (Table J), so that the lineages within this population may be expected to have accumulated fewer deleterious recessives. It is no surprise that self progeny from this population performed as well as intrapopulational crosses from other populations. They were, however, inferior to Gnarlbine Rock intrapopulational crosses, in keeping with expectations. The superiority in all cases (including Gnarlbine) of intrapopulational hybrid seedlings over selfed seed from the same population strongly suggests that competition amongst seedlings in natural seed beds would ensure the survival of only the most heterozygous types. Interpopulational crosses were somewhat unpredictable and exhibited competitive abilities generally inferior to 711 intrapopulational crosses, but superior to 71 1 seifs (apart from the Gnarlbine Rock seifs; Table 2). Different arrays of deleterious mutations would have accumulated in the various genetically diverged and reproductively isolated 711 populations, so that crosses between them would be expected to result in heterotic hybrids. However, the unpredictable performance of the interpopulational cross progenies as compared to the deterministic performance of other hybridity classes indicates that these hybrids combine non-coadapted genomes, are somewhat unstable, and may be expected to exhibit less than optimal performance. These results mirror previous work in which varying degrees of heterosis were generated in progeny of artificial interpopulational I. petraea hybrids (Beltran & James 1974). Dry weights of parental, FI and F2 material, grown without competition under constant experimental conditions were compared. In all cases, the FI yielded greater dry weight than the mid-parental value, and generally exhibited considerable heterosis over the poorer performing parent. In some cases the FI performed better than both parents. These results conform to expectations when viewed in terms of stigma protrusion. The poorest performing parents (Iron Knob, Rainy Rock) had very low stigma protrusion frequencies (Table 1) and were highly inbred, so it was not surprising that FIs performed much better than the seifs of these parents. The most outbred populations in the study (Yackyackine ["495"] and Disaster Rock) generally did not exhibit heterosis for dry weight accumulation when crossed to plants from other populations. We conclude that inbreeding populations of 71 1 /. petraea are somewhat debilitated by the accumulation and fixation of deleterious recessive mutations and they show significant levels of heterosis following both inter- 224 Journal of the Royal Society of Western Australia, 80(3), September 1997 and intrapopulational crossing. Inbreeding does not purge the deleterious recessives from the gene pool. In addition, a higher level of outbreeding in 711 /. pctraea populations leads to the incorporation of only low levels of recessive lethals into the gene pool, a minimum genetic load and little, though definite, inbreeding depression. These conclusions refine the generally accepted relationships between the breeding system, genetic load and inbreeding depression and are best understood in terms of the role of sexuality in genetic repair or cleansing. The Pigeon Rock Population The Pigeon Rock population is a very significant one in the evolution of I. petraea since it is in this population that complex hybridity evolved. Complex hybridity combines autogamy, a balanced lethal system and permanent hybridity for complexes generated by chromosomal interchanges (reciprocal translocations) (Darlington 1958; James 1965; Carson 1967; Holsinger & Ellstrand 1984). Pigeon Rock is composed of plants fitting two broad genetic types, with a number of lineages in each. Ancestral structural homozygote (711) and derived complex hybrids co-exist in stable proportions of about 1:2 respectively (James et al. 1990). Models for the evolution of the genetic system (James et al. 1990; James et al. 1991) propose that I. petraea on Pigeon Rock was initially 711 with an acquired autogamous habit (Table 1). This meant the population comprised essentially genetically isolated lineages, predisposed to the accumulation of deleterious mutations. The genetic load approached 127/128 as outlined above, with each supergenic allele carrying a lethal equivalent. In this genetic environment seed aborting lethals were generated, in the form of duplications and deficiencies, by chromosome segment transposition, probably mediated by transposable elements. Abortion of seed due to these deficiencies had the benefit of reducing allocation of maternal resources into incompetent, homozygous seeds. In Pigeon Rock complex hybrids, the initial seven supergenic loci have been reduced to five by coalescence of three of the bivalents into a single supergenic locus. During meiosis there are 4 bivalents plus a ring of six (®6) visible at metaphase 1. The ©6 consists of 2 complexes, N and S, each of whose 3 chromosomes are placed alternately in the ©6, and held in place by chiasmata at their homologous ends. The duplications described above facilitated the origin of complex generating interchages by enabling recombination between homologous segments in otherwise non- homologous chromosomes (James 1970; James et al. 1991). Each ©6 plant is an NS heterozygote which produces only N and S gametes. Gamete viability depends on regular segregation of the N and S chromosomes to opposite poles at anaphase-1 of meiosis. NN homozygotes are avoided since the N complex is not transmissible in pollen. SS homozygotes are eliminated by seed abortion, elaborated from the seed abortion system which existed prior to complex hybridity (James et al. 1991). The NS ©6 complex hybrid is superior to the 711 structural homozygote in that it potentially represents a 4-fold increase in the number of fully heterozygous progeny (1/25 compared to 1/27). The genetic load is thus reduced to 31/32. The proposed model for the evolution of complex hybridity is supported by a number of lines of evidence. Firstly, mathematical models suggest that, given the constraints seen at Pigeon Rock, complex hybridity will rise to the level of prominence seen there (James et al. 1990). Secondly, allozyme marker heterozygosity, which would also act as a marker for deleterious mutation Pigeon Rocks PR vs 3 Mile Rock ©10 Figure 2. Competition experiments involving Pigeon Rocks selfed 711, intrapopulational 711 x 711 crossed, selfed ©6 and selfed 3 Mile Rock ©10 progenies. Methods similar to those described in Table 2. 225 Journal of the Royal Society of Western Australia, 80(3), September 1997 heterozygosity, is noticeably greater in the 06 than the 711 (James et al. 1983). Thirdly, a series of competition experiments (Playford 1987) similar to those described above showed that the various levels of hybridity at Pigeon Rock were rated as might be expected from Table 2. Selfed 711 material from Pigeon Rock performed extremely poorly when pitted against its intrapopulational 711 x 711 rivals (Fig 2). The selfed 711 was inferior almost to the same extent against the selfed 06 (Fig 2), a finding of obvious importance for the establishment and maintenance of the genetic system at Pigeon Rock. 711 x 711 crosses were superior to the 06, but these would occur only very rarely, if at all, in the natural population and would not appreciably affect the 711:0 6 balance. Finally, preliminary results of a molecular phylogeny using Random Amplified Polymorphic DNA strongly support the hypothesis that the 0 6 originated from a Pigeon Rock 711 lineage exhibiting elevated seed abortion. Resolution of the mechanism of evolution of complex hybridity on Pigeon Rock is particularly instructive. It demonstrates circumstances in which recessive lethal genes which operate early in seed development may accumulate in inbred lineages, a somewhat unexpected and controversial phenomenon (cf Husband & Schemske 1996). Elaboration of Larger Ringed Complex Hybrids There are many populations of complex hybrid /. petraea, all extending in a south-westerly direction from Pigeon Rock (James 1965; James 1970). It is thought that a series of migrations occurred, starting with the Pigeon Rock ©6. Hybridisation of an immigrant ©6 complex hybrid from Pigeon Rock with plants of the invaded 711 population would generate heterotic hybrids which were also heterozygous for numerous duplications and deficiencies arising from the seed abortion systems in the Pigeon Rock parent. The duplications facilitated further interchanges and allowed the coalescence of the genome into fewer supergenic loci, while the deficiencies formed the basis of a new balanced lethal system. Thus, interpopulational hybridisation involving the Pigeon Rock 06 complex hybrid was able to generate new complex hybrids with larger rings in the recipient population. Migration of the new, larger ringed complex hybrid to the next 711 population led to further elaboration of the complexes so that the number of chromosomes incorporated into the ring progressively increased, in general, the further from Pigeon Rock. Usually each population is characterised by a single meiotic configuration and a very restricted array of genotypes (James 1970; James et al. 1983), indicating that a single, most fit complex hybrid type has displaced all others. Complex hybrid populations at the extreme of the genetic system's range are 014. The ©14s have a single supergenic locus and a genetic load reduced to 1 /2. Competition experiments have again demonstrated the feasibility of this model. Complex hybrids performed somewhat better than interpopulational 711 x 711 crosses, and markedly better than 711 seifs (Table 2). The 3 Mile Rocks ©10 out competed both 711 and ©6 material from Pigeon Rock (Fig 2) indicating that, at least in this case. the larger ring confers more vigour than the smaller one. The 010 was also victorious over Pigeon Rock 711 intrapopulational crosses (Table 2), which was perhaps not according to expectations, but which might be explained on the grounds of a generally high level of relatedness and genetic debilitation among the Pigeon Rock 7IIs. Complex hybridity is a highly specialised system and very stable once established in a population. Crosses between plants from different complex hybrid populations display negative heterosis in the FI (Beltran & James 1974), suggesting that a level of hybrid vigour or heterosis at least equivalent to that exhibited by 711 x 711 interpopulational crosses (see Table 2) is already conserved within plants. In addition, it may be expected that the two complexes associated in a naturally occurring complex hybrid are highly selected for mutual co-adaptation whereas complexes from different populations, which have not been mutually selected, are not coadapted and perform less well together. The selective advantage of complex hybridity has been explained, in the past, in terms of a "pursuit of hybridity" (Darlington 1939; James 1965; Carson 1967). Detailed study of the genetic properties and competitive abilities of natural populations of /. petraea distributed over the granite rocks of Western Australia has led to identification of the masking of deleterious mutations as the underlying basis of hybrid superiority and to a detailed appreciation of the pathways followed in this evolutionary pursuit. CD Darlington's concept of the "pursuit of hybridity" in evolutionary progress has always been controversial, but we can be certain that it is a real and powerful phenomenon. Adaptive Responses of Complex Hybrids The pursuit of hybridity in highly inbreeding populations of 7. petraea has led to the evolution of complex hybridity. In this genetic system, the seven pairs of chromosomes have been coalesced to form two complexes which are mutually coadapted alleles of a single supergene. These complexes are distributed from parent to offspring without change, for there is essentially no recombination within their chromosomal components, and their components cannot assort independently as is possible in the primitive seven bivalent forming types. Thus, recombinational capability, which is often taken to be the primary value of sexual reproduction in normal diploid sexual species, and the source of evolutionary capability, is highly suppressed in these derived complex hybrids. However, the complex hybrids exhibit quite striking evolutionary adaptations in important components of their reproductive biology (summarised in Table 3). First, the complexes for which the complex hybrids are heterozygous increase in size, from three chromosomes in the Pigeon Rock ©6, through to seven in the © 14s. The changes, from outbreeding to inbreeding through to the adoption of complex hybridity and increasing ring size, arc mirrored by corresponding changes in heterozygosity, genetic debilitation and competitive ability (Table 3). 226 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 3 Summary of Isotoma petraea breeding system and morphological adaptive responses to changes in genetic system. Primitive Forms Pigeon Rock Complex Hybrids Cytology 711 711 711 ©6 ©10-014 Stigma Protrusion High Low Very Low Very Low Variable Observed Heterozygosity Low Low Low Medium High Competitive Ability High Low Low Medium High Debilitation Low High High Low? None Expected Interpopulational Heterosis Low High n/a n/a Negative Hybrid Instability Yes Yes ? ? Yes Flower Size Large Large Large Medium Small Ovule Number Low Low Low Low High Seed Aborting Lethals Low Low Low-Medium Medium-High High, Early Second, complex hybrid plants are characterised by a smaller flower size than 7IIs (James 1982a), a trend evidenced even within the Pigeon Rock population where there is a distinct floral dimorphism dependent on the genetic system (Table 3). A reduction in flower size is commonly associated with autogamic pollination (Jarne & Charlesworth 1993) and presumably represents the loss of a function (attraction of pollinating vectors) where such a function is irrelevant. Third, the accumulation of chromosomes into interchange rings in complex hybrids results in meiotic irregularities and reduced gametic fertility (Darlington 1958; James 1970; Cleland 1972). Gamete fertility, 100% in 711 forms, is reduced to about 70% in 06s and about 20% in ©14s. Fully half of the selfed progeny from the remaining gametes would be homozygous for the complexes and therefore not viable, so the potential fecundity of the ©14s is about 10% that of the 711s. Despite this loss of fecundity, the proportion of fully heterozygous progeny in complex heterozygotes is still higher for all ring sizes than for 7IIs (James 1970). The loss of fecundity is compensated for, in complex heterozygotes, by a substantial increase in ovule numbers (Table 3). A study of seven 711 and sixteen complex hybrid populations (Kiew 1969) showed that the 711 plants had an average of 1130 ovules per flower, while the complex hybrids averaged 1545. The Pigeon Rock ©6 had an average ovule number of 1090, a figure close to the average for the 711 populations, and to that for alethal Pigeon Rock 7IIs (JD Bussell unpublished data). Fourth, the complex homozygote elimination system shows significant evolutionary development amongst /. petraea complex hybrids (Lavery & James 1987). The recessive seed aborting lethals which removed the SS complex homozygotes in the Pigeon Rock ©6s were relatively late acting so that the aborted seed exhibited a well developed but collapsed testa. As rings become larger, and as distance from Pigeon Rock increases, the time of seed abortion becomes earlier and earlier, so that in some large ringed forms, the aborted seeds are barely distinguishable from unfertilised ovules (Table 3). Earlier seed abortion allows fewer resources to be invested into incompetent seed. In some populations, homozygote elimination is based on complementary gametic lethality so that no resources are invested into incompetent complex homozygotes. Distribution and Generation of Variation The profound genetic system change, from relatively open sexuality to essentially invariant complex hybridity, might well be expected to be associated with a redistribution of variation within the population system (e.g. Darlington 1958; Jarne & Charlesworth 1993). This has been studied by examining the distribution of phenotypic variation within and between populations of /. petraea (Kiew 1969; SF1 James unpublished data). Patterns of variation were found to be dependent on the genetic system and on the selective importance of the character under examination, as detailed below. The distribution of variance for ovule number, a character directly linked to the reproductive capability and fitness of plants, is strikingly different in the 7IIs and complex hybrids (Fig 1). In the 7IJs, most variance for ovule number occurs between populations, there is little between plants and very little between flowers within plants. Amongst the complex hybrids, on the other hand, most variance for ovule number occurs between plants within populations - the level at which selection occurs. Complex hybrids also have a significant amount of variation within plants. The complex hybrids' uniformly higher ovule number suggests that directional selection for ovule number is very strong in complex hybrid populations and less so among the 71 Is . Presumably this elevated ovule number is a response to the considerably reduced probability of an ovule becoming a seed, as described earlier. In contrast, for metric characters (including leaf shape and the length of the peduncle) which have no obvious bearing on reproductive capability, the variance within the primitive 7IIs is distributed almost equally between populations, between plants within populations, and within plants (Fig 1). Amongst the derived complex hybrids, however, the variance between populations is relatively larger, and the variance between plants within populations, and within plants, is reduced. The expression of phenotypic attributes is a complex process determined by the interaction of many genes. Characters which show constant expression, and therefore low levels of within plant variation, are said to exhibit a high degree of canalization, or homeostasis. Characters exhibiting higher levels of within plant variation are determined by less well canalized gene systems. Evolutionary change of a homeostatic character 227 Journal of the Royal Society of Western Australia, 80(3), September 1997 can only be achieved when the canalizing gene systems are destabilised so that selectable variation becomes exposed in the population. In general, heterozygous genotypes tend to exhibit more homeostatic control than homozygous genotypes (e.g. Mitton & Grant 1984; Palmer & Strobeck 1986; Jarne & Charlesworth 1993). However, wide crosses, as well as inbreeding, may lead to the destabilisation of developmental processes (Clarke 1994; Freeman et al. 1994). These observations may explain how the complex hybrids, with generally optimum levels of enforced heterozygosity, have a relatively lower within plant variance for metric characters than the inbreeding 7IIs (which are more homozygous for a range of deleterious mutations), and yet have a higher proportion of within plant variance for ovule number than the 711s (Figure 1). In any population there is a fine balance between rigid control of developmental processes and maintenance of adaptive capacity. 7. petraea offers a superb resource for studies into the effect of varying degrees of inbreeding and homozygosity on developmental processes and phenotypic variability in plants. Despite abandonment of recombination often being described as an evolutionary dead end (Darlington 1958; Carson 1967; Wagner & Gabriel 1990), important selectable diversity may still be generated in I. petraea , even in complex hybrids. Perhaps, along with hybridisation and the vestiges of sexual recombination and segregation, other more primitive mechanisms are utilised for genetic diversification and genome reorganisation. In particular, mobile genetic element activity destabilises the genome, it may affect particular parts of the genome more than others, and it may be promoted by "genomic shock", where mutually uncoadapted genomes are combined in wide crosses. Conclusions The network of granite rocks characterising large areas of the Australian continent are of particular importance as museums for relict flora, fauna, and of evolutionary processes. They provide a living collection of discrete yet integrated natural evolutionary experiments which generally so far have been sampled in only the most cursory fashion. Resolution of the evolutionary pathway leading to complex hybridity in 7. petraea has provided a unique perspective on the consequences of inbreeding. Competition experiments reported herein, and previous work (Beltran & James 1974), have demonstrated distinct competitive advantages at critical stages in the proposed evolutionary pathway. At a more fundamental level, the 7. petraea story supports the theory that biparcntal sexual reproduction and meiotic recombination are primarily concerned with genetic repair. Impediments to these processes result in the accumulation of deleterious mutations, genetic load, inbreeding depression and debilitation of whole populations. Importantly, and contrary to classical expectations, inbreeding, and especially intense inbreeding, does not purge the genetic load. Load in l. petraea may be managed either by outbreeding or, in inbreeding populations, by the early elimination of incompetent homozygotes and strict control of recombination via complex hybridity. Interpopulational hybrids and the naturally occurring complex hybrids demonstrate considerable variation (decanalization) for morphological and reproductive characteristics and for competitive ability. These hybrids are comparable in many ways to the mutator genotypes generated by interracial hybrids of Drosophila (Thompson & Woodruff 1978), some of which induce mobile genetic element activity (Fontdevila 1988). It seems likely that mutator genotypes, generated by interfacing separately co¬ adapted gene pools, release an innovative capability which is more primitive than the genetic innovation associated with diploid sexuality, and which has been, and still is, of profound importance in organic evolution. This harks back to Darlington (1956) who, in discussing the nature of variation in natural systems, concluded that the genome's ability to generate biological capability and variability must be more important in enabling biological evolution than the environmental factors promoting it. The possible role of mobile genetic elements in mutator genotype activity in 7. petraea remains to be investigated. It is hoped that this paper has presented a picture of the granite rocks scattered throughout arid Australia, not so much as self-contained inselbergs to be managed and conserved as isolated units, but as networking subdivisions of species' metapopulations and as places of active, documentable evolutionary change. They are ideally suited, as museums of evolutionary processes, for the study of genetic behaviour, variation and evolutionary responses. 1. petraea has proved to be most illuminating in the study of these phenomena, and will no doubt continue to be so. 7. petraea has been instrumental in the development of an understanding of the presence and action of seed aborting lethal genes in many of Western Australia's signature floral genera, including Stylidium , Anigozanthos, Drosera, Laxmannia and Eucalyptus (reviewed by James 1982a,b, 1992, 1996; Burbidge & James 1991). Recently, Kennington & James (1997) have outlined a mechanism whereby recessive lethal and mildly deleterious genes may interact to prevent purging of load in eucalypt species. Study of 7. petraea has also helped elucidate notions of genomic coalescence as a potent evolutionary force (reviewed by James 1992). Pigeon Rock, set in the midst of superb, relatively undisturbed country, is of considerable importance as the putative origin of complex hybridity in 7. petraea. This inselberg warrants special conservation status, including removal of the present stockyards, exclusion from the surrounding pastoral lease and preclusion of vermin. In fact, we feel that Pigeon Rock warrants World Heritage listing. Acknowledgments : We thank N Cohen, I Playford, L P Kiew and R Tinetti for data currently residing in unpublished honours dissertations at the Department of Botany, University of Western Australia. Thanks are due also to M Waycott, J Kennington, and H Stace for discussions and comments on the manuscript. References Bell G 1988 Uniformity and diversity in the evolution of sex. In: The Evolution of Sex: An Examination of Current Ideas (eds R E Michod & B R Levin). Sinauer Associates, Sunderland, Massachusetts, 126-138. 228 Journal of the Royal Society of Western Australia, 80(3), September 1997 Beltran I C & James S H 1974 Complex hybridity in Isotoma petraea. IV. Heterosis in interpopulational hybrids. Australian Journal of Botany 22:251-264. Bernstein H, Hopf F A & Michod R E 1988 Is meiotic recombination an adaptation for repairing DNA, producing genetic variation, or both? In: The Evolution of Sex: An Examination of Current Ideas (eds R E Michod & B R Levin). Sinauer Associates, Sunderland, Massachusetts, 139-160. Burbidge A H & James S H 1991 Postzygotic seed abortion in the genetic svstem of Stylidium (Angiospermae: Styhdiaceae). The Journal of Heredity82:319-328. Carson H L 1967 Permanent heterozygosity. Evolutionary Biology 1:143-168. Charlesworth D, Morgan M T & Charlesworth B 1990 Inbreeding depression, genetic load, and the evolution of outcrossing rates in a multilocus system with no linkage. Evolution 44:1469-1489. Clarke G M 1994 The genetic basis of developmental stability. I. Relationships between stability, heterozygosity and genomic coadaptation. In: Developmental Instability: Its Origins and Evolutionary Implications (ed T A Markow). Kluwer Academic Publishers, Dordrecht, 17-25. Cleland R E 1972 Oenothera: Cytogenetics and Evolution. Academic Press, London. Cohen N 1982 Hybridity. Its effects on growth, variation & survivability in Isotoma petraea. Honours Thesis. Department of Botany, University of Western Australia, Perth. Darlington C D 1939 Evolution of Genetic Systems. Cambridge University Press, Cambridge. Darlington C D 1956 Natural populations and the breakdown i f classical genetics. Proceedings of the Royal Society, London, Series B 145:350-364. Darlington C D 1958 Evolution of Genetic Systems. Oliver & Boyd, Edinburgh. Fisher R A 1930 The Genetical Theory of Natural Selection. Clarendon Press, Oxford. Fontdevila A 1988 The evolutionary potential of the unstable genome. In: Population Genetics and Evolution (ed G de Jong). Springer-Verlag, Berlin, 251-263. Freeman D C, Graham J H & Emlen J M 1994 Developmental instability in plants: symmetries, stress and epigenesis. In: Developmental Instability: Its Origins and Evolutionary Implications (ed T A Markow). Kluwer Academic Publishers, Dordrecht, 99-121. Holsinger K E & Ellstrand N C 1984 The evolution and ecology of permanent translocation heterozygotes. American Naturalist 124:48-71. Husband B C & Schemske D W 1996 Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50:54-70. James S H 1965 Complex hybridity in Isotoma petraea. I. The occurrence of interchange heterozygosity, autogamy and a balanced lethal system. Heredity 20:341-353. James S H 1970 Complex Hybridity in Isotoma petraea. II. Components and operation of a possible evolutionary mechanism. Heredity 25:53-78. James S H 1978 Some causes of seed sterility in certain native Australian plants. Proceedings of the International Plant Propagators' Society 28:389-398. James S H 1982a Coadaptation of the genetic system and the evolution of isolation among populations of Western Australian native plants. In: Mechanisms of Speciation (ed C Barigozzi). Alan Liss, New York, 461-470. James S H 1982b The relevance of genetic systems in Isotoma petraea to conservation practice. In: Species at Risk: Research in Australia (ed R H Groves & W D L Ride). Australian Academy of Science, Canberra, 63-71. James S H 1992 Inbreeding, self-fertilisation, lethal genes and genomic coalescence. Heredity 68:449-456. James S H 1996 Seed abortion and the evolution of genetic systems in some Australian native plant groups. In: Gondwanan Heritage: Past, Present and Future of the Western Australian Biota (ed S Hopper, J Chappill, M Harvey & A George). Surrey Beatty & Sons, Chipping Norton, 90-99. James S H, Playford J & Sampson J F 1991 Complex hybridity in Isotoma petraea. VIII. Variation for seed aborting lethal genes in the ©6 Pigeon Rock population. Heredity 66:173- 180. James S H, Sampson J F & Playford J 1990 Complex hybridity in Isotoma petraea. VII. Assembly of the genetic system in the ©6 Pigeon Rock population. Heredity 64:289-295. James S H, Wylie A P, Johnson M S, Carstairs S A & Simpson G A 1983 Complex hybridity in Isotoma petraea V. Allozyme variation and the pursuit of hybridity. Heredity 51:653- 663. Jarne P & Charlesworth D 1993 The evolution of the selfing rate in functionally hermaphrodite plants and animals. Annual Review of Ecology and Systematics 24:441-466. Kenneally K F & Lowrie A 1994 Rediscovery of the presumed extinct triggerplant Stylidium merrallii (Styhdiaceae) with an amended description of the species and its conservation status. Western Australian Naturalist 19:269-277. Kennington W J & James S H 1997 The effect of small population size on the mating system of a rare clonal mallee, Eucalyptus argutifolia (Myrtaceae). Heredity 78:252-260. Kiew L P 1969 The occurrence, analysis and evolutionary exploitation of genetic variance in ovule number in Isotoma petraea. Honours Thesis. Department of Botany, University of Western Australia, Perth. Lande R & Schemske D W 1985 The evolution of self¬ fertilisation and inbreeding depression in plants. 1. Genetic models. Evolution 39:24-40. Lavery P & James S H 1987 Complex hybridity in Isotoma petraea. VI. Distorted segregation, gametic lethal systems and population divergence. Heredity 58:401^108. Lynch M & Gabriel W 1990 Mutation load and the survival of small populations. Evolution 44:1725-1737. Maynard Smith J 1988 The evolution of recombination. In: The Evolution of Sex: An Examination of Current Ideas (eds R E Michod & B R Levin). Sinauer Associates, Sunderland, Massachusetts, 106-125. Mitton J B & Grant M C 1984 Associations among protein heterozygosity, growth rate, and developmental homeostasis. Annual Review of Ecology and Systematics 15:479-499. Muller H J 1932 Some genetic aspects of sex. American Naturalist 66:118-138. Muller H J 1964 The relation of recombination to mutational advance. Mutation Research 1:2-9. Palmer A R & Strobeck C 1986 Fluctuating asymmetry: measurement, analysis, patterns. Annual Review of Ecology and Systematics 17:391^121. Playford J 1987 Components in the development of complex hybridity in Isotoma petraea. BSc (Hons) Thesis, Department of Botany, University of Western Australia, Perth. Thompson J N & Woodruff R C 1978 Mutator genes - pacemakers of evolution. Nature 274:317-321. Wagner G P & Gabriel W 1990 Quantitative variation in finite parthenogenetic populations: what stops Muller's ratchet in the absence of recombination? Evolution 44:715-731. Zouros E 1994 Associative overdominance: evaluating the effects of inbreeding and linkage disequilibrium. In: Developmental Instability: Its Origins and Evolutionary Implications (ed T A Markow). Kluwer Academic Publishers, Dordrecht, 37-48. 229 Journal of the Royal Society of Western Australia, 80:231-233, 1997 The ant communities of Sanford Rock Nature Reserve, Westonia, Western Australia A I Doronila & J E D Fox School of Environmental Biology, Curtin University of Technology, GPO Box U1987, Perth WA 6845: email: idoronil@info.curtin.edu.au Abstract This study investigated the ant communities present in the Sanford Rock Nature Reserve (325 mm mean annual rainfall, 300 km East of Perth), an 800 ha Reserve comprising a complex mosaic of granite outcrop and tumbled boulders with surrounding flat areas carrying woodlands or shrublands. Nine locations in the reserve and two others in a eucalypt woodland 9 km away were sampled with pitfall traps and hand collection during March of 1994-1996. A total of 87 ant species from 22 genera was recorded. Three genera, Iridomyrmex (22 species), Camponotus (12 species) and Melophorus (11 species) represented 51.7% of the total species composition, which is typical of ant communities in arid and semi-arid Australia. Ant species richness (>30) was greatest in Eucalyptus woodlands with mixed shrub understorey species. The Callitris and Acacia low shrub thicket were represented by 18 and 21 species respectively. Thickets with Allocasuarina or Eucalyptus crucis had the lowest species richness (< 15). These thickets were in deep soil pockets which were moist and contained large amounts of litter. Hot climate specialist ants were poorly represented in the species poor areas. Introduction Westonia is a small agricultural and gold mining district 285 km east-north-east of Perth, Western Australia (31° 40' S, 118° 35' E). The district has been extensively cleared for wheat growing and sheep grazing and the nearest piece of remnant vegetation is 9 km north-east of the town at the Sanford Rocks Nature Reserve (800 ha). The reserve is in very good condition as it has been fenced off from cattle and sheep grazing and there have been no fires in nearly all of the reserve since at least 1927 (Muir 1979). This area experiences a Mediterranean climate typical of the wheatbelt regions. Summer thunderstorms occur occasionally but most of the rainfall occurs during winter. The average annual rainfall is 325 mm. The role of fauna within mature and developing ecosystems situated on previously disturbed land is often neglected. There is a paucity of long-term studies on fauna recolonization (Majer 1989). Ants are one of the most important invertebrate groups in Australia especially in seasonally arid regions (e.g. Westonia). They are useful as bio-indicators of change in the environment (Andersen 1990; Majer 1983). Ant community structure is influenced by interaction with their habitat, mainly with the vegetation complex (Andersen 1986). The aim of this study is to assess the ant communities present on several of the vegetation communities and to determine the factors influencing the patterns of their community organisation. © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 Methods Site selection Ants were sampled annually from 1994-1996 on eleven transects. Nine sites were sampled in the reserve. These were; (Eg) grass-eucalypt woodland boundary Eucalyptus capillosa woodland with sparse herb understorey of Amphiphogon strictus; (Sg) salmon gum woodland with E. saltnonophloia, £. sheathiana and £. myriadena with an Acacia and chenopod understorey; (Th) Acacia scrub thicket on rock edge; (Gw) £. salubris and salmon gum, E. saltnonophloia woodland with a sparse chenopod understorey; (Al) Allocasuarina huegeliana thicket; (Yg) £. loxophleba, E. myriadena and E. saltnonophloia woodland with a sparse understorey of Acacia , Melaleuca, Bon/a sp; (Cx) E. crucis thicket with dense understorey of Kunzea and Calothamnus; (Ov) overhang with mixed Acacia and Myrtaceae shrubs; and (Ct) Callitris columeliaris thicket. Two other locations were sampled 9 km away from the reserve. The disturbed site (Ewd) is situated at the eastern end of the Westonia town-site where the hospital was situated 70 years ago. It is a Eucalyptus woodland of 15-20 m in height. The upper canopy is dominated by E. longicornis and E. salubris with a sparse understorey of Acacia hemiteles and chenopod shrubs. The undisturbed woodland (Ewu) is situated at the town commons. The vegetation is a Eucalyptus woodland dominated by E. salubris and the understorey is a dense association of Acacia and Melaleuca species. Sampling methods Ants were sampled using pitfall traps (1.8 cm diameter Pyrex®) which were partly filled with a 30:70 glycerol : alcohol mixture). Each location was sampled with a transect of ten traps with 10 m spacing and operated for a 1 week period during mid-March. Hand collections 231 Journal of the Royal Society of Western Australia, 80(3), September 1997 Table 1 The number of ant species collected in the study at Sanford Rock Nature Reserve. Area1 Sanford Rock Nature Reserve Non reserve Thickets Eucalypt woodland Cx Ov A1 Th Ct Eg Gw Yg Sg Ewu Ewd Myrmeciinae M yrmecia 1 1 1 Ponerinae Bothroponera 1 1 1 RJiytidoponera 2 1 1 3 3 3 2 3 3 4 Odontomachus 1 1 1 1 Myrmicinae Crematogaster 1 1 Aphaenogaster 1 1 1 Cardiocondyla 1 Monomorium 1 2 5 4 2 6 1 7 6 4 6 Meranoplus 1 1 2 2 1 Pheidole 2 1 2 1 2 Tetramorium 3 1 1 2 1 3 Podomynm 1 1 Pseudomyrmecinae Tetraponera 1 1 1 Dolichoderinae Iridomyrmex 4 5 1 2 5 6 10 6 9 12 8 Papyrius 1 1 Formicinae Calomynnex 1 1 1 1 1 1 Camponotus 2 1 2 4 5 5 7 8 6 7 Melophorus 1 2 5 3 7 7 5 8 3 7 Notoncus 1 Ophistopsis 1 Polyrachis 1 1 1 1 1 Stigmacros 1 TOTAL SPECIES 9 11 13 18 21 31 32 33 45 39 43 1 ants listed by subfamily (bold) and genus (italic). were also performed for uniform periods during the day. Ants were sorted to genera and a number or letter code was allocated to each specimen which applies only to this study. The vegetation composition, cover, density, percent leaf litter cover and percent bare ground were measured at equidistant intervals along each transect. Results and Discussion A total of 87 ant species from 22 genera was recorded (Table 1). Three genera, Iridomyrmex (22 species), Camponotus (12 species) and Melophoms (11 species) represented 51.7% of the total species composition which is typical of ant communities in arid and semi-arid Australia (Greenslade 1976; Andersen 1983). Classification of ants into functional groups (Andersen 1990; Greenslade 1979) shows that dominant ant species ( Iridomyrmex ) contributed 20-30% to the diversity of the communities with high species richness (Fig 1). Dominant species also contributed > 60% of the total catch where they were present in high frequencies. Ant species richness (>30) was greatest in Eucalyptus woodlands with mixed shrub understorey species. The Acacia low shrub and Callitris thicket were represented by 18 and 21 species respectively. Thickets with Allocasuarina or Eucalyptus crucis had the lowest species Location □ Dominant J Subordinate ■ Opportunist □ Generalized myrmicines ] Climate specialist □ Cryptic □ Large solitary Figure 1. Classification of ant communities present in the 11 sites by functional groups. 232 Journal of the Royal Society of Western Australia, 80(3), September 1997 richness (< 15). These thickets were in deep soil pockets which were moist and contained large amounts of litter. Hot climate specialist ants e.g. Melophorus and Meranoplus species were poorly represented in the species-poor areas. Ordination of the presence and absence data of ants in each transect was performed, and axes 1, 2 and 3 represented 24.6, 14.8 and 13.1% of the total variance, respectively. Three main groupings could be identified from the ant assemblages; • sites with eucalypt woodlands but with two identifiable subgroups of the reserve and non-reserve areas; Callitris thicket appeared to be similar to these sites; • the Acacia scrub and Allocasuarina thickets were similar; and • the £. crucis and overhang area were similar. Factors which appeared to be important to ant species richness were tested. A negative correlation was observed between the number of ant species and percentage litter cover (r = -0.559) and vegetation cover (r = -0.542). Many factors such as soil moisture, depth of litter, den itv of ground vegetation and soil type determine tc availability and suitability of nest sites and foraging behavior. These factors consequently affect the diversity of ants in an area (Greenslade 1979). Acknowledgments: We thank the students involved in the Terrestrial Ecology 301 field trips who participated in various aspects of the study. References Andersen A N 1983 Species diversity and temporal distribution of ants in the semi-arid mallee region of north-western Victoria. Australian Journal of Ecology 8:127-137. Andersen A N 1986 Diversity, seasonality and community organization of ants at adjacent heath and woodland sites in south-eastern Australia. Australian Journal of Zoology 34:53-64. Andersen A N 1990 The use of ant communities to evaluate change in Australian terrestrial ecosystems: a review and a recipe. Proceedings of the Ecological Society of Australia 16:347-357. Greenslade P J M 1976 The meat ant Iridomyrmex purpureus (Hymenoptera: Formicidae) as a dominant member of ant communities. Journal of the Australian Entomological Society 15:237-240. Greenslade P J M 1979 A Guide to the Ants of South Australia. South Australian Museum, Adelaide. Majer J M 1983 Ants: bio-indicators of minesite rehabilitation, land-use and land conservation. Environmental Management 7:375-383. Majer J M 1989 Animals in Primary Succession: The Role of Fauna in Reclaimed Lands. Cambridge University Press, Cambridge. Muir B G 1979 Some nature reserves of the WA wheatbelt. Part 14 Westonia Shire. Department of Fisheries and Wildlife, Perth. 55pp. 233 Journal of the Royal Society of Western Australia, 80:235-237, 1997 Acacia williamsiana (Fabaceae: Juliflorae): A new granitic outcrop species from northern New South Wales J T Hunter Department of Botany, University of New England, Armidale NSW 2351 email: jhunter@northnet.com.au Abstract Acacia williamsiana J.T. Hunter a new rare species (2RCa) restricted to granitic outcrops within the New England Batholith is described. This species is known from four disjunct localities within the North Western Slopes botanical division in New South Wales. Introduction During the compilation of a preliminary checklist of the flora of Kings Plains National Park on the north western slopes of New South Wales by John Williams, an unidentifiable specimen of Acacia was collected. This taxon was collected again from Kings Plains National Park by the author during an extensive survey of the granitic outcrop flora of the New England Batholith. Further collections of material that match this taxon were made during this survey on granitic outcrops within Severn River Nature Reserve and the proposed Kwiambal National Park. Subsequently a population was also discovered on a granite outcrop by the author, during an inspection of a new fire trail within the Torrington State Recreation Reserve. Acacia williamsiana at all localities was restricted to granitic outcrops or around the base of outcrops. This taxon often dominated habitats where it was found, sometimes forming monospecific stands. Materials and Methods This study is based on examination of materials collected by the author and others at all known localities of this taxon. All measurements were made from dried specimens. Taxonomy Acacia williamsiana J.T. Hunter, sp. nov. Acacia williamsiana , species nova similis A. bidgaensis Tind. & Stuart J. Davies a qua ovarium glabris, pedunculus brevissimus, et juvenus phyllodium glabris. Typus : New South Wales: North Western Slopes: on the banks of Kings Plains Ck, in Kings Plains National Park, north-west of Glen Innes, J.T. Hunter 4122 & P.J. Clarke, 1 November 1996 ( holo : NSW; iso : AD, BRI, HO, MEL, NE, PERTH). Tall shrub to small tree 2 — 8 m tall, often spreading © Royal Society of Western Australia 1997 Granite Outcrops Symposium, 1996 widely when young then becoming fruticose and eventually erect; Bark fissured and flaky, grey to dark grey-brown; branchlets angular, flattened to ellipsoid in section, glabrous, ± glaucescent, yellow-orange to red- brown. Phyllodes borne singly, diphyllous; phyl lodes on young plants broad elliptic to obovate, sometimes slightly falcate, often held horizontally, 1.3 — 7.5 cm long, 1.3 — 2.5 cm wide, pale green almost glaucous in appearance, glabrous; longitudinal nerves 60—100 per phyllode, with 3 — 5 more prominent; apex obtuse with a broad ± oblique callous mucro; mature phyllodes oblanceolate, linear, elliptic to narrow elliptic, rarely slightly falcate, held erect, 4.5 — 12 cm long, 0.4— 1.1 cm wide, pale green to subglaucous, glabrous or rarely pilose and glabrescent; longitudinal nerves 20 — 60 per phyllode, with 3 — 5 more prominent; apex acute with a distinct ± oblique callous mucro, 0.5 — 3.5 mm long; gland 0 or 1, obscure ± hairy, 0 — 0.5 mm from pulvinus, orifice elliptical; pulvinus 0.5 — 3 mm long, wrinkles transverse, pale; stipules 2, 0.3 — 1 mm long, triangular, often caducous, ± pilose. Inflorescences spicate, borne in pairs one on either side of the phyllodes, sessile, initially curved, 1.5 — 4 cm long in fruit, glabrous or rarely hairy basally, >150 flowers per spike, pale yellow; bracts 2, glabrous, pilose or pubescent, 1.2— 2.5 mm long, 1—1.7 mm wide, ± caducous; pedicels lacking; buds broad ovate; stipe 0.5 — 1.5 mm long in fruit; bracteole 0.7 — 0.9 mm long, pale brown, distinctly clawed, the lamina perpendicular to the claw, tomentose apically and abaxially. Flowers 5-merous; calyx fused, tube 0.7 — 1.2 mm long, lobes minute ca. 0.1 mm long, silver woolly to tomentose, more so basally; corolla valvate, fused about half way or more, tube 1 — 1.5 mm long, lobes 0.8 — 1.2 mm long, glabrous, minutely clawed; stamens with filaments 1.7 — 2.5 mm long; anthers versatile, 0.1 — 0.15 mm long; ovary ellipsoid, 0.2— 0.5 mm long, 0.2— 0.4 mm wide, green, hairy, viscid; style 1—2.5 mm long, stigma minutely lobed. Legumes linear, falcoid but not curled, ± constricted between seeds, 3.5 — 9 cm long, 2 — 4 mm wide, brown, ± glaucescent, wrinkled, margins thickened, turgid when young and fresh becoming flat. Seeds 6—12 arranged longitudinally in legume, ellipsoid, 0.6—1 mm long, 0.3 — 0.6 mm wide, dark brown to black; funicle filiform, folded 2 — 3 times. Fig. 1. Specimens examined. New South Wales: North Western Slopes: Kings Plains Ck, Kings Plains National Park, /. T. Hunter 1866 & V.H. Hunter, 27 April 1994 (NE); 235 Journal of the Royal Society of Western Australia, 80(3), September 1997 Figure 1. A — H, Acacia zvilliamsiana. A, adult branch with flower buds. B, young branch. C, adult branch with flower buds. D, venation pattern of phyllode. E, pulvinus and gland. F, adult branch with pods. G, flower. H, inflorescence. Scale bars; A, B, C, F = 4 cm; D = 7 mm; E = 4 mm; G = 2 mm; H = 13 mm. Drawn from J.T. Hunter 3933 (A); J.T. Hunter 2888 (B); J.T. Hunter 2934 (C, D, E); J.T. Hunter 4113 (F); J.T. Hunter 4114 (G, H). Drawn by D Mackay. 236 Journal of the Royal Society of Western Australia, 80(3), September 1997 Kings Plains, 1 km N of Falls, 50 km NW of Glen Innes, J.T. Hunter 2888 , 9 March 1995 (NE); 10 km W of Torrington on the Buther Rd, J.T. Hunter 3931 — 3935, 4 May 1996 (NE JTH 3931; BRI JTH 2932; NSW JTH 3933; MEL JTH 3935); 29.3 km from Emmaville on the Gulf Rd, C. Nano & /.£>. Williams s.n., 13 December 1996 (NE); Severn River Nature Reserve, NE of Glen Innes, near Pindari Dam, J.T. Hunter 3112, 13 June 1995 (NE); Severn River Nature Reserve, NE of Glen Innes, near Pindari Dam, J.T. Hunter 3142, 14 June 1995 (NE); McIntyre Falls, Kwiambal National Park, W of Ashford, J.T. Hunter 4114 & P.J. Clarke , 1 November 1996 (BRI, MEL, NE, NSW, PERTH); McIntyre Falls, Kwiambal National Park, W of Ashford, J.T. Hunter 4113 & P.J. Clarke, 1 November 1996 (NSW). Distribution. This species has been found in four disjunct localities on the north western slopes of New South Wales; Kings Plains National Park 40 km north¬ west of Glen Innes, Severn River Nature Reserve 50 km north-west of Glen Innes, Torrington State Recreation Reserve 60 km north-north-west of Glen Innes and the proposed Kwiambal National Park and adjacent Severn State Forest 30 km west of Ashford. Habitat. Found in open and exposed situations in shrubland and low woodland on and around the base of granite and porphyry outcrops between 280 — 1100 m altitude. The mean annual rainfall in these areas is between 640 — 800 mm. This taxon is found within crevices of bare rocky slopes and shallow soil surrounding the bare slopes. This species often forms thick stands when mature, and can dominate or co¬ dominate communities in which it is found. Associated species include Callitris endlicheri (Pari.) F.M. Bailey, Eucalyptus prava L. Johnson & K. Hill, £. caleyi Maiden, £. dealbata A. Cunn. Ex Schauer, Allocasuarina iiwphloa (F. Muell. & Bailey) L.A.S. Johnson, A. brachystachya L.A.S. Johnson, Micromyrtus grandis J.T. Hunter, and Astrotricha roddii Makinson. Flowering and fruiting. September to December Conservation status. A ROTAP code of 2RCa (Briggs & Leigh 1988) is suggested. This species has a disjunct distribution and is known only from four localities and from a restricted habitat, on and around the base of granitic outcrops. Therefore this species is considered to be rare. All known populations are within some form of reservation under the control of the New South Wales National Parks and Wildlife Service. Etymology. The specific epithet is in honor of Mr John B. Williams former lecturer and now honorary fellow at the University of New England, Armidale, NSW. Mr Williams has worked extensively in the north¬ east of New South Wales, particularly on granite flora and was the first to find and note this species as a potential new taxon. Notes. This taxon has a diverse appearance both in terms of habit and phyllode form. Comparatively young plants have a divaricate habit with phyllodes that are short and broad (Fig 1). Adult plants have a tall and straight habit with long thin phyllodes that are held erect. Forms at intermediate stages are often present. This species is killed by fire and seedlings appear readily afterwards. Seedlings, however, do not seem to be present unless a disturbance such as fire has occurred. Consequently individual populations are usually the same cohort grown after single wildfire events. Only where fire has passed differentially over outcrops are mixed aged populations found. The affinities of this species are unclear at this stage. However, it is possible that they are near A. bulgaensis Tind. & Stuart J. Davies and A. diphylla Tind. Conclusion Acacia williamsiana is a distinct species endemic to granitic outcrops of the New England Batholith. Other members of the Juliflorae growing on the New England Batholith are granitic outcrop endemics, such as A. pycnostachya F. Muell. and A. pubiflora Pedley. Others of the Juliflorae while not restricted to outcrops, are commonly found on them. Further work on the phylogenetic and biogeographical relationships of this species and its relatives is warranted. Acknowledgments : The author wishes to thank the director of the NSW National Parks and Wildlife Service for allowing collection of materials within service areas. B Briggs is thanked for allowing access to specimens at NSW. Thanks also to J B Williams for discussions, the staff of the Glen Innes office of the NSW National Parks and Wildlife Service for their assistance, D Mackay for illustrations, and J J Bruhl and P J Clarke for supervision. The author acknowledges the receipt of an Australian Postgraduate Award scholarship. References Briggs J D & Leigh J FI 1988 Rare or Threatened Australian Plants. Australian National Parks and Wildlife Service, Canberra. Special Publication 14. 237 Journal of the Royal Society of Western Australia, 80:239-247, 1997 History and management of Culham Inlet, a coastal salt lake in south-western Australia E P Hodgkin 86 Adelma Road, Dalkeith, WA 6009 email: ehodgkin@cygn us.uzoa.edu.au Manuscript received August 1996; accepted Map 1997. Abstract When Culham Inlet was first flooded by the Holocene rise in sea level it was an estuary, but in historic times it has been a salt lake closed by a high sea bar. It is in an area of low rainfall and episodic river flow and sometimes all water is lost by evaporation to below sea level. With above average rainfall in 1989 and 1992, high water levels in the Inlet flooded farm paddocks and threatened to break the bar and a road along it from Hopetoun to the Fitzgerald River National Park. In 1993 the bar was breached to release flood water, and the Inlet was briefly an estuary. Engineering measures designed to restore road access and prevent flooding are examined for their potential to restore the Inlet to its pre-1993 condition of a productive ecosystem. Recent clearing in the catchments of Culham Inlet and nearby estuaries in the south coast low rainfall area has increased river flow to them and appears to have caused their bars to break more frequently. Introduction In historic times Culham Inlet has been a coastal lagoon on a semi-arid part of the south coast of Western Australia (Fig 1), separated from the sea by a high bar (Fig 2) that is only known to have broken naturally once, in 1849. The bar was broken artificially in 1920, but for over 70 years since then the Inlet has absorbed river flow without the bar breaking, until 1993. Culham Inlet was a © Royal Society of Western Australia 1997 239 Figure 1. Culharn Inlet; the catchment with 400 mm isohyet (left) and plan of the Inlet (right). Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 2. Culham Inlet bar/dune area, January 1990; water level at 2.5 m AHD. Scale 1:8000 Photograph from the Department of Land Administration, Perth; Copy Licence 509/96. salt lake where water depth varied from 5 m to none, depending on episodic flow from two saline rivers and evaporation. Ecologically, it supported a restricted range of 'estuarine' fauna, sometimes with an abundance of a few species of fish and large numbers of waterbirds (Hodgkin & Clark 1990). A road along the bar gave access from Hopetoun (population 206, 1991 Census) on the east to the Fitzgerald River National Park to the west, but in 1989 and 1992 floods closed the road and inundated nearby low-lying paddocks. The town's growing tourist industry suffered and there was social and economic pressure to ensure reliable road access and prevent flooding. In May 1993 the high water level in the Inlet again caused flood¬ ing and threatened to breach the bar. An attempt to release flood water without the bar breaking failed; it broke and the Inlet was effectively an estuary for a few weeks until the breach closed. Several engineering measures to rebuild the road and prevent flooding were proposed, and one was implemented in 1996-97. The unpredict¬ ability of rainfall, the limited data on river flow, and the probability that recent clearing in the catchment has con¬ siderably increased river flow to the Inlet all make it dif¬ ficult to assess the environmental consequences of the various proposals to rebuild the road and the extent to which the Inlet can now revert to being a productive coastal lagoon. The Environment Culham Inlet is an 11 km2 lagoon about 1 m deep below mean sea level (Fig IB). Seven km of the Phillips River are scoured in places to 4 m below sea level, and there is a river delta at about sea level. All levels quoted here are relative to Australian Height Datum (AHD) which is 0.089 m above MSL (an accurate tidal datum was established in January 1990). The mean daily tide range is 0.67 m (MHHVV to MLLW) at Hopetoun. Inlet water level varies greatly with river flow and evaporation, from +4 m to -1 m (no water). The salinity of lagoon water varies from 10 to >70 ppt. Tributary river water is seldom <5 ppt, and stagnant river pools can be hypersaline to sea water (Hodgkin & Clark 1990). The barrier between the lagoon and the sea is a 1 km long bar, now with a dune built on it (Fig 2). The western part is a 10-15 m high dune, 100 m wide, with trees and shrubs. The eastern 400 m is 4 to 5 m high, 40-60 m wide, and sparsely vegetated. Catchment rainfall ranges from 500 mm at the coast to 350 mm inland (Fig 1A). The only long-term (100 years) rainfall data are for Ravensthorpe and Hopetoun, both just outside the catchment Winter rainfall is relatively reliable with 60-70% falling in the 6 months May to October (Fig 3), but the summer rainfall means are boosted by infrequent heavy falls (>100 mm). There can 240 Journal of the Royal Society of Western Australia, 80(4), December 1997 mm Rainfall mm Rainfall ■ MEAN □ MEDIAN Figure 3. Mean and median monthly rainfall at Ravensthorpe and Hopetoun (1901-1993). Data from the Bureau of Meteorol¬ ogy, Perth. be episodic 2-3 day falls of 50 to 100 mm in most months. A 50 mm fall, such as that of 6-8 October 1992, was estimated to recur every 7 years on average and that of 28-29 May 1993 every 2.5 years (Anon 1993a). Mean annual pan evaporation is 1754 mm (at Esperance), but surface evaporation from the Inlet is probably only 85% of pan evaporation, as found in Peel Inlet by Black & Rosher (1980) and in reservoirs by Hoy & Stephens (1979). An estimate based on water level data in Culham Inlet and rainfall at Hopetoun (1990-1994) indicates that evaporation from the Inlet was about 1100 mm over the six summer months from November to April. Two rivers flow to the Inlet (Fig 1), Phillips River with a catchment of 2100 km2 and Steere River with 485 km2. The Phillips catchment rises from sea level to about 300 m with extensive areas of sandplain and Precambrian rocks and a narrow belt of Quaternary coastal deposits (Thom et al. 1977; Thom & Chin 1984). Vegetation is predominantly mallee in inland areas and mallee-heath towards the coast (Beard 1972, 1973). The rivers are not gauged, but effective annual flow probably varies from zero to >50 x 106m3, most of it as episodic events over a few days following heavy rain, as shown by the record of water level in the Inlet kept by R Cooper (Figs 4 & 5). The flow pattern is similar to that in the Pallinup River (Fig 6), the nearest river with a long gauged record. mm Rainfall 150 100 50 1989 Ravensthorpe Hopetoun 1990 BJU. 1991 bhH all 1992 1993 R 466 H 603 R 527 H464 R 353 H 327 R 677 H 664 1989 1990 1991 1992 1993 \ - 1 - 1 - 1 - i i i Q 3 X ° < 0 Figure 4. Top. Monthly and annual rainfall at Ravensthorpe and Hopetoun, 1989-1993 (mean annual rainfall: Ravensthorpe 424 mm, Hopetoun 506 mm). Data from the Bureau of Meteorology, Perth. Bottom. Water level in Culham Inlet 1989-1993. Data from R Cooper. 241 Journal of the Royal Society of Western Australia, 80(4), December 1997 O X < a) > o Q CO CO CD The volume of river flow also varies greatly with soil moisture in the catchment. Nearly 300 mm of rain from February to July 1992 caused little flow from the dry catchment and no rise in Inlet water level (Fig 4). However that rain saturated the catchment, and in August to early October a further 300 mm of rain caused a rise in water level of 3.4 m. Only about 50 mm of rain on two days in October caused a rise of 1.5 m (about 15 x 106 m3 of river flow). The capacity of the Inlet to accept such flows without the bar breaking depends on the initial water level in the Inlet, which was only 0.5 m in July 1992 but was already 3.3 m in May 1993 (Fig 5). The area of cleared catchment (estimated from aerial photographs) increased from about 10% in 1968 to 40% by 1988. There is no record of river flow to Culham, however mean annual flow is estimated to be 3.5 x 106 m3 and to have increased from 1.3 x 106 m3 before clearing (data from Surface Water Branch, Water and Rivers Commission, Perth), and it may continue to increase for some time. The loss of deep-rooted vegetation, reduced moisture percolation from soil compacted by stock, and rising groundwater levels all contribute to increased runoff from the semi-arid catchment. The Biota Before the bar broke in 1993, Culham Inlet was periodically a highly productive ecosystem. Commercial fishers took large catches of black bream (Acanthopagrus butcheri) in seasons when the water level remained relatively high and salinity was <45 ppt (32 tonnes in 1990-91; 61 t in 91-92; 77 t in 92-93; Fisheries Department, CAESS, Perth) and recreational fishers also took large numbers of bream. A goby (Pseudogobius olorum) and two hardyheads (Atherinosoma elongata and Leptatherina wallacei) were common. There were large waterbird populations with 25 recorded species. The few euryhaline estuarine species of macro-invertebrates were abundant: the encrusting tubeworm Ficopomatus enigmaticus and the false mussel Fluviolancitus subtorta (Trapeziidae), a few polychaete worms, two amphipods and midge larvae (Hodgkin & Clark 1990). After the brief period while the Inlet was tidal, additional estuarine- marine invertebrate species (Ccratonereis sp, Spisula trigonella) were found to be common and a few opportunistic species of fish were caught, notably sea mullet Mugil cephalus and herring Arripis georgianus (Bennett & George 1994). The Inlet has a narrow fringe of salt-tolerant flora dominated by the paperbark Melaleuca cuticularis backed by coastal moort (Eucalyptus platypus var heterophylla), with yate (E. occidental^ var occidentals), Acacia cyclops, and Eucalyptus tetragona on higher ground (Bennett & George 1994). The paperbarks and coastal moorts along the eastern shore of the Inlet died following the prolonged 1989-90 flooding with saline water (about 17 ppt). The previously limited areas of samphire ( Sarcocornia quinqueflora) have expanded greatly while the water level has been low since 1993. The aquatic Ruppia megacarpa was present but seldom abundant. History of Culham Inlet Before the post-glacial rise in sea level, Culham Inlet would have been a valley with a river flowing through it; there is about 25 m depth of sand below present sea level at the bar (Anon 1993b). By 6500 years BP the rising sea level had flooded the Inlet, at least to its present level, and it remained an open estuary until about 3500 BP as evidenced by the abundant sub-fossil fauna of estuarine- marine molluscs dated 3660 ± 185 BP (Hodgkin & Clark 1990). With only episodic river flow from the semi arid catchment the bar built to a height at which it retained river flow, perhaps at first with periodic breaks as now at 'normally closed' estuaries such as Hamersley Inlet (Fig 1), before becoming a 'permanently closed' estuary (Lenanton 1974; Hodgkin & Lenanton 1981) i.e. a coastal salt lake. The western bar, now with a high dune, may have closed first and the eastern low bar later. The bar is known to have opened naturally in 1849 (Gregory 1849) and is said to have broken in the 1870s. It was opened artificially in 1918 or 1920 when floods threatened to break it during eight years of above average rainfall (1913-20), and it is reported to have remained open for about three months. The break appears to have opened a 200 m wide gap through the eastern low bar where sea water is seen seeping into the almost dry Inlet in an aerial photograph of January 1981 (Fig 7). Above average rainfall in 1955 (** 660 mm) again threatened the bar, and water is reported to have 'trickled out'. By September 1988, runoff from above average winter rain had raised the Inlet water level to about 3 m, with 0.8 m of water over the road along the bar at its lowest point. Again in 1989 there was above average winter 242 10 10 10 10 50 40 30 20 10 10 10 10 50 40 30 20 10 10 10 10 10 10 50 40 30 20 10 10 30 20 10 20 10 30 20 10 30 20 10 0 • 6. frorr and Journal of the Royal Society of Western Australia, 80(4), December 1997 )6 m3 974 M 21 975 976 JZZL 977 979 980 981 , T77\ . T771 , 1982 52 983 J YZL 23 984 985 986 Y7A ■ V77K . 77A 987 0.5 106 988 ll II 989 1990 16 37 991 28 992 y p V7X M r777 58 993 ■ 11 1 I 103 FMAMJJASOND Month inthly and annual flow (106 m3) in the Pallinup auged catchment 602001 (3655 km2). Data from the vers Commission, Perth. rain, the Inlet water level was about 3.3 m (Fig 4), roads and farmland were flooded, and paperbark trees and coastal moorts along the eastern shore of the Inlet died. In May 1990 a pipe of 900 mm internal diameter was installed in the eastern end of the bar, with an overflow level at 2.2 m (Anon 1993a). Rainfall was well below average in 1991, there was little river flow and Inlet water level continued to fall for the first seven months of 1992 to about 0.4 m. Early rain saturated the catchment soils and river flow from 300 mm of rain in August to October raised Inlet water level to 3.9 m (Fig 5). Water seeped through the bar/dune along its whole length and flowed strongly where the 1990 pipe discharged onto limestone shore rock. By mid March 1993, evaporation and seepage through the bar had lowered water level to 2.2 m, which was still a dangerously high level for the beginning of winter (Fig 5). Above average rainfall in February to early May on a saturated catchment raised the water level to 3.3 m by 9 May. A cut was hastily made through the western dune to lower the water level without breaking the bar. However on 28-29 May, before the planned spillway could be completed, about 50 mm of rain brought river flow that raised the Inlet water level nearly half a metre in 48 hours, to 3.7 m. Water tore through the cut and opened a 70 m wide breach through the dune and beach to about 3 m below sea level, and then scoured a channel against the dune (Fig 8). The water level fell 3 m in 3 days but it was a week before the Inlet was tidal. Within six weeks the huge volume of sand and shells carried out to sea had rebuilt a wide beach to about 2.3 m above sea level and closed the Inlet. Since the break, rainfall has been below average and Inlet water level has only briefly been above sea level. In February 1997 the Inlet was a salt pan, and when seen in April 1997 sea water was flowing into it through the eastern bar and through the beach at the breach in the western bar. The beach, at 2.5 m, was still effectively the bar and outside the line of the dune. If the cut had not been made, flow from the 50 mm of rain of 28-29 May could have raised water level to 4.5 m and broken the eastern low bar (50 mm of rain in October 1992 caused a rise in water level of 1.4 m). The 300 mm of rain in August-October 1992 would also have broken the bar if the water level had not been so low (0.5 m) in July. There have been previous occasions when the bar was threatened by well above average rainfall and river flow without it breaking: e.g. from 1917-1920 (when it was broken artificially), in 1955 and from 1958-1960. But now, as noted above, river flow is probably considerably greater than on those occasions. Management Background to management Following the floods of 1988-89, there was pressure from farmers and the Hopetoun community to prevent flooding and ensure road access from Hopetoun to the Fitzgerald River National Park. A proposal to cut the bar was deferred, and in 1990 the pipe was installed to lower the water level and reduce the risk of flooding. This did not have the capacity to cope with major floods and several proposals for the controlled release of flood water 243 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 7. Culham Inlet bar/dune area, January 1981. Scale 1:15 000. Photograph from the Department of Land Admin¬ istration, Perth; Copy Licence 519/97. at the eastern low bar were examined. The high Inlet water level (3.9 m) in October 1992 again caused extensive flooding and threatened to break the bar. Continued high water levels early in 1993 made it urgent to implement the planned controlled release of Inlet water through the western high dune before winter, but this was overtaken by river flow from the storm of May 28-29 and the water broke through. After the break, several public meetings were held at Hopetoun to discuss the future of Culham Inlet. Everyone wanted continuous road access from Hopetoun to the Park, but inevitably there were conflicts of interest and little agreement as to the level to which Inlet water should be allowed to rise. Some low-lying paddocks are subject to flooding at 2 m and farmers did not want their land flooded, commercial fishers were concerned at the potential loss of the profitable black bream fishery (Anderson & Cribb 1994) and wanted to maintain the highest possible level, as did local residents concerned to preserve the ecosystem. The compromise reached at a public meeting in July 1993 was that the Inlet should hold as much water as possible consistent with minimal flooding of paddocks and access roads, and that the road from Hopetoun to the Park should not be closed for more than about five days once in five years (Anon 1993a). A critical water level for management is 3.5 m, above which there will be unacceptable flooding. Management options Five engineering measures to restore road access and manage the bar were proposed and assessed. All provided for flood water to be released at the site of the 1993 break in the bar and the roadway to be rebuilt at 4 m AHD. The first four involved rebuilding the road along its former alignment adjacent to the bar. Option 5 diverted the road inland from the bar to protect the roadway from wave action and reduce scour by flood flow (Fig 8). The five management options were; 1. Build a 120 m long floodway section of roadway at 2 m (or 2.5 m) to release flood water (Anon 1993a). This allowed for the road to be closed briefly during floods, with a probability of once in five years. 2. Rebuild the road as an effectively impermeable bar¬ rier (Anon 1993b). 3. Rebuild the road with a 40 m long sacrificial section at 3.5 m to release flood water (Anon 1993b). 4. Build a 60 m long bridge over a sacrificial section of roadway at about 3.5 m to release flood water (Anon 1993b). 5. Rebuild the road in an arc about 500 m inland from the break in the dune at 4 m (Fig 8), with a 100 m long relief floodway at 3.7 m, eleven 1.8 m diameter culverts through the road embankment at an invert level of 1 m to release flood water, and one smaller culvert at -1 m to equalise water levels between the Inlet and the pool between the roadway and the bar 244 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 8. Top. Culham Inlet bar/dune area with the new road, January 1997. Sea water is seeping into the almost dry Inlet through the beach and through the scour channel and low level pipe. Photo from the Department of Land Administration; Copy Licence 530/98. Bottom. Explanatory diagram: A, culverts (11 at 1 m above AHD); B, culvert (1 at 1 m below AHD; C, the 1993 scour channel; D, site of the 1990 pipe. (Anon 1995). If, as anticipated, the bar rebuilt to 3.5 m it would have to be broken when flood water reached that level. Option 1 was not pursued when Option 4 was proposed. Options 2 and 3 were rejected because of the probability that floods would breach the road, perhaps once in 20 years, and access to the Park interrupted for long periods while it was being restored. Further study of Option 4 was not pursued because of the risk of "significant and uncontrolled failure [of the roadway] at less than acceptable frequencies" (Anon 1995). Option 5 has now been implemented and the road was opened officially on 15 April 1997. Environmental considerations The prime environmental objective for management should be to make Culham Inlet as healthy and productive a water body as possible and restore the Inlet to as near to its pre-1993 condition as is now practicable. compatible with there being minimal flooding of paddocks and access roads. To achieve this the Inlet must hold as much water as possible, for as long as possible, and with a salinity less than about 70 ppt. There is probably little river flow to the Inlet in three out of four years, mean annual evaporation is about 1500 mm, the Inlet is only 1 m deep below AHD, the water becomes >70 ppt during long periods of low river flow, and all water can be lost by evaporation. The capacity of the Inlet to accept a river flow event depends on the area of the Inlet (11 km2), the volume of flow («* 11 x 106 m3 for every 1 m depth), the initial water level, and the height of the retaining barrier. The volume and timing of river flow depend on rainfall and the moisture content and compaction of the catchment soils. Rainfall and soil moisture are unpredictable and so too must be the volume and frequency of river flow, while the depth and salinity of water retained after flood flow will depend on the height of the barrier. On past experience, the 245 Journal of the Royal Society of Western Australia, 80(4), December 1997 combination of events which produced the high water levels of 1989, 1992 and 1993 might be expected to recur once in about fifty years. However, river flow may have at least doubled as the result of catchment clearing during the last 30 years, and the previous 50-year probability of a 3.5 m high bar breaking may now be a 20 to 30-year probability, and a 2.5 m bar breaking every five to seven years. The potential for management to achieve the above environmental objective for Culham Inlet (i.e. with as much water as possible for as long as possible and the salinity less than about 70 ppt) will depend on the height of the retaining barrier and how often it breaks. Option 1 would retain flood water to the level of the floodway (2 m or 2.5 m) with a maximum depth of 3 to 3.5 m of water. The Inlet would again be a closed salt lake, but the water would become hypersaline and all water dry up more often than in the past. Options 2, 3 and 4 have the potential to retain water to 3.5 m (Option 2 to 4 m) with the capacity to return the Inlet close to its pre-1993 condition until the water level reaches 3.5 m and the road/bar is breached. All Inlet water would then be lost down to sea level, the Inlet would be briefly an estuary as in May 1993, and would only revert to its previous condition after the barrier was rebuilt. The Option 5 road and pipes are now in place. However four years after the bar was breached, the beach was still effectively the bar; it was at 2.5 m and still seaward of the dune line. In this situation it can be breached (either from the Inlet or by wave action) so frequently that a stable bar is unlikely to rebuild higher on the dune line naturally. When the bar breaks all water will be lost down to 1 m and leave only 2 m depth of water, water that can become hypersaline and dry up. The Option 5 proposal envisaged the bar rebuilding to 3.5 m on the dune line (Anon 1995) and at that height it would retain water to a depth of 4.5 m and Culham Inlet would have the potential to be a healthy and productive a waterbody again — while the bar held. Discussion Culham Inlet was an estuary until about 3500 years ago, but probably for several hundred years it has been a coastal salt lake with a high sea bar that last broke naturally in 1845. The water level varied from -1 m to 4 m AHD (0 m to 5 m deep) and salinity from 10 ppt to brine. In 1989 and 1992 river flow raised water levels in the Inlet so high that the stability of the bar was threatened, and the road along it was impassable. In May 1993 the water level was already so high that the Inlet no longer had the capacity to accept that month's river flow, and the bar broke. Following the break, the priority for management was to ensure road access to the National Park and limit paddock flooding. The road has been rebuilt; but what is the future of the Culham Inlet ecosystem? Can it be a healthy and productive environment again? The health and productivity of Culham Inlet depend mainly on the depth and salinity of the water. Depth and salinity are dictated by rainfall and river flow, and the height of any retaining barrier and the frequency with which it breaks. It was anticipated that the bar would rebuild to 3.5 m at which height a stable bar on the dune line could be expected break with a frequency of only once in 20 to 30 years. But in May 1997 the beach was still the bar, at 2.5 m and seaward of the dune line, and it can be expected to break relatively frequently, perhaps once every five to seven years, and with such frequent breaks it is unlikely to rebuild higher naturally. When the bar breaks, the water will be lost down to 1 m, only 2 m deep, and will become hypersaline and dry up more often than in the past — considerably more frequently with a 2.5 m beach/bar than with a 3.5 m stable bar that can retain water to a depth of 4.5 m. The bar could be rebuilt to 3.5 m from the excess of sand now on the wide beach; there would be short-term flooding of farm land before the bar broke, but this would seem to be a small price to pay for Culham Inlet to have the potential to be as productive an ecosystem as is now possible. The combination of well above average rainfall in 1992, especially in the cleared upper catchment, the saturated catchment, the high Inlet water level and the high rainfall of May 1993 may have been unusual for the time of year, but the 50 mm of rain of 28-29 May could be expected to recur once in 2.5 years. It was the large volume of river flow from that rain that broke the bar; it was beyond the available capacity of the 11 km2 Culham Inlet to accept at that time. The nearby Jerdacuttup Lakes (Fig 1) has three times the area to absorb flow and would not have been threatened. The last time the Culham bar is known to have been threatened was in 1955 when only about 10% of the catchment was cleared, but now with about 40% cleared the river flows of 1992-93 were probably much greater. Coastal lakes and estuaries in the south coast low rainfall area are the receiving waters for river flow from catchments that have been extensively cleared since the 1960s. The bars of two 'normally closed' estuaries, Beaufort Inlet and Stokes Inlet, have broken more frequently following clearing in the catchments; and flood flow has increased sediment transport to their shallow basins (Hodgkin & Clark 1988, 1989). Landcare groups are now implementing management measures in the catchments to reverse the rising water tables and prevent salinisation, and it is to be hoped that these measures will in time reduce runoff and river flow, and soil erosion and sediment transport to the coastal lakes and estuaries. Acknowledgments: Assistance from staff of the Department of Environ¬ mental Protection and the Marine and Harbours Branch of the Depart¬ ment of Transport and access to their files is acknowledged. I am grateful to R Cooper for permission to use his data on water levels in Culham Inlet, and to M Cliff for information on fish catches. W S Andrew, R Clark, and R C J Lenanton have kindly read and discussed drafts of the paper. The support of members of the Hopetoun community is greatly appreciated, especially that of H and R Taylor. References Anderson C &: Cribb A 1994 Fish, floods and tourism at Culham. Western Fisheries, Autumn: 26-27. Anon 1993a Culham Inlet outlet management works design. Unpublished Report. Kinhill Engineers, Victoria Park, West¬ ern Australia Anon 1993b Proposed structure at Culham Inlet. Main Roads. Perth, Western Australia. Anon 1995 Proposed structure at Culham Inlet. Main Roads. Perth, Western Australia. 246 Journal of the Royal Society of Western Australia, 80(4), December 1997 Beard J S 1972 The vegetation of the Newdegate and Bremer Bay areas. Western Australia. Vegmap Publications, Sydney. Beard J S 1973 The vegetation of the Ravensthorpe area, West¬ ern Australia. Vegmap Publications, Sydney. Bennett K & George K 1994 Biological study of Culham Inlet. Report to the Minister for the Environment. State Govern¬ ment of Western Australia, Perth.. Black R E & Rosher J E 1980 The Peel Inlet and Harvey Estuary system hydrology and meteorology. Western Australian De¬ partment of Conservation and Environment, Perth. Bulletin 89. Gregory A C 1849 Report of an examination about Doubtful Island Bay for coal, and a place to embark it, by Commander of the Colonial Schooner "Champion". Mss. Battye Library, Perth. Hodgkin E P & Clark R 1988 Beaufort Inlet and Gordon Inlet, Estuaries of the Shire of Jerramungup. Western Australian Environmental Protection Authority, Perth. Estuarine Stud¬ ies Series Number 4. Hodgkin E P & Clark R 1989 Stokes Inlet and other Estuaries of the Shire of Esperance. Western Australian Environmental Protection Authority, Perth. Estuarine Studies Series Number 5. Hodgkin E P & Clark R 1990 Estuaries of the Shire of Ravensthorpe. Western Australian Environmental Protection Authority, Perth. Estuarine Studies Series Number 7. Hodgkin E P & Lenanton R C 1981. Estuaries and Coastal Lagoons in Western Australia. In: Estuaries and Nutrients (eds B J Neilson and B J Cronin). Humana Press, Clifton N J, 307-321. Hoy R D & Stephens S K 1979 Field study of lake evaporation — analysis of data from phase 2 storages and summary of phase 1 and 2. Australian Water Resources Council Technical Paper Number 41. Australian Government Publishing Service, Canberra. Lenanton R C J 1974 Fish and Crustacea of the Western Aus¬ tralian South Coast Rivers and estuaries. Fisheries Research Bulletin 13. Department of Fisheries & Fauna, Perth. Thom R & Chin R J 1984 Bremer Bay, Western Australia. 1:250 000 Geological Series — Explanatory Notes. Geological Survey of Western Australia, Perth. Thom R, Lipple S L & Sanders C C 1977 Ravensthorpe, Western Australia. 1:250 000 Geological Series — Explanatory Notes. Geological Survey of Western Australia, Perth. 247 Journal of the Royal Society of Western Australia, 80:249-254, 1997 Pre-contact human skeletal remains from Useless Loop, Western Australia N G Jablonski1 & S Bowdler2 1 Department of Anthropology, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118-4599 USA email: njabonski@calacademij.org 2Centre for Archaeology, Department of Anthropology, The University of Western Australia, Nedlands, WA 6907 email: sboiudler@cyllene.uiva.edu.au Manuscript received November 1996 ; accepted July 1997 Abstract A human skull found in 1992 near Useless Loop, Western Australia is described here as being that of a pre-European contact Aboriginal Australian. The skull was that of a gracile male of approximately 50 years of age at the time of death. Teeth recovered with the skull showed heavy wear, and lesions in the alveolar bone of the jaws suggested that the individual possibly suffered from periodontal disease and, probably, at least one painful abscess at the time of death. The morphology of the individual was similar to that of other, contemporary populations of Aboriginal people from the central region of Western Australia. Determination of the absolute age of a sample of cranial bone by accelerator mass spectrometry yielded a probable age of 2730 400 yr bp. This find therefore represents one of a very few sets of pre-European contact human remains in Western Australia to have been recovered from a known location. Introduction Archaeological research in the region of Shark Bay, Western Australia, over the last ten years has revealed a thirty thousand year history of occupation (Bowdler 1990a, b,c, 1995). The evidence has consisted primarily of stone artefacts and food refuse, collected or excavated from open midden and rock shelter sites. Few pre-Euro¬ pean human remains were known from the region, and none have ever been excavated from a primary context, or securely dated. The find reported here, while not com¬ pletely satisfactory in terms of depositional context, at least represents pre-European human skeletal remains from a known location. In October 1992, a human cranium was found on the track north of the township of Useless Loop which had been graded four months previously (Bowdler, unpub¬ lished observations). It was presumed that the cranium had been disturbed from its original resting place by the grader, and could have been moved by up to 400 m. The spot of the find was marked by a wooden stake. The skull was found half-buried in a grader windrow on the side of the track. The spot is about 10-15 m above sea level, on a fairly narrow neck of the Prong, and no more than 100 m from the coast in a straight line. It is, in fact, the sort of place where it would not be unexpected to find a prehistoric site, an open midden scatter of the kind common in this region (Bowder 1990a, b). On careful ex¬ amination near the find, a human incisor was found which, from its state of wear, appeared to be that of a prehistoric Australian Aboriginal person. Fourteen meters to the south, also on the side of the road, were found a chalcedony flake and a piece of baler shell. On © Royal Society of Western Australia 1997 the track and its sides, about 50-100 m to the south, was evidence of a scattered midden site, consisting of Terebralia, baler. Turbo, and other marine shells. The ac¬ cumulation was not very dense, but was reasonably evi¬ dent. A few calcrete flakes were noted also. The soil was a light-coloured dune sand; no darker horizon was evi¬ dent anywhere. Further cranial remains, including the mandible, were recovered approximately six months later within 100 m of the original find. Because the sec¬ ond set of remains matched exactly the expected descrip¬ tion of the missing portions of the cranium and man¬ dible, the two sets were assumed to have been originally associated. We describe here these newly discovered cranial re¬ mains and compare them with those of others known from the area, and more widely in Australia. The age is estimated and the significance of the find is discussed. In accord with the wishes of the Aboriginal people in Denham, the skull was re-interred near the spot of its discovery after it had been studied. Materials and Methods Morphometry The cranium (hereafter designated the ULHK cra¬ nium) was examined and measured at Useless Loop and again at the Department of Anthropology, University of Western Australia. A standard battery of anthropometric measurements were made on the ULHK cranium for comparison with measurements for other modern human populations reported by Howells (1973, 1989). The cra¬ nium was measured using digital calipers, standard spreading calipers, and coordinate calipers. The average value of three measurement trials, rounded to the nearest 249 Journal of the Royal Society of Western Australia, 80(4), December 1997 millimeter, is presented in this report. Complete descrip¬ tions of the measurements can be found in Howells (1973). An estimate of the cranial capacity of the cranium was made by measuring the volume of the sand that filled the endocranial cavity, using a beaker (to ± 50 cm3). The computer software program CRAN1D (Wright, 1989/1992) for the automated comparison of cranial mea¬ surements was used to assess the affinities of the ULHK cranium. This program permits comparison of the di¬ mensions of any unknown human cranium with the di¬ mensions of 2524 female and male crania from around the world based on measurements reported by Howells (1973, 1989). The IDCRAN2 routine of CRANID was used to analyse the average values of the measurements. The first analysis in this routine provides a catalogue of the 50 nearest neighbors of the cranium in question. The sec¬ ond analysis in this routine (the K-means cluster analy¬ sis) indicates that the composition of the group that is closest in shape to the cranium in question. Dating Two dating strategies were employed. First, a radio¬ carbon date was obtained for marine shell remains (‘ Terebralia sp) gathered from the nearby midden site. It seemed a reasonable assumption that the skull could have been associated with the prehistoric site. Second, following consultation with Aboriginal people in Denham, a piece of bone from the nasal area of the skull itself was submitted for radiocarbon dating to the Qua¬ ternary Dating Research Centre at the Australian Na¬ tional University. Accelerator mass spectrometry radio¬ carbon dating with the MUD accelerator of the Research School of Physical Sciences and Engineering was con¬ ducted by J Head. Two fractions, bone apatite and col¬ lagen, were dated separately. Results Description of the remains The human skeletal remains consisted of a single hu¬ man cranium and mandible designated ULHK for pur¬ poses of this report. Selected standard views of the ULHK remains are provided in Figures 1-4. Nearly all elements of the cranium were recovered during the course of the investigation. Because the man¬ dible and fragments of the upper facial skeleton were recovered several months after the cranium had been re¬ turned to Useless Loop, no attempt was made to recon¬ struct the complete skull. The elements of the upper fa¬ cial skeleton that were separated from the main mass were as follows; the maxilla (except orbital process), the pterygoid plates of the sphenoid, the zygomatic bone and the zygomatic process of the temporal bone. The damage to the cranium that separated the bones of the upper right facial skeleton occurred post mortem, and was al¬ most certainly caused by the road grader which disin¬ terred the cranium. The supraciliary ridge on the right side has been punctured post mortem , revealing a large frontal sinus (Fig 1). The puncture hole was almost cer¬ tainly produced by the action of the road grader. No fractures were evident. A comparison of the measurements of ULHK cranium with those of other samples from Western Australia Figure 1. ULHK cranium, norma frontalis. Note damage to upper facial skeleton and supraciliary ridge on right side. Approxi¬ mately one-half actual size. Figure 2. ULHK cranium, norma basalis. Note the small hole in the alveolar bone of the maxilla in the molar region possibly resulting from a dental infection. Approximately one-half actual size. 250 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 3. ULHK cranium, norma lateralis (left). Note bleaching of bone on top of calvaria and in the occipital region. Approximately one-half actual size. (Table 1) indicates that the Useless Loop specimen was very similar in size to other males from the state, and especially to those from the central region of Western Australia (defined by Margetts & Freedman (1977) as around Port Hedland, North West Cape, Shark Bay and Cue). Several teeth were associated with the skull, but only three were measurable (Table 2). In the upper dentition, only the heavily worn stump of the right upper third molar was found in its alveolus. The heavily worn left upper central incisor and left canine were recovered at the site by sieving. The alveoli of the upper left central incisor, lateral incisor, canine, and first premolar were well preserved and clearly held teeth at the time of death. The alveoli of the upper left first molar and third molar were shallow and may have held teeth at death, but the alveolus of the second molar appears to have been re¬ sorbed long before death. A small natural opening in the alveolar bone in the second to third molar region (Fig 2) communicated with the antrum of the maxillary sinus; this suggests that one of those teeth may have been ab¬ scessed at the time of death or that an oro-antral fistula had occurred following the loss of an infected molar. In one of the fragments of the right facial skeleton that was recovered, a similar lesion was found in the floor of the maxillary sinus above the upper third molar. The mandible was complete and did not show signs of dental abscesses or excessive periodontal disease. Pre¬ served on the right side of the mandible were the alveoli of the lateral incisor, the canine, third premolar, the fourth premolar, and the first molar. Four heavily worn teeth were recovered on the left side of the mandible; the canine, third premolar, fourth premolar (stump only, worn to the cervix) and third molar. The alveoli of the first and second molars were also preserved. Resorbed alveoli indicated that several teeth had been lost ante mortem on the left side, the central and lateral incisors; on the right side, the central incisor, the first or second mo¬ lar, and the third molar had been lost. Judging from the advanced state of wear of the teeth and the absence and/or diseased state of the molar teeth, the individual's ability to chew was no doubt compro¬ mised. The cranium was uniformly medium brown in colour except for two areas, on the top of the calvaria and on the left parietal and occipital bones (Figs 1, 3, 4), that appear to have been bleached by recent exposure to the Figure 4. ULHK cranium, posterior view. Approximately one- half actual size. 251 Journal of the Royal Society of Western Australia, 80(4), December 1997 Table 1 Measurements (in mm) of the ULHK cranium compared to those of the following other modem human cranial samples; 1) the mean and range of Western Australian Aboriginal male and female crania [WA- M and WA-F] reported by Margetts & Freedman (1977); 2) the mean for central Western Australian male crania [CWA-MJ reported by Margetts & Freedman (1977); 3) and the centroid of the distribution of populations of living Homo sapiens measured by Howells (1973, 1989). The standard abbreviations for the measurements are as follows; glabello-occipital length (GOL); nasio-occipital length (MOL); basion- nasion length (BNL); basion-bregma height (13BH); maximum cranial breadth (XCB); maximum frontal breadth (XFB); biauricular breadth (AUB); biasterionic breadth (ASB); basion-prosthion breadth (BPL); nasion-prosthion height (NPH); nasal height (NLH); orbital height [left] (OBH); orbital breadth [left] (OBB); bijugal breadth (JUB); nasal breadth (NLB); palate breadth (MAB); bimaxillary breadth (ZMB); zygomaxillary subtense (SSS); bifrontal breadth (FMB); nasio-frontal subtense (NAS); biorbital breadth (EKB); interorbital breadth (DKB); cheek height (WMH); nasion-bregma chord (FRC); nasion-bregma subtense (FRS); bregma-lambda chord (PAC); bregma-lambda subtense (PAS); lambda-opisthion chord (OCC); and lambda-opisthion subtense (OCS). The values from Margetts &: Freedman (1977) have been rounded to the nearest mm. ULHK WA-M mean (range) WA-F mean (range) CWA-M mean Homo sapiens centroid GOL 192 187 (169-202) 177 (163-191) 188 179 NOL 187 - - - 177 BNL 100 101 (91-116) 97 (76-129) 101 99 BBH 133 131 (121-144) 128 (120-143) 130 132 XCB 139 131 (117-143) 130 (120-143) 131 137 XFB 110 - - - 113 AUB 122 118 (106-131) 113 (104-127) - 121 ASB 110 107 (97-117) 104 (94-122) - 107 BPL 98 103 (95-119) 97 (62-107) 104 98 NPH 73 69 (60-87) 65 (53-75) 70 66 NLH 52 49 (42-77) 46 (40-59) 51 50 OBH 33 33 (28-44) 32 (26-38) 34 34 OBB 40 40 (30-44) 32 (26-38) 40 39 JUB 114* 118 (105-134) 109 (98-125) - 115 NLB 29* 28 (23-38) 26 (21-30) 28 26 MAB 29* - - - 63 ZMB 96* 92 (80-107) 88 (80-103) 93 95 SSS 20 - - - 23 FMB 103 101 (86-110) 97 (91-111) - 97 NAS 20 - - - 16 EKB 100 100 (93-110) 95 (89-101) - 97 DKB 24 22 (18-28) 20 (17-23) 22 21 WMH 24 - - - 23 FRC 108 112 (99-124) 106 (95-116) - 109 FRS 23 25 (20-31) 26 (21-34) - 25 PAC 112 116 (104-128) 109 (97-126) - 111 PAS 23 - - - 24 OCC 101 94 (82-105) 95 (85-108) - 96 OCS 33 - - - 28 ‘denotes a measurement made on the assumption of bilateral symmetry Table 2 Measurements (mm) of the three measurable teeth from the ULHK remains, compared with mean and range (in brackets) of Western Australian Aboriginal males and females reported by Freedman & Lofgren (1981). Mesiodistal (MD) and buccolingual (BL) dimensions are given. Tooth ULHK ULHK WA Male W A Female MD BL MD BL MD BL Upper left lateral incisor 6.52 7.45 7.24 (5.8-8.6) 7.03 (5. 8-7.9) 7.10 (5.2-8.3) 6.67 (5.6-7.8) Upper left canine 7.19 9.67 8.01 (6.8-9.5) 9.18 (7.6-10.1) 7.81 (7.4-8.4) 8.44 (7.6-9.1) Lower left canine 7.50 11.10 7.28 (6.0-8.4) 8.67 (7.5-9.8) 6.91 (6.2-7.6) 7.79 (7.2-8.7) 252 Journal of the Royal Society of Western Australia, 80(4), December 1997 sun. The cranial cavity was filled with dry, compact sand, and the rootlets of some plants were found adhering to the endocranial surface, to the orbits and to the outside of the base of the skull. No soft tissues or hair were found. The bone of the cranium and mandible was well preserved and quite sturdy, but there was no indication that fossilization had begun. Relatively little mineral con¬ tent had been lost during its period of interment, a con¬ dition that would have been consistent with burial in a well-drained location. The length of time that the cra¬ nium had been in the ground could not be accurately estimated by simple visual examination. The volume of the endocranial cavity was approximately 1450 ml ± 50 ml, which is within the range for modern human males. Estimation of age at death Sutural fusion and obliteration on the ectocranial sur¬ face of the cranium was advanced and the same appears to have been true on the endocranial surface, as far as could be seen. The central parts of the coronal suture (at bregma), the rostral half of the sagittal suture and the central part of the lambdoidal suture were nearly com¬ pletely obliterated. Estimation of age at death from the skeletal remains of adults is fraught with uncertainties and is particularly difficult in cases such as this where no postcranial remains (including the potentially useful pubic symphysis) have been recovered. On the basis of the status of the cranial sutures and dentition, the minimum age of this individual is judged to have been 50 years (estimated range 45-55 yr). The sturdiness of the cranial bones and the absence of marked thinning of these bones would further suggest that the individual was not of an advanced age. Sex of the individual At the time that the cranium was first examined, the first author judged it to be female because the cranium bore relatively gracile muscular markings despite being sturdily built. Further detailed study, however, sug¬ gested that the skull was almost certainly that of a male, after examination of other cranial remains of Aboriginal individuals at the Western Australian Museum (Perth) and the Natural History Museum (London) and after fol¬ lowing the methods of sex determination described by Larnach & Freedman (1964). Unfortunately, because no further skeletal remains were found with the skull, con¬ firmation of sex using elements of the postcranium was not possible. Racial affinity The presence of a suite of several morphological fea¬ tures of the cranium clearly indicates that the individual was an Aboriginal Australian. These features include the marked development of the supraorbital region, a low forehead rising gently to bregma, a broad interorbital region, rectangular orbits, short nasal bones, a large nasal opening, a calvaria showing a "gabled" or "barn-roof" outline (Fig 4), and the presence of an occipital "bun" (Fig 3). These results of visual inspection of the cranium were reinforced by the quantitative analysis. Unusual features In the lambdoidal suture were found six sutural (wormian) bones or Ossa Incae, two of which can be clearly seen in Fig 3. Such bones represent accessory cen¬ tres of ossification and are present as common variations in human crania world-wide. The cranial base was slightly asymmetrical. The right occipital condyle bore a posterolateral extension not seen in its antimer (Fig 2) and the space occupied by the jugu¬ lar bulb was about 50% larger on the right than on the left. Such asymmetries are common to all human groups and probably had no effects on the individual during life. Quantitative morphological analysis The measurements of the ULHK cranium were com¬ pared to those of other Aboriginal crania from Western Australia as compiled by Margetts & Freedman (1977) and to the centroid of the distribution of Howells' (1973, 1989) entire sample of crania from modern human groups. In most cases, the raw measurements of the cra¬ nium ULHK fit comfortably within the range for male crania from Western Australian Aborigines. The first analysis of the IDCRAN2 procedure indi¬ cated 50 crania from Howells' sample that were most similar to the ULHK cranium (note that the comparative sample of Howells' on which CRANID is based includes measurements from southern and eastern, but not west¬ ern Australian Aboriginal skulls). This list revealed that the ULHK cranium was similar to that of many diverse human groups; the single nearest "morphological neigh¬ bor" of the ULHK cranium was a female cranium from the Tolai population in New Britain but this should not be construed to indicate that the ULHK cranium was of an individual from New Britain. This similarity is not surprising in light of the recognized morphological simi¬ larities of Melanesian and Aboriginal Australian popula¬ tions (Howells 1973, 1989). The K-means cluster analysis indicated clearly that the groups that best matched the ULHK cranium were South Australian and Tasmanian Aboriginals, although the cranium also showed strong affinities with groups from New Britain and East Africa. The results of the quantitive morphological analysis support the visual assessment presented in the previous section that the ULHK skull was that of a male Australian Aboriginal. Absolute age of the ULHK remains Dating of the marine shell from the midden site gave an age of 7400 ± 70 bp, which if corrected for the oceanic reservoir effect produces a date of 6950 ± 70 bp. This date is entirely consistent with many other midden sites in the Shark Bay region with Terebralia shells. We cannot, however, assume that the skull was in fact associated with this site. With respect to bone from the skull itself, the apatite fraction gave an age of 2730 ± 400 yr bp, and the collagen fraction produced an age of 750 ± 250 yr bp. The older of the two bone dates would be generally considered the better date because most contamination comes from younger sources of carbon (J Head, Australian National University, personal communication). The shell-derived date is, technically speaking, a more reliable date for the event which it dates, which is the gathering of these particular molluscs. It also fits well 253 journal of the Royal Society of Western Australia, 80(4), December 1997 with other dates from these shell species collected and excavated from shell midden sites in the Shark Bay re¬ gion. The bone date must be preferred, despite its techni¬ cal imprecision, as a date for the death of the individual from which the bone was derived. Discussion The results of visual analysis of the ULHK cranium and analysis of the measurements of the cranium using CRANID indicate that the ULHK skull was that of a male Aboriginal Australian, who was probably about 50 years old at the time of death. The state of preservation of the bone of the cranium was generally good, and all the damage sustained to the cranium on the right side of the facial skeleton and right supraciliary ridge was consis¬ tent with damage caused by the road grader which un¬ earthed the cranium. The ULHK cranium represented an individual who suffered from extreme dental wear and reasonably serious dental disease near the time of death. Measurements of the ULHK cranium were found to be very similar to those of other samples of male crania from Western Australia. In their comparative study of a large series of modern Aboriginal crania from Western Australia and coastal New South Wales, Margetts & Freedman (1977) noted progressive morphometric sepa¬ ration between equivalent parts of the two states. That is, greater similarities between the northern populations, and progressively fewer between the central and south¬ ern populations, were found. This result, they indicated, was suggestive of a north to south migration down the east and west coasts. An alternative explanation can also be offered. The greater similarities between the northern populations of eastern and western Australia may have been due to greater genetic interchange between these populations because of fewer major ecogeographical bar¬ riers between them. The increasingly greater differences between central and southern populations may have been due to a greatly reduced chance of genetic inter¬ change between the populations resulting from greater distances and more severe ecological barriers separating them. The ULHK skull appears to represent an individual who was buried at the top of a hill in such a way that the skull came to protrude from the side of the track cut through the crest of the dune. It was then scooped up by the grader and redeposited at the spot of the find in the track. It is likely that the postcranial remains remain in situ in the side of the track. Recommendations made in an original (Bowdler, 1992, unpublished) report to the former Aboriginal Sites Department (Western Australian Government, Perth) included the future possibility of find¬ ing further skeletal materials at the same site. The ULHK skull was found close to a midden dated to about 7000 yr bp, but the skull itself appears to be around 2700 or possibly even 750 yr old. Whichever date is more accurate, the skull clearly predates the arrival of Europeans in Western Australia. Illustrated descriptions of such finds are rare. The ULHK skull represents an important example of central Western Australian Ab¬ original morphology that contributes to our understand¬ ing of human variability as it existed in Australia during the Holocene, prior to European settlement. Acknowledgements: SB would like to thank the members of the Yagdallah Club in Denham and particularly S Gosper, for their help and for grant¬ ing permission to publish the findings of this study. J Head is thanked for expediting the AMS date, and S McGann for assistance in the field. People at Useless Loop were particularly helpful and friendly, especially G and S Privett, R O'Keefe and H Bielawski. NGJ would like to thank M Lofgren (Western Australian Museum, Perth) and C Stringer (Natural History Mu¬ seum, London) for access to skeletal remains in their care. References Bowdler S 1990a Archaeological research in the Shark Bay region: an introductory account. In: Research in Shark Bay: Report of the France-Australia Bicentenary Expedition Committee (eds P F Berry, S D Bradshaw &BR Wilson). Western Aus¬ tralian Museum, Perth, 1-12. Bowdler S 1990b Before Dirk Hartog: prehistoric archaeological research in Shark Bay, Western Australia. Australian Archae¬ ology 30:46-57. Bowdler S 1990c The Silver Dollar site. Shark Bay: an interim report. Australian Aboriginal Studies 1990, 2:60-63. Bowdler S 1995 The excavation of two small rockshelters at Monkey Mia, Shark Bay. Australian Archaeology 40:1-13. Freedman L & Lofgren M 1981 Odontometrics of Western Aus¬ tralian aborigines. Archaeology in Oceania. 16:87-93. Howells W W 1973 Cranial Variation in Man. A study by multi¬ variate analysis of patterns of difference among recent human populations. Papers of the Peabody Museum of Archaeology and Ethnology, Harvard University. Vol 67. Howells W W 1989 Skull Shapes and the Map. Papers of the Peabody Museum of Archaeology and Ethnology, Harvard University. Vol 79. Lamach S L & Freedman L 1964 Sex determination of aboriginal crania from coastal New South Wales, Australia. Records of the Australian Museum 26:295-308. Margetts B M and Freedman L 1977 Morphometries of Western Australian Aboriginal skulls. Records of the Western Australian Museum 6:63-105. Wright R 1989/1992 Identifying the origin of a human cranium: Computerized assistance by CRANID. Computer program and manual published and distributed by author (c/0 Depart¬ ment of Anthropology, University of Sydney, Sydney NSW 2006). 254 Journal of the Royal Society of Western Australia, 80:255-262, 1997 Status of a shallow seagrass system, Geographe Bay, south-western Australia K McMahon1, E Young1'2, S Montgomery1, J Cosgrove1, J Wilshaw1 & D I Walker13 department of Botany, The University of Western Australia , Nedlands, WA 6907 ^Current address: DEEB, Monash University, Clayton VIC 3168 fl email: diwalker@cyllene.moa.edu.au Manuscript received December 1996; accepted March 1997 Abstract Geographe Bay, southwestern Australia, is a shallow open embay ment with sandy substrata dominated by the seagrass Posidonia sinuosa, which has greater than 60% cover of the bay. The surrounding agricultural catchments are sandy and low-lying, with extensive drains contributing anthropogenic nutrients (255 tonnes of N and 34 tonnes of P seasonally) to Geographe Bay. Potential for eutrophication of this system and the subsequent effects on the biota, particularly the health of seagrass communities, has been of concern by local residents. This study documented physical parameters in the water column and the distribution and status of the benthic biota from 1993 to 1995. There were seasonal changes in water temperature (ranging between 14.8 °C in winter to 21.6 °C in summer), and irradiance (ranging between 0 in winter to 850 pmol m'2 s1 in summer). These changes followed expected seasonal patterns in a mediterranean climate. Above¬ ground biomass of the dominant seagrass Posidonia sinuosa (115 to 470 g nr2) was similar to that reported for unpolluted systems in southern Australia, as was epiphyte load (0.1 to 26 g irr2). Geographe Bay showed few symptoms of eutrophication despite seasonal increases in nutrient loads Introduction Geographe Bay is a 100 km wide north-facing embayment, situated between Cape Bouvard and Cape Naturaliste, on the south western coast of Australia (Fig 1). It is a relatively protected bay with a gentle sloping bathymetry (2 m km'1). A Holocene sediment veneer overlies Pleistocene limestones and clays (Paul & Searle 1978). The shallow subtidal region (2-14 m) is characterised by sandy substrata interspersed with lime¬ stone pavements and reef. There are extensive beds of seagrass Posidonia sinuosa Cambridge & Kuo, with greater than 60% coverage throughout the bay. Amphibolis antarctica (Labillardiere) Sonder el Ascherson ex Ascherson, often occurs on the periphery of P. situtosa meadows and limestone pavements (Walker el al. 1987). Land bordering Geographe Bay has been cleared ex¬ tensively for agriculture over the last 150 years. A net¬ work of drains was developed to facilitate agricultural and urban development (Fig 1). Cattle farming with lesser amounts of sheep farming and potato and lucerne crops, are the predominant agricultural practices. The soils in the district have poor nutrient retention capaci¬ ties and are highly leached with low ambient nutrient concentrations (McComb & Davis 1993). The establish¬ ment of agriculture has relied extensively on the addition of nutrients, especially trace elements, phosphatic and nitrogenous fertilisers (McComb & Davis 1993). Much phosphorus is lost to drainage from leaching sandy soils (Hodgkin & Hamilton 1993). This results in an increase in nutrients draining from the Geographe Bay catchment. These have been calculated as an annual total of 255 © Royal Society of Western Australia 1997 tonnes of nitrogen and 34 tonnes of phosphorus (Holmes 1995). Increases in nutrient concentrations above a particular level in a system often result in eutrophication (Stirn 1994). Detrimental effects on coastal systems over¬ enriched with nutrients have occurred worldwide (Nixon 1993) and can be identified as symptoms of eutrophication. These include biological factors such as; • the presence of phytoplankton amd macroalgal blooms especially an increase in frequency, duration and extent of the blooms (Lukatelich & McComb 1989; Malone 1991); • high algal epiphyte biomass often measured as epiphyte biomass to seagrass leaf biomass; • absence of or low biomass of benthic macrophytes (Cambridge el al. 1986; Silberstein el al. 1986; Neverauskas 1987, 1988); • shifts in species composition (Lukatelich & McComb 1989); • decrease in diversity of organisms (Walker et al. 1991); and • increase in diseases of fish and water fowl (Kemp et al. 1984); and physical factors such as; • high light attenuation coefficient; and • low dissolved oxygen concentrations (Dubravko et al. 1993). With nutrient enrichment, phytoplankton can rapidly increase biomass and bloom (Ryther & Dunstan 1971). Chlorophyll a is a relative estimate of phytoplankton 255 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 1. Map of Geographe Bay showing the location of the study sites; 1, Dunsborough; 2, Toby's drain; 3, Buayanup drain; 4, Vasse-Diversion drain; 5, Vasse-Wonnerup Estuary; 6, Forrest Beach; 7, Capel River; 8, 5 Mile Brook diversion. biomass (Round 1981). Concentrations greater than 1 pg H are higher than background levels in the oligotrophic, temperate waters of south Western Australia, and indicate blooms of phytoplankton (Anon 1993). Cyanobacteria may also bloom under conditions of nutrient enrichment (Hallegraeff 1992). Under nitrogen-limited conditions with an adequate phosphorus supply, cyanobacteria can proliferate due to their ability to fix and supply their own nitrogen requirements (Round 1981). Biomass of opportunistic epiphytes may be boosted under conditions of nutrient enrichment, if other variables are favourable for growth (Pickering et al. 1993). The epiphyte to leaf biomass ratio (E/L) is a indicator of excessive epiphyte loads on seagrasses (Penhale & Smith 1977). Values of E/L greater than 0.3 for P. sinuosa indicate excessive epiphyte biomass with a deleterious shading effect on the seagrass (Neverauskas 1987). The presence of submerged aquatic vegetation can be an indicator of water quality (adequate light penetration) and nutrient status i.e. low nutrient concentrations (Walker & McComb 1992; Dennison et al. 1993). In eutrophic systems where light is often reduced for extended periods, there may be decreases in the biomass of benthic macrophytes (Neverauskas 1988; Sand-Jensen & Borum 1991). The changes in seagrass biomass can be an indicator of the extent of eutrophication (Hillman et al. 1989). For the seagrass Posidonia sinuosa in a monospecific meadow down to 4 m deep, biomass ranges between 250 and 500 g nr2 (Hillman & Morrison 1994). A decrease from this range may indicate that seagrasses growing in the same depth and under similar conditions are under stress (Neverauskas 1987). There have been recent concerns about the health of the Geographe Bay system. Losses of seagrass cover in some nearshore areas were documented in the 1970s (Conacher 1993), but the study showed some indication of subsequent recovery. Anecdotal evidence of macroalgal blooms in summer are of concern as potential symptoms of eutrophication. The aim of this study was to document physical and biological conditions in Geographe Bay. Physical profiles of light, temperature and salinity were recorded. Biological measures of seagrass biomass, algal epiphyte load, chlorophyll a concentration, the presence or absence of cyanobacterial and diatom aggregations and the presence or absence of seagrass wrack were taken. These were interpreted to assess whether there were any symptoms of eutrophication in Geographe Bay. Methods Geographe Bay has a temperate, mediterranean type climate, characterised by warm, dry summers and cool, wet winters (Walter 1979). The annual rainfall is 800 mm 256 Journal of the Royal Society of Western Australia, 80(4), December 1997 a year, with 85% of the rain falling between May and October (Fahrner & Pattiaratchi 1995). In summer the winds are generally easterly in the morning at speeds of 5 m s'1 with a south-southwesterly sea breeze in the afternoon, with speeds up to 15 m s'1. Winter weather patterns are dominated by the passage of westerly cold fronts associated with low pressure systems which cross the coast every 7 to 10 days. These result in strong, sustained west to north-westerlies of up to 15 m s*1. An average of 15 days of winter storms occur every year (Fahrner & Pattiaratchi 1995). Water movement in Geographe Bay is mainly wind driven, as the tidal range is small, generally less than 1 m. Fahrner & Pattiaratchi (1995) predicted a predominantly northerly transport of water along the perimeter of the bay from Cape Naturaliste to Myalup in westerly, south¬ westerly and southerly winds. South-easterly winds result in the offshore transport of water and easterly winds, in the offshore transport of water in a south-westerly direction. North-westerly winds would result in the onshore transport of water with weak and variable currents. The average flushing time off Busselton is predicted as 3 to 5 days for easterly, southerly and south westerly winds. Longer flushing times, up to 14 days, occur when south-easterly and north-westerly winds dominate (Fahrner &c Pattiaratchi 1995). In recent years, a cold current (the 'Capes Current7) has been documented in the south west of Western Australia (A Pearce & C Pattiaratchi, pers comm); it originates on the western side of Cape Naturaliste, impinges on the nearshore of Geographe Bay, and moves northward, along the coast¬ line to Perth. Eight sites were chosen to survey physical and biological conditions in nearshore Geographe Bay (Fig 1). Six were potential impact sites, off-shore from drainage systems (Toby's drain, 33 ° 37.797 'S 115° 10.794 'E; Buayanup drain, 33 ° 38.531 'S 115 ° 14.933 7E; Vasse- Diversion drain, 33 ° 38.339 'S 115 0 19.303 'E; Vasse- Wonnerup Estuary, 33 ° 36.116 'S, 115 ° 25.401 'E; Capel River, 33 ° 30.194 7S 115 ° 31.362 7E; 5 Mile Brook Diversion, 33 0 27.572 'S, 115 ° 33.593 'E) and two were used as reference sites with no direct terrestrial run-off (Dunsborough, 33° 36.425 'S, 115 ° 07.112 'E; Forrest Beach, 33 ° 34.265 7S 115 ° 26.783 7E). Over the summers of 1993/1994 and 1994/1995, surveys were undertaken every fortnight. In April, July and September 1994, four sites were sampled more intensively (Buayanup drain, Vasse Diversion drain, Vasse-Wonnerup Estuary and Dunsborough). In July 1994, only Dunsborough and Vasse-Wonnerup Estuary were sampled due to rough weather conditions. Each site was located in 5 m of water, approximately 500 m from shore, in the middle of a seagrass bed. At each site, light and temperature profiles were recorded, and benthic communities surveyed. In July 1994, benthic communities were not surveyed due to turbid conditions. The information for the eight sites sampled intensively over the summers was used as general habitat descriptions. The data from the four sites studied over the whole year were used to compare physical and biological conditions. Light and temperature Temperature (°C) and light intensity (pmol m'2 s'1) were measured with a Flamon Yeokal temperature /salinity bridge and a Licor light meter with an underwater sensor. A vertical profile was carried out at 1 m intervals from the bottom to 1 m from the surface, then at 0.5 and 0.1 m from the surface. Light measurements were used to cal¬ culate a vertical light attenuation coefficient as the slope of the line between log of light and depth (Kirk 1994). Secchi disk depth (Zsj) readings were made and attenua¬ tion coefficients were then calculated using the formula Kd = 1.44/Zsd where Kdis the attenuation coefficient (Kirk 1994). Benthic communities A 300 m equilateral triangular transect (100 m each side) was swum with SCUBA at each site to document benthic communities and physical features. Observations were made at 10 m intervals along the transect. The type of substratum was recorded as sand, low pavement rock or high rock and the percentage cover of benthic biota was noted. This percentage cover was a visual estimate of the amount of substratum covered by seagrass and/or macroalgae over a 10 m distance and the value was estimated at 1 m distance above the transect, taking into consideration 1 m either side of the transect. Macroalgae were collected for identification. The presence or absence of monospecific diatom and blue-green aggregations and their relative abundances along the transect were also noted. The number of times each group was present along the transect was summed and then divided by the total number (30) of possible observations. The relative presence value ranged between 0 for no occurrence to 1 for occurrence along the whole transect. This information provided a general habitat description of the area. P. sinuosa and epiphyte biomass estimates Seagrass dry weight biomass was estimated at each site by stratified sampling (Walker 1988). Six 20 x 20 cm quadrats were placed haphazardly within a 5 m radius, in the middle of the dominant seagrass bed, at each site. All the above-ground material originating from inside the quadrat was collected and frozen. On return to the laboratory, each sample was scraped with an industrial razor blade to remove the erect algal epiphytes (encrusting epiphytes were not removed) and separated into red (Rhodophyta), green (Chlorophyta) and brown (Phaeophyta) algae. Seagrass and epiphytic material were dried for 24 hours and then weighed to give an estimate of dry weight in grams (Neverauskas 1987). These values were extrapolated to dry weight (dw) biomass of seagrass and epiphytes in grams per unit area (g m'2). The epiphyte to leaf (E/L) ratio was calculated for each site and sampling period. This is the ratio of epiphyte dry weight biomass to seagrass dry weight biomass (Neverauskas 1987). As extensive amounts of seagrass wrack accumulated over winter; in July the cover of seagrass wrack on the beach was quantified along a 1.5 km stretch of shoreline westward from the entrance of the Vasse-Wonnerup Estuary. Large accumulations of seagrass wrack (mainly Posidonia sinuosa) were present on the sandy beach out to 100 m from shore, between Vasse-Wonnerup Estuary and Vasse Diversion drain in July. The cover was estimated at 30 m intervals along a ten metre wide strip from the shore-line up the beach. Along a 350 m stretch, the volume of wrack was determined by measuring the 257 Journal of the Royal Society of Western Australia, 80(4), December 1997 height, width and depth of each wrack 'dump". The area from which this total volume of seagrass may have originated was calculated by estimating dry weight biomass of the wrack from the volume of wrack present (Hansen & Brearley, pers comm) and then relating this to average biomass (g m2) in the area offshore. Results Physical conditions Drain flow and water colour. During summer (Decem¬ ber to February) and in April, the water in Geographe Bay was calm and clear; the bottom was visible at 5 m depth. None of the drains or the estuary were flowing. In July, all drains and the estuary were flowing rapidly, discharging a large volume of brown-coloured fresh water. Vasse Diversion Drain was flowing fastest followed by the Vasse-Wonnerup Estuary and then Buayanup drain. The water offshore was extremely turbid, the bottom was not visible in 1 m of water, and brown discolouration was visible 3 km out to sea. This discolouration extended from Toby's drain to Forrest Beach. There was no discolouration in the water column at Dunsborough, the reference site. In September, the estuary had ceased to flow but the two drains were flowing more slowly, with Buayanup drain flowing faster than Busselton drain. Temperature. Water temperature peaked at the end of summer (February) with a mean of 21.6 ± 0.4 °C. The mean minimum temperature was recorded in July at 14.8 ± 0.4 °C (Table 1). There was a gradient of temperatures along the coast in summer with 1 °C cooler temperatures in the south-west of the bay in comparison to the north¬ easterly sites. Light. The amount of light above the seagrass canopy, at 5 m depth, also peaked in summer (mean of 620 pmol nr2 s4) and was lowest in winter (Table 1). Turbidity increased dramatically in winter at sites associated with drains. The attenuation coefficient was higher in July and September than in January and April (0.22 - 0.96). There was a slight increase in attenuation coefficient (0.13 - 0.35) in winter at the site not associated with a drain (Table 1). Water column phytoplankton concentration. The water column chlorophyll a concentration was generally low over the year (Table 2). The highest chlorophyll a Table 1 Temperature, light and attenuation coefficients in Geographe Bay. Temperature is averaged over four sites. Both temperature and irradiance were measured at the bottom of the water column, above the seagrass canopy. Site 1 is Dunsborough, the site without an associated drain. Sites 3, 4 and 5 are all associated with drains, and are Buayanup drain, Vasse-Diversion drain and Vasse-Wonnerup Estuary respectively. Values are mean ( ± standard deviation, with a sample size of 4, except July with a sample size of 2). Temperature Irradiance (pmol nr2s4) Light attenuation (°C) Site Sites 1 3 4 5 Mean 1 3 4 5 Mean Jan. 1994 20.8 ± 0.4 830 400 565 700 624 ± 106 0.113 0.101 0.081 0.217 0.128 ± 0.04 Apr 1994 21.3 ±0.2 234 269 197 292 248 ± 24 0.134 0.125 0.054 0.117 0.108 ± 0.02 July 1994 14.8 ns ns ns ns 0.350 ns ns 0.960 0.655 Sep 1994 15.5 ± 0.1 50 43 41 4 35 ±12 0.215 0.340 0.28 0.385 0.305 ± 0.04 Jan 1995 21.6 ±0.3 550 570 430 420 493 ± 45 0.062 0.008 0.105 0.167 0.086 ± 0.04 Table 2 Temporal and spatial variation in P. sinuosa biomass, epiphyte biomass and chlorophyll a concentration at Geographe Values are mean (± standard error, with a sample size of 6). Bay in 1994-995. Month P. sinuosa biomass (g m : [) Algal epiphyte biomass (g nr2) Chlorophyll , a (pg l-1) Sites 1 3 4 5 1 3 4 5 1 3 4 5 Jan 1994 234 ± 32 469 ± 40 287 ± 47 234 ± 32 5 ± 2 1 ± 1 8 ± 3 26 ± 9 0.12 0.22 0.22 0.20 April 1994 113 ± 19 282 ±31 180 ± 38 162 ± 27 10 ± 3 12 ± 2 9 ± 2 2 ± 1 0.08 0.14 0.18 0.12 July 1994 152 ±20 ns ns 116 ±15 24 ± 6 ns ns 0±0 ns ns ns 0.63 Sept 1994 154 ± 26 196 ± 17 152 ± 20 119 ±24 17 ± 5 3 ± 1 1 ± 1 0±0 0.12 0.17 0.23 0.49 Jan 1995 159 ± 21 281 ± 28 244 ± 29 271 ±11 2 ± 1 5 ± 2 11 ±2 1 3 ± 1 0.09 0.27 0.13 0.13 258 Journal of the Royal Society of Western Australia, 80(4), December 1997 Species recorded at eight Table 3 study sites in Geographe Bay, 1993 and 1994. Site 1 2 3 4 5 6 7 8 Seagrass Posidonia sinuosa + + + + + + + + Posidonia angustifolia + + + Amphibolis antarctica + + + + + 4- + + Amphibolis griffithii + + + + + Halophila oval is + + + + + Macroalgae Chlorophyta Caulerpa cactoides + Cauierpa sp + + + + Codium sp + Halimeda cuneata + + Rhipiliopsis sp + Caulocystis uvifera + Dictyopteris muelleri + + Ecklonia radiata + + Hydroclathrus clathratus + Lobophora variegata + Padina sp + + + Scaberia aghardii + + + + Sargassum sp + + + + Rhodophyta Asparagopsis amiata + + Dictymenia sonderi + Gelidium asperum + Gelidium sp + Gigartina disticlm + Laurencia clavata + Osmundaria prolifer a + + Phacelocarpus alatus + Epiphytic algae Cladophora montagneana + + + + + + Cladophora dalmatica + + + + + + + Polycerea nigrescens + + + + + Pachydictyon paniculatum + + + + + + + Pachydictyon polycladum + + + + + + + Metagoniolithon chara + + + + + + Lenormandia marginata + + Chondria sp + Microalgae Mastogloia sp + + + + + + Chroococcus sp + + + + + + + + concentration was recorded in winter at Vasse- Wonnerup Estuary (0.63 pg l'1) and the lowest in summer at Dunsborough (0.08 pg l'1). The average concentration over the year at the four sites was 0.22 ± 0.05 pg T1. Benthic communities - general description. Seagrass and algal species found in Geographe Bay are detailed in Table 3. The benthic substrata of Geographe Bay can be divided into three main categories; sandy substrata, a combination of sandy substrata and low relief reef and a combination of low and high relief reef. Sandy substrata were present in the south-western portion of the bay between Dunsborough and Vasse-Wonnerup Estuary. These areas were dominated by monospecific meadows of Posidonia sinuosa, with Amphibolis antarctica and Amphibolis griffithii (J Black) den Hartog often on the periphery of the meadows. Closer to Dunsborough the cover of Amphibolis spp increased and a number of large meadows occurred. The brown macroalga, Scaberia agardhii Greville, was also present in patches. 259 Journal of the Royal Society of Western Australia, 80(4), December 1997 Sites with sandy substratum and low relief reef were found north of Wonnerup, around Forrest Beach, with a combination of seagrasses ( P . sinuosa, A. antarctica, A. griffithii and Halophila ovalis (R Brown) JD Hooker) in small patches. Dominant macroalgae included Osmundaria prolifera Lamouroux, Scaberia aghardii , Sargassum spp, Caulcrpa spp and Padina spp. Plate corals and sponges were also present. At Capel and 5 Mile Brook Diversion, there was low and high relief reef. There were small patches of the seagrasses, A. griffithii and A. antarctica, at Capel. Posidonia angustifolia was found on sandy bottoms sur¬ rounding high relief reef, further north. Most common benthic macroalgae were the kelp, Ecklonia radiata (C Agardh) J Agardh, and Sargassum spp. Many small red turf algae were present along with erect red and green algae such as Caulerpa spp. Three main types of epiphytes were noted in the bay; macroalgal epiphytes (Table 3), diatom aggregations and cyanobacterial aggregations. There was temporal varia¬ tion in these epiphytes, with Cladophora Kuetzing (Chlorophyta), Pachydictyon and Polycerea (Phaeophyta), dominant in summer and red algae such as Metagoniolithon Ducker more common in winter. Amphibolis bore a greater toad of red algal epiphytes than did Posidonia. An extensive bloom of Cladophora occurred in December of 1993 and 1994 at Capel River and Forrest Beach. This disappeared after one month. Mastogloia Hustedt (Bacillariophyta) aggregations were found in the south-western portion of the bay. They increased in cover over summer. After a storm in summer, the majority of the aggregations were gone and remained absent throughout the rest of the year. Chroococcus Nageli (cyanobacteria) aggregations were present at all sites. They increased in cover over summer, especially at Vasse River Diversion drain and Vasse- Wonnerup Estuary, but were not present during the rest of the year. P. sinuosa biomass. Highest biomass was recorded in summer and lowest in winter (Table 2). The mean maxi¬ mum biomass for P. sinuosa occurred during summer (470 g m'2 at Buayanup drain) and the mean minimum biomass occurred in winter (115 g nr2 at Vasse- Wonnerup Estuary). Buayanup drain had the highest biomass of the four sites examined. Epiphytic algal biomass. There were temporal and spatial variations in the epiphytic communities on the seagrass (Table 2) with two main patterns in algal epiphyte biomass. At the site without a drain, the highest epiphyte biomass was recorded in winter (24 g nr2) and the lowest in summer (2 g nr2). However at sites with drains, the highest biomass in summer (26 g rrr2), and the lowest in winter (0.1 g nr2). The E/L ratio never exceeded 0.2 throughout the year. Seagrass wrack accumulations. Cover of beached seagrass wrack ranged between 0 to 45% of the area surveyed at each ten meter interval. Along a 1.5 km stretch of beach the average cover was 24%. Along a 350 m stretch of beach, the total volume of wrack present was estimated at 81 000 m3. Discussion The marine benthic communities examined at 5 m depth in Geographe Bay were dominated by seagrass, particularly Posidonia, although sites with more reef had more Amphibolis. The seasonal variation in seagrass biomass was relatively consistent and within the range of other studies (Hillman et al 1989). Algal communities were typical of Western Australian coastal reefs and seagrass beds. There were no large accumulations of drift algae, in summer or winter. Algal epiphytes species and the change of species seasonally were typical of seagrass communities along the south-western Australian coast (Hillman & Morrison 1994) Extensive wrack deposits were observed in winter, mainly composed of seagrass as reported elsewhere on the Western Australian coastline (Kirkman & Walker 1989). The volume of wrack accumulated on a 350 m stretch of beach had an estimated dry weight biomass of 2-8.0 106 kg dw. This equates to 4.8 km2 of seagrass at maximum biomass. The area that this is likely to originate from is 15 km2 of seagrass meadow. A number of parameters measured in Geographe Bay varied over the year. The water column temperature de¬ creased from a summer and autumn maximum of 21.6 °C to a winter and spring minimum of 14.8 °C. The 6.8 °C reduction in temperature between summer and winter is expected for this latitudinal position (Dring 1992). The water temperature in Geographe Bay was 1 °C less than that recorded for Perth coastal waters at the same depth (Buckee et al. 1994). This probably reflects the cold Capes Current which moves northward along Geographe Bay and the warmer Leeuwin Current which passes Perth (Cresswell & Golding 1980). Vertical light attenuation coefficient increased in July and September, indicating that light was dissipated more quickly through the water column. Phytoplankton and organic and inorganic particles held within the water column all affect attenuation (Kirk 1994). High turbidity from suspended particles and tannins discharged from the drains increased the attenuation in July and September, at sites with drains. The accumulation and breakdown of seagrass in nearshore waters in July, especially at Vasse-Wonnerup Estuary also contributed to increased light attenuation. Light and temperature often determine the relative growth rates of algae, with lower temperatures and irradiances interacting to reduce relative growth rates (Round 1981). In July and September, at the drain sites, very few epiphytes were found on P. sinuosa. The low irradiance at the seagrass bed may have restricted epiphyte growth. Dunsborough, the site without the drain, had the highest epiphyte load in July and attenuation was not as high as the other drain sites. Algal epiphytes were most likely not limited by light at this site. Dunsborough was the only site conforming to the pattern described in the literature of higher epiphyte loads on P. sinuosa in winter (Hillman & Morrison 1994). Phytoplankton concentrations (estimated from chlorophyll a) were low through out the year at all sites, with no blooms detected. There was a noticeable increase in phytoplankton in the water column at Vasse- Wonnerup Estuary in July and September. This may be a 260 Journal of the Royal Society of Western Australia, 80(4), December 1997 reflection of higher nutrient concentrations in the water column at this time and follows the expected seasonal pattern of spring blooms. The biomass of P. sinuosa seagrass varied seasonally and was similar to values described in the literature (Hillman et al 1989). There was a three fold reduction in seagrass biomass from 350 g m 2 in summer to 120 g m'2 in winter. This seasonal variation is influenced by leaf loss through storm events and shedding of leaves after growth in summer (Patriquin 1975). Studies of net production in the seagrass P. sinuosa under different light and temperature regimes have estimated the minimum temperature and light required for positive net growth (Masini et al 1995). In relation to these studies, net production of seagrass in Geographe Bay at Dunsborough, Buayanup and Vasse-Diversion drain, would be positive from September to April, over spring, summer and autumn, with temperatures greater than 15 °C and irradiance at the top of the seagrass canopy greater than 30 pmol nr2 s'1. At Vasse-Wonnerup Estuary in September, irradiance at the top of the seagrass canopy was 4 pmol nv2 s*1 and in July this would be less because attenuation coefficients were much higher at this time. Therefore at Vasse-Wonnerup Estuary between July and September, net production is likely to be negative. Respiration would be greater than photosynthesis and a net loss of organic carbon could be expected. A mechanism for maintaining growth over winter would be required for seagrasses with net negative productivity. Pirc (1989) found that P. oceanica (L) Delile was able to store carbon and nitrogen in the rhizomes during summer and autumn, which could be mobilised and accessed for leaf growth in winter. A similar mechanism could maintain P. sinuosa in winter when light and temperature limit photosynthesis. The seagrass at Dunsborough, where light did not fall below 50 pmol nr2 s1, could maintain a positive net production throughout the year. A reduction in light reaching a seagrass bed for extended periods, by epiphyte shading or high turbidity in the water column, can cause a decrease in density of seagrass (Neverauskas 1987; Cambridge et al. 1986). Limited light during winter from high turbidity in the water column at Vasse-Wonnerup Estuary and the shading in summer by cyanobacterial aggregations at Vasse Diversion drain and Vasse-Wonnerup Estuary make these sites susceptible to reduction in seagrass density. However to test this hypothesis a longer period of study would be required. Potential symptoms of eutrophication noted in this study were blooms of Cladophora at Capel and Forrest Beach in early summer (December 1993 and 1994). These blooms are expected seasonal phenomena, where elevated nutrient concentrations from winter and warmer conditions promote algal growth. If such a bloom was to continue for extended periods, detrimental effects such as reduced oxygen concentrations and shading effects could occur. There were also aggregations of diatoms and cyanobacteria. Cyanobacteria are often present in eutrophic systems (Hallegraeff 1992). In Geographe Bay, the increase in relative presence of cyanobacteria and diatom aggregations over the summer may be attributed to warmer temperatures and calm conditions. Some cyanobacteria can fix nitrogen and therefore can proliferate in nitrogen-poor conditions if required phosphorus is available (Smith 1984). Cyanobacterial aggregations were dominant around Busselton and diatom aggregations were dominant in the southern end of the study area, near Dunsborough. When strong wrind and wave action were present, the blooms were dispersed. There were no blooms in winter and after summer storms; all diatom aggregations disappeared. Aggregations of diatomaceous mucopolysaccharides ("slime") have also been observed in the mediterranean sea, in summer, under low nutrient concentrations in the water column, less than 1 pg l1 nitrate (Stim 1994; I M Munda, pers. comm.). The nitrate concentration varied from 1 to 3 pg H in the waters of Geographe Bay in summer (unpublished observation). Conclusions This study has identified seasonal variation in physical and biological conditions in the nearshore waters of Geographe Bay. They conform to the expected seasonal patterns in a mediterranean shallow seagrass system. The seagrasses at most sites are in a healthy condition, without exceptional epiphytic loads. Blue-green and diatom aggregations were the only symptom of eutrophication noted in this study and require further study. Acknowledgments: This study was funded by the Water Authority of Western Australia, South-West Division. Additional funding was provided by a UWA Jennifer Arnold Memorial Scholarship to K McMahon. D Lord contributed to this research and provided comments on the manuscript. We are grateful to G Moore, E Cole, C Sim, T Carruthers, J Verduin, A Markey, M Coyne, A Maskew, M Guns tan, M Delval, M Jury and N Marba for valuable assistance with field work. References Anon 1993 Western Australian Water Quality guidelines for fresh and marine waters. Environmental Protection Author¬ ity, Perth. Bulletin 711. Buckee J, Rosich R & van Senden D 1994 Perth Coastal Waters Study: Water Quality Data. Water Authority of Western Aus¬ tralia, Perth. Cambridge M L, Chiffings A W, Brittan C, Moore L & McComb A J 1986 The loss of seagrass in Cockbum Sound, Western Australia. II. Possible causes of seagrass decline. Aquatic Botany 24:269-285. Conacher A 1993 Geographe Bay: time series analysis of changes in seagrass extent over time. Unpublished Report, the Environmental Protection Authority, Perth. Cresswell G & Golding T J 1980 Observations of a south-flowing current in the south-eastern Indian Ocean. Deep Sea Research 27:449-466. Dennison W C, Orth R J, Moore K A, Stevenson J C, Carter V, Dollar S, Bergstrom W & Batiuk R A 1993 Assessing water quality with submersed aquatic vegetation. BioScience 43:86- 94. Dring M J 1992 The Biology of Marine Plants. Cambridge Uni¬ versity Press, Cambridge. Dubravko J, Rabalais N N, Tyrner R E & Wiseman W J 1993 Seasonal coupling between riverbome nutrients, net produc¬ tivity and hypoxia. Marine Pollution Bulletin 26:184-189. Fahmer C K & Pattiaratchi C B 1995 The physical oceanography of Geographe Bay, Western Australia. In: Geographe Bay 261 Journal of the Royal Society of Western Australia, 80(4), December 1997 Summary Report, Wastewater 2040 Strategy for the South- West Region. Water Authority of Western Australia, Perth, 3-12. Hallegraeff G M 1992 Harmful algal blooms in the Australian region. Marine Pollution Bulletin 25:186-190. Hillman K & Morrison P 1994 Determination of time series changes in marine communities. Section 2. In: Perth Coastal Waters Study: Determination of time series changes in marine communities (eds K Hillman, P Morrison, P Jernakoff & J Nielsen). Water Authority of Western Australia, Perth, 1-81. Hillman K, Walker D I, Larkum A W D & McComb A J 1989 Productivity and nutrient limitation. In: Biology of Seagrasses: A Treatise on the Biology of Seagrasses with Special Reference to the Australian Region (eds A W D Larkum, A J McComb & S A Shepherd). Elsevier, Amsterdam, 635-668. Hodgkin E P & Hamilton B H 1993 Fertilisers and eutrophica¬ tion in south-western Australia: setting the scene. Fertiliser Research 36:95-103. Holmes, R M 1995 Contaminants inputs inventory of Geographe Bay. In: Geographe Bay Summary Report, Wastewater 2040 Strategy for the South-West Region. Water Authority of Western Australia, Perth, 13-18. Kemp W M, Boynton W R, Twilley R R, Stevenson J C & Ward L G 1984 Influences of submersed vascular plants in upper Chesapeake Bay. In: The Estuary as a Filter (ed V S Kennedy). Academic Press, New York, 367-394. Kirk J T O 1994 Light and Photosynthesis in Aquatic Ecosys¬ tems. Cambridge University Press, Cambridge. Kirkman H & D I Walker 1989 Western Australian seagrass. In: Biology of Seagrasses: A treatise on the biology of seagrasses with special reference to the Australian region (ed A W D Larkum, A J McComb & S A Shepherd). Elsevier/North Holland, Amsterdam, 157-181. Lukatelich R J & McComb A J 1989 Seasonal changes in macro¬ phyte abundance and composition in a shallow southwest¬ ern Australian estuarine system. Waterways Commission, Perth. Malone T C 1991 River flow, phytoplankton production and oxygen depletion in Chesapeake Bay. Modem and ancient continental shelf anoxia. Geological Society of America Special Publication 83-93, Masini, R J, C J Simpson, J L Cary & A J McComb 1995 Effects of light and temperature on the photosynthesis of temperate meadow-forming seagrasses in Western Australia. Aquatic Botany 49:239-254. McComb A J & Davis J A 1993 Eutrophic waters of southwest¬ ern Australia. Fertiliser Research 36:105-114. Neverauskas V P 1987 Monitoring seagrass beds around a sewage sludge outfall in South Australia. Marine Pollution Bulletin 18:158-164. Neverauskas V P 1988 Response of a Posidonia community to prolonged reduction in light. Aquatic Botany 31:361-366. Nixon S W 1993 Nutrients and coastal waters: too much of a good thing? Oceanus (Summer): 38-47. Pa triquin D G 1975 'Migration' of blowouts in seagrass beds at Barbados and Carriacou, West Indies, and its ecological and geological implications. Aquatic Botany 1:163-189. Paul M J & J D Searle 1978 Shoreline movements Geographe Bay, Western Australia. Proceedings of the Fourth Australian Conference on Coastal & Ocean Engineering Publisher, Adelaide, 207-212. Penhale P A &: W O J Smith 1977 Excretion of dissolved organic carbon by eelgrass (Zostera marina ) and its epiphytes. Lim¬ nology and Oceangraphy 22: 400-407. Pickering T D, Gordon M E & Tong L J 1993 Effect of nutrient pulse concentration and frequency on growth of Gracilaria chilensts plants and levels of epiphytic algae. Journal of Applied Phycology 5:525-533. Pirc H 1989 Seasonal changes in soluble carbohydrate, starch, and energy content in mediterranean seagrasses. Marine Biol¬ ogy 10:97-105. Round F E 1981 The Ecology of Algae. Cambridge University Press, Cambridge. Ryther J H & Dunstan W M 1971 Nitrogen, phosphorus, and eutrophication in the coastal marine environment. Science 271: 1008-1013. Sand-Jensen K & Borum J 1991 Interactions among phytoplank¬ ton, periphyton, and macrophytes in temperate freshwaters and estuaries. Aquatic Botany 41:137-175. Silberstein K, A W Chiffings & A J McComb 1986 The loss of seagrass in Cockburn Sound, Western Australia. III. The effect of epiphytes on productivity of Posidonia australis Hook. F. Aquatic Botany 24: 355-371. Smith S V 1984 Phosphorus versus nitrogen limitation in the marine environment. Limnology and Oceanography 29:1149- 1160. Stirn J 1994 Man-made eutrophication in the Mediterranean Sea. Mediterranean 4:8-23. Walker D I 1988 Methods for monitoring seagrass habitat. Victorian Institute of Marine Sciences, Melbourne. Working Pa¬ per Number 18. Walker D I, Hutchings P A & Wells F E 1991 Seagrass, sediment and infauna - a comparison of Posidonia australis, Posidonia sinuosa and Amplubolis antarctica, Princess Royal Harbour, South-Western Australia I. Seagrass biomass, productivity and contribution to sediments. In: Proceedings of the 3rd International Marine Biological Workshop: The Flora and Fauna of Albany, Western Australia (eds F E Wells, D I Walker, H Kirkman & R Lethbridge). Western Australian Museum, Perth, 597-610. Walker D I, Lukatelich R J & McComb A J 1987 Impacts of proposed developments on the benthic marine communities of Geographe Bay. Environmental Protection Authority, Perth. Technical Series 20. Walker D I & A J McComb 1992 Seagrass degradation in Aus¬ tralian coastal waters. Marine Pollution Bulletin 25: 191-195. Walter H 1979 Vegetation of the Earth in Relation to Climate and the Eco-physiological Conditions. Springer-Verlag, New York. 262 Journal of the Royal Society of Western Australia, 80:263-280, 1997 Foraminifera from Exmouth Gulf, Western Australia D W Haig Department of Geology & Geophysics, University of Western Australia, Nedlands WA 6907 email: dim ig@geol. u zua .edu.au Manuscript received January 1997; accepted September 1997. Abstract Two hundred and thirty-six benthonic species and six planktonic species are identified among Foraminifera present in Holocene sediment of Exmouth Gulf, in a water depth of 5-30 m. The benthonic microfauna comprises 20 agglutinated species (including 9 Lituolida, 1 Trochamminida, and 10 Textulariida), 74 porcellaneous species (Miliolida), and 142 hyaline species (including 4 Spirillinida, 27 Lagenida, 37 Buliminida, and 75 Rotaliida). Abundant species, at least at one site, include Ammolium australiensis, Textularia foliacea, Textularia lateralis , and Textularia oceanica among the agglutinated types; Parahauerinoides fragilissimus , Pencroplis pertusus, PlanispirineUa exigua, Pseudoniassilina australis , Quinqueloculina arenata, Q. philippinensis , Q. sp 8, Sigmoihauerma involuta , Sorites marginalis , Triloculina tricarinala among the porcellaneous species; and Ammotzia parkinsoniana, Amphistegina lessonii, A. sp cf A. papillosa , A. radiata, Asterorolalia gaimardi , Cibicides sp cf C. refulgens, Discorbinoides patelliformis, Elphidium sp cf E. advenum, E. crispum, E. sp 1, Heterostegina depressa, Operculina ammonoides , Pararotalia nipponica, and Rosalina cosymbosella among the hyaline (Rotaliida) species. Introduction There are no comprehensive published records of foraminiferal faunas from inner neritic environments along the 2500 km Western Australian coast between Derby and Augusta (Fig 1), and this region remains one of the least documented for Foraminifera on any continent (see review by Murray 1991). The few records that exist detail only part of the fauna and illustrate only a few species. Betjeman (1969) presented a broad description of the distribution of Foraminifera on the western continental shelf, based on scattered samples (one in the western Exmouth Gulf) from latitudes 18 °S to 34 °S. Mention was made of 191 species and 38 of these were illustrated. From the data presented in Betjeman's (1969) paper and the material from his study housed in the Museum collection of the Department of Geology and Geophysics at the University of Western Australia, it is not possible to confirm the identifications of many of the species named but not illustrated in the article. The most comprehensive reports of modern Foraminifera in Western Australia are studies by Mackenzie (1962) and Quilty (1977) on south coast estuarine and embayment faunas, and a major taxonomic study by Loeblich & Tappan (1994) who described and illustrated 946 species from sediment samples collected from the Sahul Shelf and Timor Trough (within a water depth range of 19 - 2716 m). The Sahul Shelf monograph of Loeblich & Tappan (1994) illustrates species by scanning electron micrographs which allow detailed comparison with Foraminifera found elsewhere along the shelf. The documentation of species along the coast, using a consistent nomenclature, will enhance the utility of the Foraminifera, particularly in biogeographic differentiation within this region. The purpose of this paper is to provide a comprehen¬ sive record of the foraminiferal species present in Holocene sediment in Exmouth Gulf, a major embayment on the central north-west coast of Western Australia (Figs 1, 2). This taxonomic record is a prerequisite to studies that may elucidate aspects of sedimentology, ecology, and biogeography. Figure 1. Map of Western Australia showing location of Exmouth Gulf and main sectors of the western continental shelf. © Royal Society of Western Australia 1997 263 Journal of the Royal Society of Western Australia, 80(4), December 1997 Material and Methods The species identified in the present study were recovered from 68 sediment grab samples taken throughout Exmouth Gulf (Fig 2) during the AIMS RV Lady Basten cruise 669/672, in September - October 1994. The samples come from a water-depth range of 5-30 m. A portion of each sample was washed through 2 mm, 150 pm, and 63 pm sieves. At least 150 Foraminifera were systematically picked from the 150 pm - 2 mm sediment fraction, and comprehensive selective picks were made of rare species in all sediment fractions. Physiographic setting The environmental setting of Exmouth Gulf was described by Brown (1988) who recorded the sea-floor sediment as mainly red-brown and grey-brown muddy quartz sand. Exmouth Gulf occupies a hot, semi-arid region and, according to Brown (1988), winter salinities within the Gulf may range from 35 %o to 39 %», north to south, and winter water temperatures may vary from 21.5 0 to 18 °C in the same direction (based on a 1972 survey). During the September 1994 study by McCook et al. (1995), the highest salinity noted in the Gulf was 38.5 %o. Strong currents are present, with spring-tide velocities about 0.5 m s'1 in deep areas, 1 m s1 on shallow open flats, and several metres per second in tidal channels (Brown 1988). McCook et al. (1995) noted very turbid conditions with much suspended material in the water due to rough sea conditions and strong currents. At the time of collection, the foraminiferal tests in the sand probably included specimens hundreds or thousands of years old as well as recently dead skeletons and com¬ paratively rare living individuals. The surficial sediment on the Gulf floor is doubtless reworked by currents, bioturbation, and, in some areas, by commercial trawl¬ fishing. Specimens eroded from Pleistocene rocks and in¬ corporated in the modern sediment have been discarded in this study. These are differentiated from the modern types by cement-infilled, poorly preserved tests. Record of Foraminifera The species are arranged alphabetically within Orders recognized by Loeblich & Tappan (1994). Generic defini¬ tions follow Loeblich & Tappan (1987), except for: (1) the Lagenida where genera are interpreted following Jones (1994); (2) the Miliolida where QuinquelocuHna is inter¬ preted in a broader sense (after Haig 1988); and (3) recent revisions by Revets (1990, 1991, 1993, 1996) among the Buliminida and Rotaliida. For each species, the original generic attribution is given in square brackets if the current generic designation differs from the original. This allows easy reference to Ellis & Messina's (1940 et seq) Catalogue of Foraminifera which includes a copy of the original description of the species. Previous published records of the species from Western Australian waters, confirmed by examining original material or high quality published illustrations, are given in partial synonymies. Only those species not illustrated in Loeblich & Tappan's (1994) Sahul Shelf monograph are figured here (Figs 3-7). All materials from the study, including scanning electron micrographs of all the identified species, are housed in the micropalaeontological collection of the Department of Geology & Geophysics, The University of Western Australia. Order Lituolida Ammotium australiensis (Collins); [. Ammomarginulina ]; Figure 3:1. Eggerelloides australis (Collins); [Eggerella]; Figure 3:2. Gaudryina convexa (Karrer); [Textilaria]; Figure 3:3. Gaudryina convexa (Karrer); Burdett et al, 1963 (revision of species); Quilty 1977, Fig 14, Hardy Inlet. Gaudryina sp; Figure 3:4. Haplophragmoides pusillus Collins. Haplophragmoides pusillus Collins; Loeblich & Tappan 1994, PI 7, Figs 1-7, Sahul Shelf. Nouria pwlymorpyhinoides Heron-Alien & Earland; Figure 3:5. Nouria polymorphinoides Heron-Alien & Earland; Mackenzie 1962, PI 1, Fig 3, Oyster Harbour. IReophax sp 1; Figure 3:6. Reophax sp 2; Figure 3:7. Velcroninoides jeffreysii (Williamson); [Nomonina]; Figure 3:11,12. Order Trochamminida Paratrochammina simplissima (Cushman & McCulloch); [Trocliammina]. Paratrochammina simplissima (Cushman & McCulloch); Loeblich & Tappan 1994, PI 24, Figs 1-12, Sahul Shelf. 264 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 3+ Agglutinated (1-12) and porcellaneous (13-25) Foraminifera (bar scales = 0.1 mm). 1, Ammotium australiensis (Collins) from sample 1134. 2, Eggerelloides australis (Collins) from sample 1116. 3, Gaudtyina convexa (Karrer) from sample 1027. 4, Gaudryina sp from sample 1021. 5, Nonna polymorphinoides Heron-Alien & Earland from sample 999. 6, IReophax sp 1 from sample 1114. 7, Reophax sp 2 from sample 1134. 8, Textularia kerimbaensis Said from sample 1029. 9, Textularia lateralis Lalicker from sample 1134. 10, Textularia occidentals Heron-Alien & Earland from sample 1134. 11, 12, Veleroninoides jeffreysii (Williamson) from sample 1029. 13, 14, Affinetrina bassensis (Parr) from sample 1095. 15, 20, Affinetrina sp from sample 1080. 16, Borelis schlumbergeri (Reichel) from sample 999. 17, 18, Cribromiliolinella milletti (Cushman) from sample 999. 19, Hauerina sp from sample 1024. 21, Miliolinella sp cf M. baragzvanathi (Parrj from sample 1008. 22, Miliolinella sp from sample 1024. 23, 24, Quinqueloculina barnardi Rasheed from sample 1134. 25, Quinqueloculina distorqueata Cushman from sample 1008. 265 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 4. Porcellaneous (1-24) and hyaline-spirillinid (25-27) Foraminifera (bar scales = 0.1 mm). 1, Quinqueloculina neostriatula Thalmann from sample 1128. 2,3, Quinqueloculina pittensis Albani from sample 1134. 4, 5, Quinqueloculina poeyana d'Orbigny from sample 1024. 6, 7, Quinqueloculina spl from sample 1027. 8, 9, Quinqueloculina sp 3 from sample 1134. 10, 11, Quinqueloculina sp 4 from sample 1134. 12, 13, Quinqueloculina sp 5 from sample 999. 14, 15, Quinqueloculina sp 6 from sample 1134. 16, 17, Quinqueloculina sp 7 from sample 1134. 18, 19, Quinqueloculina sp 8 from sample 1134 (different specimens). 20, 21, Sigmoilinella tortuosa Zheng from sample 999. 22, Spiroloculina angulata Cushman from sample 1027. 23, Triloculina bamardi Rasheed from sample 1024. 24, Triloculina papuaensis Rasheed from sample 1128. 25- 27, Mychostomina sp from sample 1008 (25 and 26, same specimen). 266 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 5. Hyaline Foraminifera: Spirillinida (1,2), Lagenida (3-10), Buliminida (11-25) and Rotaliida (26, 27) [bar scales - 0.1 mm]. 1, 2, Conicospirillinoides sp from sample 1091. 3, Dentalina sp 1 from sample 1024. 4, Dentalina sp 2 from sample 1010. 5. Fissurina sp 1 from sample 1134. 6, Fissurina sp 2 from sample 1029. 7, Glcmdulina sp from sample 1008. 8, Guttulina sp from sample 999. 9, Pseudoglandulina sp from sample 999. 10, Webbinella sp from sample 1029 (cemented to shell fragment). 11, Angulogerina sp 1 from sample 1067. 12, Angulogerina sp 2 from sample 999. 13, Bolivina striatula (Cushman) from sample 999. 14, Bolivina sp I from sample 999. 15, Bolivina sp 2 from sample 999. 16-18, Elongobula milletti (Cushman) from sample 1134 (different specimens). 19, Evolvocassidulina sp from sample 1021. 2o" Globocassidulina sp from sample 1095. 21, Reussella sp 1 from sample 999. 22, Reussella sp 2 from sample 1134. 23-25, Reussella sp 3 from sample 1134 (different specimens). 26, Acervulina mahabeti (Said) from sample 1021. 27, Ammonia sp from sample 1023. 267 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 6. Hyaline Foraminifera: Rotaliida (bar scales - 0.1 mm). 1, Ammonia sp from sample 1023 (same specimen as Fig 4:27). 2, 3, Anomalinulla glabrata (Cushman) from sample 999. 4, 5, Astrononion sp from sample 1008. 6-8, ICibicides sp from sample 999°. 9, 10, Cibicidoides basilanensis (McCulloch) from sample 1134. 11, 12, Cibicidoides sp from sample 1027. 13, 14, Cribrobaggina socorroensis McCulloch from sample 999. 15, 16, Cymbaloporella tabellaefomiis (Brady) from sample 1029 (chambers, shown in 16 on ventral side of test, are broken). 17, 18, Discorbinella rhodiensis (Terquem) from sample 1134 (different specimens). 19, 20, Elphidium sp cf advenum (Cushman) from sample 1134. 21, 22, Elphidium albanii Hayward from sample 1134. 23, Elphidium sp cf E. craticulatum (Fichyel & Moll) from sample 1135. 24, 25, Elphidium morlonbayensis Albani & Yassini from sample 1134. 26, 27, Elphidium oceanicum Cushman from sample 1134 (different specimens). 268 Journal of the Royal Society of Western Australia, 80(4), December 1997 Figure 7. Hyaline Foraminifera: Rotaliida (bar scales = 0.1 mm). 1, 2, Elphidium sp cf E. striatopunctatus (Fichtel & Moll) from sample 1134 (different specimens). 3, Elphidium sp, from sample 1134. 4, 5, Glabra tella sp from sample 999. 6, 7, IHaynesina sp from sample 1067. 8, 9, Lamellodiscorbis melbyae Hansen & Revets from sample 999. 10, 15, Monspeliensina sp 2 from sample 1135. 11, 12, Murryinella murrayi (Heron- Allen & Earland) from sample 999 (different specimens). 13, 14, Neorotalia calcar (d'Orbigny) from sample 999. 16-18, Nonionoides sp from sample 1095. 19, 20, Pararotalia nipponica (Asano) from sample 1134. 21, Planogypsina acervalis (Brady) from sample 999. 22, Poroeponides lateralis (Terqueum) from sample 1116. 23, 24, ?Saintclairoides sp from sample 999. 25, 26, Schwantzia sp from sample 1067. 269 Journal of the Royal Society of Western Australia, 80(4), December 1997 Order Textulariida Clavulina multicamerata Chapman. Clavulina serventyi Chapman & Parr; Mackenzie 1962, PI 1, Fig 14, Oyster Harbour. Clavulina multicamerata Chapman; Loeblich & Tappan 1994, PI 47, Figs 11-15, Sahul Shelf. Clavulina pacifica Cushman. Clavulina pacifica Cushman; Mackenzie 1962, PI 1, Fig 13, Oyster Harbour; Loeblich & Tappan 1994, PI 47, Figs 16-24, Sahul Shelf. Sahulia barkeri (Hofker); [Textularia]. Sahulia barkeri (Hofker); Loeblich & Tappan 1994, PI 32, Figs 1-8, Sahul Shelf. Textularia cushmani Said. Textularia cushmani Said; Loeblich & Tappan 1994, PI 35, Figs 1-4, Sahul Shelf. Textularia foliacea Heron-Alien & Earland. Textularia foliacea Heron- Allen & Earland; Loeblich & Tappan 1994, PI 34, Figs 6-14, Sahul Shelf. Betjeman (1969) recorded T. foliacea as a widespread species on the Western Australia shelf. Textularia kerimbaensis Said; Figure 3:8. Textularia lateralis Lalicker; Figure 3:9. Textularia oceanica Cushman. Textularia oceanica Cushman; Loeblich & Tappan 1994, PI 40, Figs 15-17, Sahul Shelf. Textularia occidentalis Heron-Alien & Earland; Figure 3:10. This may be an extreme variant of T. foliacea, differing from typical T. foliacea by a much more flaring test. Textularia sp. Textularia secasensis ; Loeblich & Tappan 1994, PI 39, Figs 8-14 (not T. secasensis Lalicker & McCulloch), Sahul Shelf. T. secasensis has a subacute peripheral margin in the adult stage that differs from the broadly rounded mar¬ gin of the present species. Order Miliolida Affinetrina bassensis (Parr); [‘ Triloculina ]; Figure 3:13,14. This species was recorded by Betjeman (1969) as Triloculina bassensis, abundant in the Abrolhos Islands and at stations north of the islands. Affinetrina sp; Figure 3:15,20. This species is related to some of the forms included originally in Miliolina kerimbatica Heron-Alien & Earland 1915 (e.g. PI 43, Fig 19 of Heron-Alien & Earland, 1915). Amphisorus hemprichii Ehrenberg. Marginopora vertebralis Quoy & Gaimard; Mackenzie 1962, PI 3, Fig 15, Oyster Harbour. ? Marginopora vertebralis Blainville; Quilty 1977, Fig 25 (juvenile specimen), Hardy Inlet. Amphisorus hemprichii Ehrenberg; Loeblich & Tappan 1994, PI 109, Figs 7-13; PI 110, Figs 6,7, Sahul Shelf. Marginopora vertebralis is distinguished from Amphisorus hemprichii by having multiple layers of chamberlets rather than two-layers in the adult stage. Betjeman's (1969) determination of M. vertebralis ap¬ parently refers to A. hemprichii, based on specimens which he mounted on slides in the UWA collection. Articulina alticostata Cushman. Articulina alticostata Cushman; Loeblich & Tappan 1994, PI 104, Figs 5-10, Sahul Shelf. This species was recorded by Betjeman (1969) as being more abundant on northern rather than southern sec¬ tors of the Western Australian shelf. Articulina mucronata (d'Orbigny); [Vertebralina]. Articulina mucronata (d'Orbigny); Loeblich & Tappan 1994, PI 104, Figs 1-4, Sahul Shelf. Betjeman (1969) recorded this species as Articulina pacifica Cushman and found that it ranged north from Exmouth Gulf to the Rowley Shelf. Articulina sp. Articulina carinata Cushman; Loeblich & Tappan 1994, Pi 104, Figs 11-18, Sahul Shelf. This species lacks the distinctly keeled periphery and numerous fine longitudinal costae that characterise A. carinata Cushman. Borelis schlumbergeri (Reichel); [Neoalveolina]; Figure 3:16. This may be the species recorded by Betjeman (1969) as Alveolinella boscii from Exmouth Gulf and the Rowley Shelf. Comuspira planorbis Schultze. Cornuspira planorbis Schultze; Loeblich & Tappan 1994, PI 56, Figs 1-7, Sahul Shelf. Coscinospira acicularis (Batsch); [Nautilus (Lituus)]. Coscinospira acicularis (Batsch); Loeblich & Tappan 1994, pi. 107, Figs 5-10, Sahul Shelf. Cribromiliolinella milletti (Cushman); [Hauerina]; Figure 3:17,18. Miliola sp A of Betjeman 1969, PI 18, Fig 23, Rowley Shelf. Dendritina ambigua (Fichtel & Moll); [Nautilus]. Dendritina ambigua (Fichtel & Moll); Loeblich & Tappan 1994, PI 108, Figs 1-4, Sahul Shelf. Edentostomina c.ultrata (Brady); [Miliolina]. Edentostomina cultrata (Brady); Loeblich & Tappan 1994, PI 63, Figs 8-12, Sahul Shelf. Euthymonacha polita (Chapman); [Peneroplis (Monalysidium)]. Euthymonacha polita (Chapman); Loeblich & Tappan 1994, PI 109, Figs 1-6, Sahul Shelf. Hauerina sp; Figure 3:19. The Exmouth Gulf specimens may be juvenile forms of H. pacifica Cushman. 270 Journal of the Royal Society of Western Australia, 80(4), December 1997 Inaequalina disparilis (Terquem); [Spiroloculina]. Inaequalina disparilis (Terquem); Loeblich & Tappan 1994, PI 64, Figs 11-18, Sahul Shelf. Mikrobelodontos bradyi (Barker); [Spiroloculina]. Spiroloculina sp A of Betjeman 1969, PI 19, Fig 16, Rowley and Dirk Hartog Shelf areas. Mikrobelodontos bradxji (Barker); Loeblich & Tappan 1994, PI 66, figs 1-8, Sahul Shelf. Miliolinella sp cf M. bamgwanathi (Parr); [Quinqueloculina]; Figure 3:21. Quinqueloculina lamarckiana d'Orbigny; Ouiltv 1977, Fig 18, Hardy Inlet. The Western Australian specimens lack the distinct obliquely curved costae of the type specimens of M. baragwanathi. Miliolinella philippinensis (McCulloch); [Pateoris]. Miliolinella philippinensis (McCulloch); Loeblich & Tappan 1994, PI 76, Figs 6-11, Sahul Shelf. Miliolinella pseudooblonga Zheng. Miliolinella circularis (Bornemann); Mackenzie 1962, PI 2, Fig 13. Oyster Harbour. Triloculinella pseudooblonga (Zheng); Loeblich & Tappan 1994, PI 88, Figs 7-18, PI 97, Figs 10-12, PI 98, Figs 1-3, 7-9, Sahul Shelf. Miliolinella quinquangula Loeblich & Tappan. Quinqueloculina sp A of Betjeman 1969, PI 19, Fig 4, Rowley, Dirk Hartog, and Rottnest Shelf areas. Miliolinella quinquangula Loeblich & Tappan 1994, PI 82, Figs 14-16, Sahul Shelf. Miliolinella sub orbicularis (d'Orbigny); [Triloculina]. Miliolinella sp B of Betjeman 1969, PI 18, Fig 24, Rottnest Shelf. Miliolinella sub orbicularis (d'Orbigny); Loeblich & Tappan 1994, PI 89, Figs 1-9, Sahul Shelf. Miliolinella sp; Figure 3:22. Nodophthalmidium gracile Collins. Nodophthalmidium gracile Collins; Loeblich & Tappan 1994, PI 57, Figs 18,19, Sahul Shelf. Nubeculina advena Cushman. Reophax scorpiurus Montfort; Mackenzie 1962, PI 1, Fig 2, Oyster Harbour. Nubeculina advena Cushman; Loeblich & Tappan 1994, PI 59, Figs 1-12, Sahul Shelf. Parahauerinoides fragilissimus (Brady); [Spiroloculina]. Parahauerinoides fragilissimus (Brady); Loeblich & Tappan 1994, PI 87, Figs 1-6, Sahul Shelf. As Hauerina fragilissitna, this species was noted by Betjeman (1969) as having a preference for the tropi¬ cal-subtropical sectors of the Western Australian shelf. Peneroplis pertusus (Forskal); [Nautilus]. Peneroplis pertusus (Forskal); Mackenzie 1962, PI 3, Fig 1, Oyster Harbour; Loeblich & Tappan 1994, PI 110, Figs 1-5, Sahul Shelf. Peneroplis planatus (Fichtel & Moll); McKenzie 1962, PI 2, Fig 27, Oyster Harbour. Betjeman (1969) recorded both P. pertusus and P. planatus (species considered by the present author to be synonymous) as more abundant on the northern rather than the southern sectors of the Western Aus¬ tralian shelf. Planispirinella exigua (Brady); [Hauerina]. Planispirinella exigua (Brady); Loeblich & Tappan 1994, PI 57, Figs 7,8, Sahul Shelf. Betjeman (1969) recorded this species from the Dirk Hartog Shelf and Rowley Shelf. Pseudomassilina australis (Cushman); [Massilina]. Massilina secans var. tenuistriata Earland; Mackenzie 1962, PI 2, Fig 25, Oyster Harbour. Pseudomassilina australis (Cushman); Loeblich & Tappan 1994, PI 91, Figs 1-3, Sahul Shelf. Not Pseudomassilina australis (Cushman); Mackenzie, 1962. Pseudomassilina robusta Lacroix. Pseudomassilina robusta Lacroix; Loeblich & Tappan 1994, PI 90, Figs 1-4, Sahul Shelf. This species is similar to Pseudomassilina sp B of Hottinger et al. (1993, PI 45, Figs 1-10). Pseudopyrgo milletti (Cushman); [Biloculina]. Pseudopyrgo milletti (Cushman); Loeblich & Tappan 1994, PI 89, Figs 10,11, Sahul Shelf. Pyrgo denticulata (Brady); [Biloculina]. Pyrgo denticulata (Brady); Loeblich & Tappan 1994, PI 92, Figs 1,2, Sahul Shelf. Pyrgo striolata (Brady); [Biloculina]. Pyrgo striolata (Brady); Loeblich & Tappan 1994, PI 92, Figs 9-15, Sahul Shelf. Pyrgoella tenuiaperta (Huang); [Biloculinella]. Pyrgoella tenuiaperta (Huang); Loeblich & Tappan 1994, PI 94, Figs 10-14, PI 99, Figs 10-17, Sahul Shelf. This species is the same as Pyrgoella sp A of Haig (1988, PI 4, Figs 7,8) and Hottinger et al. (1993, PI 53, Figs 7,8). Quinqueloculina adiazeta Loeblich & Tappan. Quinqueloculina adiazeta Loeblich & Tappan 1994, PI 85, Figs 1-18, Sahul Shelf. Contrary to Loeblich & Tappan's (1994, p. 48) proposi¬ tion, this species is not the same as Q. cf Q. berthelotiana of Haig (1988, p. 233, PI 4, Figs 23-26), but resembles Q. cf Q. rugosa of Haig (1988, p. 234, PI 8, Figs 1-5). Quinqueloculina agglutinans d'Orbigny. Quinqueloculina agglutinans d'Orbigny; Betjeman 1969, Rowley Shelf and Dirk Hartog Shelf. Agglutinella agglutinans (d'Orbigny); Loeblich & Tappan 1994, PI 70, Figs 1-9, Sahul Shelf. Quinqueloculina arenata Said. Siphonaperta ammophila (Parr); Mackenzie 1962, Oyster Harbour. 271 Journal of the Royal Society of Western Australia, 80(4), December 1997 Agglutvnella arenata (Said); Loeblich & Tappan 1994, PI 69, Figs 9-11; PI 70, Figs 10-15; PI 74, Figs 10-13, Sahul Shelf. Quinqueloculina bamardi Rasheed; Figure 3:23,24. Quinqueloculina crassicarinata Collins. Quinqueloculina crassicarinata Collins; Loeblich & Tappan 1994, PI 77, Figs 4-12, Sahul Shelf. Quinqueloculina distorqueata Cushman; Figure 3:25. Quinqueloculma ebumea (d'Orbigny); [Triloculina]. Miliolinella oblonga (Montagu); Mackenzie 1962, Oyster Harbour. Pseudolachlanella slitella Langer; Loeblich & Tappan 1994, PI 73. Figs 16-18; PI 101, Figs 1-3, Sahul Shelf. Quinqueloculina neostriatula Thalmann; Figure 4:1. Qumqneloculina sp C of Betjeman 1969, PI 19, Fig 5, Rowley Shelf. Quinqueloculina parkeri (Brady); [Miliolina]. Lachlanella parkeri (Brady); Loeblich & Tappan 1994, PI 74, Figs 1-6, Sahul Shelf. Betjeman (1969) recorded this species from the Abrolhos Islands and areas north on the Western Aus¬ tralia shelf. Quinqueloculina philippinensis Cushman. Qumqneloculina philippinensis Cushman; Loeblich & Tappan 1994, PI 81, Figs 1-10, Sahul Shelf. A broad range of ornament and apertural neck develop¬ ment is included here (similar to the range shown by the Papuan specimens illustrated by Haig 1988). Betjeman (1969) recorded the species from the Dirk Hartog Shelf and Rowley Shelf. Quinqueloculina pittensis Albani; Figure 4:2,3. Quinqueloculina subarenaria Cushman; Mackenzie 1962, PI 2, Fig 22, Oyster Harbour. Quinqueloculina poeyana d'Orbigny; Figure 4:4,5. Quinqueloculina poeyana d'Orbigny; Mackenzie 1962, Oyster Bay. The Papuan specimens identified as Q. poeyana by Haig (1988, PI 7, Figs 15-17) do not belong to this spe¬ cies (see Loeblich & Tappan 1994, p. 51). Quinqueloculina quinquecarinata Collins. Quinqueloculina quinquecarinata Collins; Loeblich & Tappan 1994, PI 79, Figs 13-18, Sahul Shelf. Quinqueloculina sulcata d'Orbigny. Quinqueloculina sulcata d'Orbigny; Loeblich & Tappan 1994, PI 82, Figs 1-6, Sahul Shelf. Not Quinqueloculina sulcata d'Orbigny; Mackenzie 1962. Quinqueloculina tropicalis Cushman. Quinqueloculina tropicalis Cushman; Loeblich & Tappan 1994, PI 78, Figs 13-15, Sahul Shelf. Quinqueloculina vandiemeniensis Loeblich & Tappan. Quinqueloculina vandiemeniensis Loeblich & Tappan; Loeblich & Tappan 1994, PI 83, Figs 1-3, Sahul Shelf. Quinqueloculina sp 1; Figure 4:6,7. ? Quinqueloculina granulocostata Germeraad; Mackenzie 1962, PI 2, Fig 18, Oyster Harbour. The costate ornament resembles that in Q. granulocostata, but the elongate apertural tooth differs from the bifid tooth in Q. granulocostata. Quinqueloculina sp 2. Quinqueloculina costata d'Orbigny; Mackenzie 1962, PI 2, Fig 17, Oyster Harbour. Quinqueloculina undulata d'Orbigny; Loeblich & Tappan 1994, PI 81, Figs 11-13, Sahul Shelf. In Papuan Lagoon assemblages, this species was misidentified as Q. poeyana by Haig (1988, p. 234, PI 7, Figs 15-17). Quinqueloculina sp 3; Figure 4:8,9. This species may be the species identified by Loeblich & Tappan (1994, p. 49, PI 80, Figs 13-15) as Q. incisa Vella. It may be related to Quinqueloculina patagonica d'Orbigny. Quinqueloculina sp 4; Figure 4:10,11. This species is close to the type recorded from the Papuan Lagoon as Qumqneloculina sp A by Haig (1988, PI 9, Figs 1-3). Quinqueloculina sp 5; Figure 4:12,13. This species was recorded from the Papuan Lagoon as Quinqueloculina sp C by Haig (1988, PI 9, Figs 7-10). Quinqueloculina sp 6; Figure 4:14,15. This species was recorded from the Papuan Lagoon as Quinqueloculina cf Q. oblonga (Montagu) by Haig (1988, PI 6, Figs 26-29). Quinqueloculina sp 7; Figure 4:16,17. This species may be related to Quinqueloculina multimarginata Said. Qumqneloculina sp 8; Figure 4:18,19. Quinqueloculina bosciana d'Orbigny; Mackenzie 1962, PI 2, Fig 16, Oyster Harbour. In chamber arrangement and apertural detail, this spe¬ cies resembles Pitella haigi Langer; but lacks distinct pseudopores. Rupertianella rupertiana (Brady); [Miliolina]. Rupertianella rupertiana (Brady); Loeblich & Tappan 1994, PI 106, Figs 1-14, Sahul Shelf. Schlumbergerina alveoliniformis (Brady); [Miliolina]. Schlumbergerina alveoliniformis (Brady); Loeblich & Tappan 1994, PI 72, Figs 9-11, Sahul Shelf. Sigmamiliolinella australis (Parr); [Quinqueloculina]. Sigmamiliolinella australis (Parr); Loeblich & Tappan 1994, PI 100, Figs 1-3, Sahul Shelf. Betjeman (1969) recorded " Miliolinella australis" as rare but widespread on the Western Australian shelf. Sigmoihauerina involuta (Cushman); [Hauerina]. Sigmoihauerina involuta (Cushman); Loeblich & Tappan 1994, PI 100, Figs 8-12, Sahul Shelf. 272 Journal of the Royal Society of Western Australia, 80(4), December 1997 Sigmoilinella tortuosa Zheng; Figure 4:20,21. Cycloforma collinsi Langer and Adelosina pascuaensis Koutsoukos & Falcetta are probably junior subjective synonyms of S. tortuosa . Sorites marginalis (Lamarck); [Orbulites]. Sorites marginalis (Lamarck); Loeblich & Tappan 1994, PI 112, Figs 1-5, Sahul Shelf. Spiroloculina angulata Cushman; Figure 4:22. Spiroloculina angulata Cushman; Mackenzie 1962, Oys¬ ter Harbour. Quilty (1977) noted this species from Hardy Inlet. Spiroloculina hadai Thalmann. Spiroloculina hadai Thalmann; Loeblich & Tappan 1994, PI 66, Figs 11-15, Sahul Shelf. Spiroloculina scrobiculata Cushman. Spiroloculina milletti Wiesner; Mackenzie 1962, pl.2. Fig 4, Oyster Harbour. Spiroloculina scrobiculata Cushman; Loeblich & Tappan 1994, PI 67, Figs 12-16, Sahul Shelf. Spiroloculina subimpressa Parr. Spiroloculina subimpressa Parr; Loeblich & Tappan 1994, PI 68, Figs 9-15, Sahul Shelf. Triloculina bamardi Rasheed; Figure 4:23. Triloculina littoralis Collins. Triloculina littoralis Collins; Loeblich & Tappan 1994, PI 95, Figs 11-13, Sahul Shelf. Triloculina papuaensis Rasheed; Figure 4:24. Triloculina tricarinata d'Orbigny; Mackenzie 1962, Oyster Harbour. T. papuaensis differs from T. tricarinata by the presence of a apertural neck. Triloculina tricarinata d'Orbigny. Triloculina tricarinata d'Orbigny; Loeblich & Tappan 1994, PI 96, Figs 1-7, Sahul Shelf. Betjeman (1969) recorded this species as widespread on the Western Australian shelf. Triloculina vespertilio Zheng. Triloculina striatotrigonula Parker & Jones; Quilty 1977, Fig 17, Hardy Inlet. Triloculina vespertilio Zheng; Loeblich & Tappan 1994, PI 97, Figs 1-6, Sahul Shelf. Vertebralina striata d'Orbigny. Vertebralina striata d'Orbigny; Mackenzie 1962, PI 1, Fig 20, Oyster Harbour; Loeblich & Tappan 1994, PI 60, Figs 1-7, Sahul Shelf. Betjeman (1969) noted this species as common in Cockburn Sound (Rottnest Shelf region). Wiesnerella auriculata (Egger); [Planispirina]. Wiesnerella auriculata (Egger); Loeblich & Tappan 1994, PI 62, Figs 1-3, Sahul Shelf. Order Spirillinida Conicospirillinoides semidecoratus (Heron-Alien & Earland); [Spirillina]. Conicospirillinoides semidecoratus (Heron-Alien & Earland); Loeblich & Tappan 1994, PI 50, Figs 1-9, Sahul Shelf. Conicospirillinoides sp; Figure 5:1,2. Mychostomina sp; Figure 4:25-27. Spirillina vivipara Ehrenberg. Spirillina vivipara Ehrenberg; Loeblich & Tappan 1994, PI 54, Figs 5-10, Sahul Shelf. The Exmouth specimens fit within the range of vari¬ ability shown by Brady's (1884, PI 85, Figs 1-4) figures. Order Lagenida Dentalina sp 1; Figure 5:3. Dentalina sp 2; Figure 5:4. Fissurina contusa Parr. Fissurina contusa Parr; Loeblich & Tappan 1994, PI 136, Figs 11-16, Sahul Shelf. Betjeman (1969) noted that this species is more abundant on southern rather than northern sectors of the West¬ ern Australian shelf. Fissurina crassiporosa McCulloch. Fissurina crassiporosa McCulloch; Loeblich & Tappan 1994, PI 155, Figs 15,16, Sahul Shelf. The Western Australian specimens appear closer to the holotype (McCulloch 1977, p. 98, PI 56, Fig 16) than the paratypes which appear to have a granular rather than smooth surface and lack a distinct basal spine. Fissurina globosocaudata Albani & Yassini. Fissurina robusta Zheng; Loeblich & Tappan 1994, PI 158, Figs 5-8, Sahul Shelf. F. robusta lacks the basal protrudences of F. globosocaudata , and has opaque U-shaped areas on ei¬ ther side of test which are absent in F. globosocaudata. Fissurina sp cf F. granulocostulata Zheng. Cerebrina granulocostata (Zheng); Loeblich & Tappan 1994, PI 135, Figs 1-7, Sahul Shelf. The apertural neck in the Western Australian specimens is more extended than in F. granulocostulata. Fissurina lacunata (Burrows & Holland); [Lagena]. Cerebrina lacunata (Burrows & Holland); Loeblich & Tappan 1994, PI 135, Figs 8-15, Sahul Shelf. Betjeman (1969) noted that this species is more abundant on southern rather northern sectors of the Western Australian shelf. Fissurina subquadrata Parr. Fissurina quadrata (Williamson); Loeblich & Tappan 1994, PI 155, Figs 1-10, Sahul Shelf. F. quadrata has a short neck and lacks the marginal groove of F. subquadrata. 273 Journal of the Royal Society of Western Australia, 80(4), December 1997 Fissurina sanpedroensis (McCulloch); [Lagenosolenia]. Lagenosolenia sanpedroensis McCulloch; Loeblich & Tappan 1994, PI 161, Figs 3,4, Sahul Shelf. Fissurina trinalimarginata (Loeblich & Tappan); [Duplella] DupleUa trinalimarginata Loeblich & Tappan 1994, PI 154, Figs 4-8, Sahul Shelf. A distinct double slit aperture is not present in the Exmouth specimens and was not shown on Loeblich & Tappan's (1994) figures of the type specimens. Fissurina wrightiatta (Brady); [Lagena]. Fissurina wrightiana (Brady); Loeblich & Tappan 1994, PI 158, Figs 1,2, Sahul Shelf. Fissurina sp 1; Figure 5:5. Fissurina sp 2; Figure 5:6. Glandulina sp cf G. symmetrica McCulloch. Glandulina symmetrica McCulloch; Loeblich & Tappan 1994, PI 168, Figs 6-8, Sahul Shelf. The species differs from the bathyal G. symmetrica by lacking the series of short diverging spines at the base of the test. Glandulina sp; Figure 5:7. Guttulina bartschi Cushman & Ozawa. Guttulina bartschi Cushman & Ozawa; Loeblich & Tappan 1994, PI 145, Figs 5-15, Sahul Shelf. Guttulina regina (Brady, Parker & Jones); [Polymorphina]. Guttulina regina (Brady, Parker &: Jones); Loeblich & Tappan 1994, PI 146, Figs 1-3. Guttulina sp; Figure 5:8. Lagena dorbignyi R.W. Jones. Lagena dorbignyi R.W. Jones; Loeblich & Tappan 1994, PI 138, Figs 6-9, Sahul Shelf. Lagena oceanica Albani. Pygmaeoseistron oceanicum (Albani); Loeblich & Tappan 1994, p. 80, PI 144, Figs 4-7, Sahul Shelf. Lenticulina domantayi (McCulloch); [Robulus]. Lenticulma domantayi (McCulloch); Loeblich & Tappan 1994, PI 121, Figs 1-8, Sahul Shelf. Nodosaria catesbyi d'Orbigny. Pyramidulina catesbyi (d'Orbigny); Loeblich & Tappan 1994, PI 116, Figs 10-12, Sahul Shelf. Oolina melosquamosa McCulloch. Favulina melosquamosa (McCulloch); Loeblich & Tappan 1994, PI 151, Figs 13-17, Sahul Shelf. Procerolagena gracillima (Seguenza); [Amphorina]. Hyalinonetrion distomapolitum (Parker & Jones); Loeblich & Tappan 1994, PI 137, Figs 10-12. Pseudoglandulina sp; Figure 5:9. Sigmoidella elegantissima (Parker & Jones); [Polymorphina]. Sigmoidella elegantissima (Parker & Jones); Mackenzie 1962, PI 3, Fig 4, Oyster Harbour; Loeblich & Tappan 1994, PI 148, Figs 4-12, Sahul Shelf. Webbinella sp; Figure 5:10. Order Buliminida Abditodendrix rhomboidalis (Millett); [Textularia]. Tortoplectella rhomboidalis (Millett); Loeblich & Tappan 1994, PI 216, Figs 1-6, Sahul Shelf. Angulogerina sp 1; Figure 5:11. Angulogerina sp 2; Figure 5:12 Bolivina striatula Cushman; Figure 5:13. Bolivina vadescens Cushman. Bolivina vadescens Cushman; Loeblich & Tappan 1994, PI 214, Figs 1-4,7-12, Sahul Shelf. Bolivina variabilis (Williamson); [ Textularia ]. Bolivina variabilis (Williamson); Loeblich & Tappan 1994, PI 216, Figs 7-15, Sahul Shelf. Bolivina sp 1; Figure 5:14. Bolivina sp 2; Figure 5:15. Bulimina marginata d'Orbigny. Bulimina marginata d'Orbigny; Loeblich & Tappan 1994, PI 242, Figs 1-4, Sahul Shelf. Buliminoides williamsonianus (Brady); [Bulimina]. Biiliminoides williamsonianus (Brady); Loeblich & Tappan 1994, PI 297, Figs 1-9, Sahul Shelf. Quilty (1977) noted this species in Hardy Inlet. The suprageneric classification of the species is uncertain (Revets 1990). Chrysalidinella dimorpha (Brady); [Chrysalidina]. Chrysalidinella dimorjaha (Brady); Loeblich & Tappan 1994, PI 252, Figs 7-13, Sahul Shelf. Elongobula sp cf E. cochlea (Wiesner). Floresina durrandi Revets; Loeblich & Tappan, PI 245, Figs 1-6, Sahul Shelf. Elongobula milletti (Cushman); [Buliminella]; Figure 5:16-18. Evolvocassidulina sp; Figure 5:19. Fursenkoina pauciloculata (Brady); [Virgulina]. Fursenkoina pauciloculata (Brady); Loeblich & Tappan 1994, PI 256, Figs 1-5, Sahul Shelf. The Exmouth specimens differ from typical F. paucilocidata in having a few more chambers in the adult test. Globocassidulina sp; Figure 5:20. Hopkinsinell a glabra (Millett); [Uvigerina]. Hopkinsinella glabra (Millett); Loeblich & Tappan, PI 232, Figs 1-11, Sahul Shelf. Loxostomina costatapertusa Loeblich & Tappan. Loxostomina costatapertusa Loeblich & Tappan 1994, PI 234, Figs 1-12, Sahul Shelf. Loxostomina costulata (Cushman); [Bolivina]. Loxostomina costulata (Cushman); Loeblich & Tappan 1994, PI 232, Figs 12-16, Sahul Shelf. This identification follows the interpretation of Hottinger et al. (1993) who showed a range of ornament from a few costae (as in Cushman's type) to numerous costae over the entire test. 274 Journal of the Royal Society of Western Australia, 80(4), December 1997 Loxostomina limbata (Brady); [Bolivina]. Loxostomina limbata (Brady); Loeblich & Tappan 1994, PI 233, Figs 1-8, Sahul Shelf. Specimens with few faint costae over the initial part of test are transitional to L. costulata. Millettia limbata (Brady); [Sagrina]. Millettia limbata (Brady); Loeblich & Tappan, PI 255, Figs 7-8, Sahul Shelf. Neocassidulina abbreviata (Heron-Alien & Earland); [Bolivina]. Neocassidulina abbreviata (Heron-Alien & Earland); Loeblich & Tappan 1994, p.131, PI 258, Figs 1-7, Sahul Shelf. This species is synonymous with Bolivina makiyamai Ishizaki. Haig (1993) suggested that transitional morphotypes may exist between N. abbreviata and Neocassidulina capitata (Cushman). Although N. capitata has not be found in Exmouth Gulf, morphotypes of N. abbreviata trending toward the more inflated, more loosely embracing chamber arrangement of N. capitata are present. Neouvigerina ampullacea (Brady); [Uvigerina]. Neouvigerina ampullacea (Brady); Loeblich & Tappan 1994, PI 246, Figs 9-19, Sahul Shelf. Radiatobolivina okinawaensis Hatta. Radiatobolivina okinawaensis Hatta in Hatta & Ujiie 1992, p. 205-206, PI 51, Figs la-2c, 3-5. Krebsina subtenuis (Cushman); Loeblich & Tappan 1994, PI 146, Figs 12-16; Sahul Shelf. Radiatobolivina sp. Krebsina pilasensis (McCulloch); Loeblich & Tappan 1994, PI 146, Figs 10,11. The relationship of Krebsina McCulloch to Radiatobolivina Hatta is uncertain, as is the suprageneric classification of these taxa. Reussella sp 1; Figure 5:21. Reussella sp 2; Figure 5:22 Reussella sp 3; Figure 5:23-25. Rugobolivinella elegans (Parr); [Bolivinella]. Rugobolivinella elegans (Parr); Loeblich & Tappan 1994, PI 220, Figs 1-6, Sahul Shelf. Betjeman (1969) noted the species (as Bolivinella elegans) from Exmouth Gulf and the Rowley Shelf. Sagrinella jugosa (Brady); [Textularia]. Sagrina jugosa (Brady); Loeblich & Tappan 1994, PI 237, Figs 12-17, Sahul Shelf. The generic attribution follows Revets (1996). Sagrinella sp. Sagrina zanzibarica (Cushman); Loeblich & Tappan 1994, PI 238, Figs 12-17, Sahul Shelf. The species differs from Bolivina zanzibarica Cushman in apertural details (having a high-arched aperture without projecting toothplate in contrast to an ellipti¬ cal aperture with distinct raised lip) and in having more inflated, less angulate adult chambers. Sagrinella scutata Saidova. Sagrinella scutata Saidova; Loeblich & Tappan 1994, PI 236, Figs 9, 10, Sahul Shelf. The generic attribution is uncertain. Haig (1993) included similar morphotypes in "Sagrina" gr. convallarium (Millett); however, the apparent broad range of vari¬ ability recorded in the Papuan assemblage is not present among the Exmouth Gulf specimens. Sagrinopsis fimbriata (Millett); [Bigenerina]. Sagrinopsis fimbriata (Millett); Loeblich & Tappan 1994, PI 239, Figs 1-10, Sahul Shelf. Sigmavirgulina tortuosa (Brady); [Bolivina]. Sigmavirgulina tortuosa (Brady); Loeblich & Tappan 1994, PI 261, Figs 1-10, Sahul Shelf. Betjeman (1969) noted this species is more abundant on southern sectors of the Western Australian shelf. Siphogenerina raphana (Parker & Jones); [Uvigerina]- Siphogenerina raphana (Parker & Jones); Mackenzie 1962, Oyster Harbour; Loeblich & Tappan 1994, PI 240, Figs 1-3, ?4-5, 6-11, Sahul Shelf. Quilty (1977) noted the species (as Rectobolivina raphanus) in Hardy Inlet. Siphouvigerina porrecta (Brady); [Uvigerina]. Siphouvigerina porrecta (Brady); Loeblich & Tappan, PI 247, Figs 6-11, Sahul Shelf. Valvobifarina mackinnoni (Millett); [Bifarina]. Valvobifarina mackinnoni (Millett); Loeblich & Tappan 1994, PI 254, Figs 9-12, Sahul Shelf. Order Rotaliida Acervulina mahabeti (Said); [ Planorbulina ]; Figure 5:26. Planorbulina acervalis Brady; Mackenzie 1962, PI 3, Fig 16, Oyster Harbour. The identification of this species follows Hottinger el al. (1993, p. 122,123, PI 165, Figs 1-7, PI 166, Figs 1-8). The specimen figured by Loeblich & Tappan (1994, PI 323, Figs 11-13) may be a juvenile of the species. Ammonia parkinsoniana (d'Orbigny); [Rosalina]. Ammonia parkinsoniana (d'Orbigny); Loeblich & Tappan 1994, PI 368, Figs 7-16; Sahul Shelf. ? Streblus beccarii (Linne); Mackenzie 1962, PI 3, Fig 18, Oyster Harbour. Ammonia supera Belford. Ammonia supera Belford; Loeblich & Tappan 1994, PI 370, Figs 7-9, Sahul Shelf. Ammonia sp; Figures 5:27, 6:1. This seems to be the same form illustrated by Loeblich & Tappan (1994, PI 387, Figs 4-6) as Notorotalia inornata Vella. Unlike Vella's types, the Exmouth specimens have an open umbilicus and some (e.g. Fig 5:27) have a prominent umbonal plug. A peripheral keel, prominent in N. inornata, is absent in the Western Australian species. Amphistegina lessonii d'Orbigny. Amphistegina lessonii d'Orbigny; Loeblich & Tappan 1994, PI 340, Figs 1-9, Sahul Shelf. 275 Journal of the Royal Society of Western Australia, 80(4), December 1997 Betjeman (1969) noted this species as widespread on the Western Australian shelf, and Quilty (1977) re¬ corded it from Hardy Inlet. Amphistegina sp cf A. papillosa Said. Amphistegina papillosa Said; Loeblich & Tappan 1994, PI 339, Figs 4-7; PI 341, Figs 1-7, Sahul Shelf. The Western Australian specimens are generally smaller and do not have the intensity of ornament shown by the type specimen. Betjeman (1969) recorded "Amphistegina radiata var. papillosa " from Shark Bay and areas north on the Western Australian shelf. Amphistegina radiata (Fichtel & Moll); [Nautilus]. Amphistegina radiata (Fichtel & Moll); Loeblich & Tappan 1994, PI 339, Figs 8-11, PI 341, Figs 8-10. Angulo discorbis comigatus (Millett); [Discorbina] Angidodiscorbis corrugatiformis (McCulloch); Loeblich & Tappan 1994, PI 290, Figs 8-10, Sahul Shelf. Anomalinulla glabrata (Cushman); [Anomalina]; Figure 6:2,3. Anomalinoides globulosa (Chapman & Parr); Loeblich & Tappan 1994, PI 355, Figs 4-5, 10-13 (not Figs 7,8; not PI 354, Figs 11-12). Asanonella tubulifera (Heron-Alien & Earland); [Truncatidina]. Asanonella tubulifera (Heron-Alien & Earland); Loeblich & Tappan 1994, PI 337, Figs 1-10, Sahul Shelf. Ashbrookia omata McCulloch. Ashbrookia tuberculata Ujiie; Loeblich & Tappan 1994, PI 262, Figs 1-3, Sahul Shelf. The Western Australian specimens lack the distinct radial grooves around the aperture that characterise A. tuberculata ; and have the matte granular surface of the spiral side of A. omata. Asterorotalia gaimardi (d'Orbigny); [Turbinulina]. Asterorotalia gaimardi (d'Orbigny); Loeblich & Tappan 1994, PI 372, Figs 1-7, Sahul Shelf. Astrononion sp; Figure 6:4,5. Baggina bubnanensis McCulloch. Baggina bubnanensis McCulloch; Loeblich & Tappan, PI 264, Figs 5-10, Sahul Shelf. Cancris auriculas (Fichtel & Moll); [Nautilus]. Cancris auriculus (Fichtel & Moll); Loeblich & Tappan 1994, PI 265, Figs 7-10, Sahul Shelf. The specimens figured by Loeblich & Tappan (1994, PI 266, Figs 1-13) as C. carinatus (Millett) may belong to this species; they lack the triangular cross-section in the last chamber that is characteristic of Millett's species. Caribbeanella sp cf C. ogiensis (Matsunaga); [Oinomikadoina]. Caribbeanella ogiensis (Matsunaga); Loeblich & Tappan 1994, PI 325, Figs 1-10, Sahul Shelf. In contrast to the types of C. ogiensis , the Western Aus¬ tralian specimens maintain an angulate margin in the adult stage and have a more lobulate periphery in the last few chambers. Caribbeanella elatensis Perelis & Reiss has a thicker test. Caribbeanella philippinensis McCulloch. Caribbeanella philippinensis McCulloch; Loeblich & Tappan 1994, PI 324, Figs 1-9, Sahul Shelf. Cibicides sp cf C. refulgens Montfort. Cibicides refidgens Montfort; Mackenzie 1962, PI 3, Fig 33, Oyster Harbour; Quilty 1977, Figs 42,43, Hardy Inlet; Loeblich & Tappan 1994, PI 318, Figs 7-9, Sahul Shelf. The Western Australian specimens have 9-10 cham¬ bers in the final whorl in contrast to 6 chambers in C. refidgens. The sutures on the umbilical side tend to be limbate, and the umbilical shoulder of the final cham¬ ber tends to be elevated above the shoulders of previ¬ ous chambers. Betjeman (1969) noted that " Cibicides refulgens" is common and widespread on the Western Australian shelf. ? Cibicides sp; Figure 6:6-8. A relationship between this species and Pulvinulina corticata Heron- Allen & Earland is suspected. Cibicidoides basilanensis McCulloch; Figure 6:9,10. An opaline sheen on the spiral side is a feature of the Exmouth specimens and McCulloch's (1977, p. 445, PI 187, Figs 1,2) types housed at the Smithsonian Insti¬ tution, Washington. Cibicidoides sp; Figure 6:11,12. Colonimilesia obscura McCulloch. Colonimilesia obscura McCulloch; Loeblich & Tappan 1994, PI 282, Figs 1-6, 13-14, Sahul Shelf. Cribrobaggina socorroensis McCulloch; Figure 6:13,14. Cymbaloporella tabellaeformis (Brady); [ Cymbalopora ]; Figure 6:15,16. Cymbaloporetta bradyi (Cushman); [Cymbalopora]. Cymbaloporella bradyi (Cushman); Mackenzie 1962, Oyster Harbour; Loeblich & Tappan 1994, PI 328, Figs 1-3, Sahul Shelf. Betjeman (1969) recorded this species as more abundant in northern sectors of the Western Australian shelf. Discorbinella bodjongensis (Le Roy); [Discorbis]. Discorbinella bodjongensis (Le Roy); Loeblich & Tappan 1994, PI 310, Figs 1-13, Sahul Shelf. Discorbinella rhodiensis (Terquem); [' Truncatidina ]; Figure 6:17,18. The identification follows Hottinger et al. (1993, p. 115, PI 150, Figs 5-9). Discorbinoides patelliformis (Brady); [Discorbina] . Discorbinoides minogasiformis Ujiie; Loeblich & Tappan 1994, PI 291, Figs 1-10, Sahul Shelf. The Western Australian specimens lack strong ornament on the spiral side. The specimens from Hardy Inlet that Quilty (1977, Figs 38,39) attributed to this species, do not belong here. Elphidium sp cf E. advenum (Cushman); [ Polystomella ]; Figure 6:19,20. The Exmouth specimens differ from the specimens illustrated as E. advenum by Loeblich & Tappan (1994, PI 379, Figs 1-4) in having a larger number of chambers 276 Journal of the Royal Society of Western Australia, 80(4), December 1997 in the final whorl, and more prominent retral pro¬ cesses across the sutures. The Exmouth specimens differ from Oyster Harbour types included in £. advenum by Mackenzie (1962) in having slightly fewer chambers in the final whorl and a sharper periphery, and from a specimen figured bv Quilty (1977, Fig 54) in having fewer chambers in the adult whorl. Elphidiurn albanii Hayward; Figure 6:21,22. Elphidium albanii Hayward 1997, p. 70,71, PI 6, Figs 1-5. Elphidiurn carteri Hayward. Elphidium carteri Hayward 1997, p. 71,72, PI 1, fig. 15, PI 6, Figs 8-12. Elphidium jenseni (Cushman); Loeblich & Tappan 1994, PI 381, Figs 4,5, ?l-3, Sahul Shelf. £. jenseni has more chambers in the adult whorl (around 17) than has E. carteri (12-14). Elphidium sp cf E. craticulatum (Fichtel & Moll); [Nauti¬ lus]; Figure 6:23. This species belongs to the group of Elphidium craticulatum (Fichtel & Moll), but differs from typical specimens by having interconnected vermiform ridges crossing the broad central umbo. Elphidium crispum (Linne); [Nautilus]. Elphidium crispum (Linne); Mackenzie 1962, PI 3, Fig 21, Oyster Harbour; Quilty 1977, Fig 55, Hardy Inlet; Loeblich & Tappan 1994, PI 378, Figs 4-6, Sahul Shelf. Betjeman (1969) recorded this species as widespread on the Western Australian shelf. Elphidium mortonbayensis Albani & Yassini; Figure 6:24,25. Elphidium mortonbayensis Albani & Yassini 1993, p. 30, Figs 74, 78. The Exmouth specimens have a more broadly rounded periphery than in the type specimens. Elphidium neosimplex McCulloch. Elphidium neosimplex McCulloch; Loeblich & Tappan 1994, PI 381, Figs 6-11, Sahul Shelf. Elphidium oceanicum Cushman; Figure 6:26,27. This is probably the same form as Albani & Yassini (1993, p. 21, Figs 27,28) recorded as Cribrononion schmitti (Cushman & Wickenden). The species has a greater number of chambers in the final whorl, and a greater number of more distinct retral processes cross¬ ing the sutures than has C. schmitti. Elphidium simplex Cushman. Elphidium simplex Cushman; Loeblich & Tappan 1994, PI 385, Figs 1-12, Sahul Shelf. Betjeman (1969) noted than E. simplex occurs from Shark Bay north on the Western Australian shelf. Quilty (1977) recorded the species much further south in Hardy Inlet. Elphidium sp cf E. striatopunctatus (Fichtel & Moll); [Nautilus]; Figure 7:1,2. Elphidium sp 1; Figure 7:3. Eupatellinella fastidiosa (McCulloch); [Patellinella]. Eupatellinella fastidiosa (McCulloch); Loeblich & Tappan 1994, PI 262, Figs 4-11, PI 263, Figs 1-6, Sahul Shelf. Some doubt exists are to the specific identity of the Western Australian specimens. The Exmouth specimens have a keeled margin with the keels of earlier chambers forming low ridges at the chamber contacts on the conical spiral surface. McCulloch (1977, p. 283) described the spiral surface of E. fastidiosa as smooth. Gavelinopsis praegeri (Heron-Alien & Earland); [Discorbina]. Gavelinopsis praegeri (Heron- Allen & Earland); Loeblich & Tappan 1994, PI 281, Figs 1-10, Sahul Shelf. Glabratella sp; Figure 7:4,5. ?Haynesitia sp; Figure 7:6,7. Heterolepa subhaidingeri (Parr); [Cibicides]. Heterolepa subhaidingeri (Parr); Loeblich & Tappan 1994, PI 359, Figs 1-13, Sahul Shelf. Heterostegina depressa d'Orbigny. Heterostegina depressa d'Orbigny; Loeblich & Tappan 1994, PI 389, Figs 1-6, PI 390, Figs 1-3, Sahul Shelf. Homotrema rubrum (Lamarck); [Millepora]. Homotrema rubrum (Lamarck) [ Millepora ]; Loeblich & Tappan 1994, PI 335, Figs 1-4, Sahul Shelf. Lamellodiscorbis melbyae Hansen & Revets; Figure 7:8,9. Lamellodiscorbis melbyae Hansen & Revets 1992, PI 4, Figs 4-6, 9. Discorbis dimidiatus (Jones & Parker) var. acervulinoides Parr; Mackenzie 1962, PI 3, Fig 9, Oyster Harbour. Discorbis dimidiatus (Jones & Parker); Quilty 1977, Figs 31,32, Hardy Inlet. ?Melonis sp. Anomalinoides globulosus (Chapman & Parr); Loeblich & Tappan 1994, PI 354, Figs 11-13, PI 354, Figs 11-13, PI 355, Figs 7-9 (not Figs 4,5,10-13). Millettiana millettii (Heron-Alien & Earland); [Cymbalopora). Millettiana millettii (Heron-Alien & Earland); Loeblich & Tappan 1994, PI 329, Figs 1-12, Sahul Shelf. Mitiiacina miniacea (Pallas); [Millepora]. Miniacina miniacea (Pallas); Loeblich & Tappan 1994, PI 335, Figs 5-6, Sahul Shelf. Monspeliensina sp 1. Ammonia convexa (Collins); Loeblich & Tappan 1994, PI 369, Figs 1-10, Sahul Shelf. A. convexa has a prominent umbonal plug and lacks the sutural clefts of Monspeliensina. Monspeliensina sp 2; Figure 7:10,15. Murrayinella murrayi (Heron-Alien & Earland); [Rotalia]; Figure 7:11-12. Most of the specimens figured by Loeblich & Tappan (1994, PI 293, Figs 1-10) seem atypical of this species. Neoconorbina terquemi (Rzehak); [Discorbina]. Neoconorbina terquemi (Rzehak); Loeblich & Tappan 1994, PI 284, Figs 1-12, Sahul Shelf. 277 Journal of the Royal Society of Western Australia, 80(4), December 1997 Betjeman (1969) recorded the species as more abundant on northern sectors of the Western Australian shelf. Quilty (1977) noted it from Hardy Inlet. Neorotalia calcar (d'Orbigny); [Calcarina]; Figure 7:13,14. Betjeman (1969) recorded " Calcarina calcar" from the Abrolhos Islands and areas north on the Western Aus¬ tralian shelf. Nonion sp cf N. subturgidurn (Cushman); [Nonionina]. Nonion subturgidurn (Cushman); Loeblich & Tappan 1994, PI 343, Figs 1-9, Sahul Shelf. The Western Australian species has a greater number of chambers (11-12) in the adult whorl than has N. subturgidium (8-9). Nonionoides sp; Figure 7:16-18. Nummulites venosus (Fichtel & Moll); [Nautilus]. Nummulites venosus (Fichtel & Moll) [Nautilus]} Loeblich & Tappan 1994, PI 388, Figs 5-9, Sahul Shelf. Bejeman (1965) recorded " Operculinella venosa" as more abundant on northern rather than southern sectors of the Western Australian shelf. Operculina ammonoides (Gronovius); [Nautilus]. Assilina ammonoides (Gronovius); Loeblich & Tappan 1994, PI 387, Figs 7-9, PI 388, Figs 1-4, Sahul Shelf. Operculina heterostcginoides Hofker. Operculina heterostcginoides Hofker; Loeblich & Tappan 1994, PI 390, Fig 4, Sahul Shelf. Orbitina carinata Sellier de Civrieux. Orbitina carinata Sellier de Civrieux; Loeblich & Tappan 1994, PI 275, Figs 7-12, Sahul Shelf. Pararotalia nipponica (Asano); [Rotalia]} Figure 7:19,20. Calcarina calcar d'Orbigny; Quilty 1977, Figs 52,53, Hardy Inlet. The species identification follows Ujiie (1966) who placed spinose Rotalia ozawai Asano into synonymy with P. nipponica. Planogypsina acervalis (Brady); [. Planorbulina ]; Figure 7:21. Planorbulinella larvata (Parker & Jones); [Planorbulina]. Planorbulinella larvata (Parker & Jones); Mackenzie 1962, PI 3, Fig 17, Oyster Harbour; Loeblich & Tappan 1994, PI 327, Figs 1-7, Sahul Shelf. Poroeponides lateralis (Terqueum); [Rosalina]; Figure 7:22. The Sahul Shelf specimens figured by Loeblich & Tappan (1994, PI 269, Figs 1-9) as Eponides cribrorepandus (Asano & UchiO) belong to P. lateralis as recognized by Hottinger el al. (1991). The figured Exmouth specimen represents an end member of the P. lateralis plexus. Rosalina cosymbosella Loeblich & Tappan 1994. Rosalina cosymbosella Loeblich & Tappan 1994, p. 140, PI 287, Figs 1-3, Sahul Shelf. Rosalina globularis d'Orbigny. Rosalina globularis d'Orbigny; Loeblich & Tappan 1994, PI 286, Figs 7-15, Sahul Shelf. Rotorbis auberi (d'Orbigny); [Rosalina]. Discorbis mira Cushman; Mackenzie 1962, PI 3, Fig 10, Oyster Harbour. Rotorbis auberi (d'Orbigny); Loeblich & Tappan 1994, PI 278, Figs 1-8. ?9-ll, Sahul Shelf. Betjeman (1969) recorded "Discorbis mira" as a wide¬ spread although rare species on the Western Australian shelf. ISaintclairoides sp; Figure 7:23,24. The generic designation is tentative. Loeblich & Tappan (1987) provisionally placed Saintclairoides McCulloch among the ceratobulimines (Order Robertinida), and suggested that the systematic position of the genus depended on additional information regard¬ ing the aperture and internal features of the test. The Exmouth specimen appears to lack an internal partition. Schwantzia sp; Figure 7:25,26. Siphonina tubulosa Cushman. Siphonina tubulosa Cushman; Loeblich & Tappan 1994, PI 299, Figs 1-10, Sahul Shelf. Siphoninoides laevigatas (Howchin); [Truncatulina]. Siphoninoides laevigatas (Howchin); Loeblich & Tappan 1994, PI 300, Figs 1-4, Sahul Shelf. Sphaerogypsina globula (Reuss); [Ceriopora]. Gypsina globulus (Reuss); Quilty 1977, Fig 60, Hardy Inlet. Sphaerogypsina globula (Reuss); Loeblich & Tappan 1994, PI 334, Figs 4-6, Sahul Shelf. Stomatorbina conccntrica (Parker & Jones); [Pulvinulina]. Stomatorbina conccntrica (Parker & Jones); Loeblich & Tappan 1994, PI 273, Figs 1-7, Sahul Shelf. Quilty (1977) noted this species, as Mississippina conccntrica , from Hardy Inlet. Order Globigerinida Globigerinoides ruber (d'Orbigny); [Globigerina]. Globigerinoides ruber (d'Orbigny); Loeblich & Tappan 1994, PI 203, Figs 1-9, PI 206, Figs 10-12, Sahul Shelf. Globigerinoides trilobus (Reuss); [Globigerina]. Globigerinoides trilobus (Reuss); Loeblich & Tappan 1994, PI 206, Figs 1-6, Sahul Shelf. Globorotalia menardii (Parker, Jones & Brady); [Rotalia]. Globorotalia menardii (Parker, Jones & Brady); Loeblich & Tappan 1994, PI 183, Figs 1-6. Globoturborotalita rubescens (Hofker); [Globigerina]. Globoturborotalita rubescens (Hofker); Loeblich & Tappan 1994, PI 208, Figs 1-12, Sahul Shelf. Neogloboquadrina humerosa (Takayanagi & Saito); [Globorotalia J. Neogloboquadrina humerosa (Takayanagi & Saito); Loeblich & Tappan 1994, PI 199, Figs 1-6, Sahul Shelf. Tinophodella ambitacrena Loeblich & Tappan. Tinophodella ambitacrena Loeblich & Tappan; Loeblich & Tappan 1994, PI 192, Figs 1-9, PI 200, Figs 1-6, Sahul Shelf. 278 Journal of the Royal Society of Western Australia, 80(4), December 1997 Table 1. Distributions of major foraminiferal groups and those species present in at least one sample at frequencies <5% . Abundances are given as percentages in systematic foraminiferal counts from the 150 pm - 2 mm sediment fraction. c £ ra O U- u c n Cl. 3 o u u TD OJ (t m 3 O ai Tv 3 c GJ to tt j J (T3 T3 ’§ £ 3 CQ 3' 3 1c o in QJ U QJ cu in T3 QJ '-n 5 '■T. § cs UJ "g _cs in y u QJ Cu in m 3 O QJ ax m ’n £ ■ SX 6 X 3 « 3 c! C3 2 .S: G GO CU V; CS ~s C! QJ /Si G rz O OS m QJ U QJ Cu s -§ 3 m 25 /■ 2 s. Cu in a R GJ « -13 s S3 .a Jkj u u Cu •3, S' § a xj tj cu cn cu in a n £ u. g- Q cs 5 a ,0 5 ,5X -S 0 1 w hJ Ch s c o -5 c e o £ c c J3 to c Jt "qj u cu rt It o ft) _c c 3 'to •S ■3 V. -3 c rj u §5 "eL ' § ■ £ Vj 2 _Q 2~ §- § s s $ G QJ C •S x3 ,‘n ■Xj /Si ,in *3. -3 § in tn G -S "3 rl 1 T, s g G • S "3 0 a < C/5 E QJ CO to < o X >-. X X to < £ £ £ O c- cs a, G • CX 1 <3 X •Aj V, X a Of Of <-r> L in iC f-H X Si X -X *0 .Vj Q sx XI 5X XI X. Sj xj X 0 a Ch m O 996 2 7 53 1 3 36 0 0 0 0 4 1 0 1 2 3 1 2 1 0 5 4 0 0 0 7 2 0 1 0 0 2 1 0 999 0 4 55 1 4 36 0 0 2 1 2 8 2 1 0 0 2 6 1 1 0 1 1 0 0 0 6 1 0 0 0 1 3 0 1002 0 3 53 1 0 42 0 0 1 0 1 10 1 1 3 1 0 1 0 1 0 7 4 0 0 0 10 1 1 2 1 0 10 0 1004 0 6 55 0 1 38 1 0 1 2 2 10 5 1 1 0 1 10 3 3 0 4 3 0 0 0 9 1 2 0 0 0 5 1 1006 0 5 50 1 0 45 0 1 2 1 1 13 2 1 0 2 0 1 2 1 1 4 1 0 0 0 6 3 2 2 1 1 11 0 1008 0 5 60 0 1 33 1 0 1 1 0 12 1 1 2 1 2 3 2 5 0 1 1 0 0 0 7 1 1 0 1 1 8 2 1010 1 10 56 2 4 27 0 0 3 5 0 3 2 0 1 5 7 2 0 2 0 3 2 0 1 0 1 1 2 1 0 2 0 1 1012 0 1 66 1 1 32 0 0 1 0 2 10 5 1 1 2 0 7 1 0 0 0 1 0 0 0 3 2 2 1 1 1 9 1 1015 1 11 49 0 1 39 1 0 3 0 2 3 3 0 2 3 4 4 0 6 4 0 1 0 0 0 6 2 2 1 2 1 7 0 1017 1 5 42 2 2 49 0 1 3 0 0 7 4 1 3 1 0 1 1 1 0 6 6 0 1 1 6 4 1 0 8 1 5 1 1021 2 8 41 1 4 47 0 1 2 2 2 5 4 1 1 0 1 2 2 2 0 3 3 0 1 2 0 1 1 0 2 0 2 3 1023 1 4 33 0 7 57 0 0 2 0 3 7 2 1 0 3 1 1 0 0 0 3 0 0 0 8 3 4 1 3 1 3 4 0 1024 2 4 36 1 5 53 0 0 1 1 0 4 1 0 1 1 2 1 1 1 1 5 3 0 0 7 1 5 1 4 1 2 3 0 1027 0 4 12 1 2 81 0 0 1 2 0 2 0 0 0 0 0 0 1 0 0 21 6 21 0 3 0 3 0 6 3 1 0 0 1029 0 13 43 0 0 44 0 3 1 3 0 11 2 2 4 6 0 0 0 2 0 3 3 0 1 4 0 11 3 4 1 2 3 0 1031 0 17 38 1 2 43 0 0 7 4 1 6 1 2 3 4 1 0 1 1 0 6 10 0 2 3 0 6 3 3 1 1 1 0 1033 0 9 33 4 1 53 0 1 1 6 0 5 0 1 3 6 0 0 0 1 0 6 6 0 0 4 0 16 2 2 0 1 0 0 1035 0 5 48 0 1 46 0 0 1 3 1 12 2 4 1 3 0 1 0 1 0 3 4 0 0 4 0 15 3 3 1 1 0 1 1040 0 8 38 1 2 52 0 1 3 3 3 7 0 1 0 3 1 0 2 3 0 8 5 0 1 1 1 2 0 0 6 1 5 1 1042 0 19 48 0 7 34 0 1 6 4 1 4 2 1 2 1 1 4 2 3 0 1 1 0 4 1 1 0 1 1 1 1 4 1 1044 0 11 55 0 4 28 0 2 2 6 3 8 5 4 0 0 2 3 1 3 0 1 1 0 1 0 3 0 2 0 2 0 1 2 1048 0 8 41 1 5 45 0 1 1 5 1 1 1 1 0 0 1 1 3 4 4 0 0 0 0 0 12 0 1 0 2 0 0 4 1050 1 30 44 1 0 26 0 2 10 6 0 7 1 1 3 6 0 1 1 1 0 6 3 0 0 1 0 0 3 3 1 0 0 1 1052 1 6 66 0 0 27 0 0 3 1 2 18 3 2 5 0 0 5 1 0 2 0 0 0 1 0 5 0 0 3 2 0 5 1 1053 0 12 47 0 3 38 0 3 4 3 1 5 8 2 0 0 1 1 4 3 3 2 0 0 3 1 3 0 3 1 1 0 6 0 1054 0 11 41 0 4 51 0 1 6 3 3 4 5 4 0 0 1 1 1 0 2 0 1 0 2 1 7 1 1 1 0 0 3 0 1056 0 14 45 1 3 37 2 2 6 1 2 2 3 1 1 0 0 2 3 2 1 1 0 0 0 0 5 0 2 0 2 0 3 1 1058 0 24 38 1 2 35 0 2 8 4 1 4 4 4 0 4 0 0 1 2 1 0 0 0 4 1 1 0 1 0 0 6 1 1 1060 0 18 42 3 3 34 0 3 10 4 1 4 2 0 3 8 0 0 1 2 0 1 2 0 2 1 1 0 0 2 1 2 0 1 1062 1 8 54 0 2 36 0 3 1 2 2 12 2 2 2 0 1 5 0 0 0 0 3 0 0 0 5 0 1 3 0 1 11 1 1064 0 23 48 0 1 29 0 2 9 7 1 8 2 2 5 7 0 2 0 1 0 2 2 0 2 2 1 0 5 5 0 1 1 2 1066 0 22 50 0 2 26 0 1 12 5 1 9 5 3 3 5 1 2 0 1 1 2 3 0 2 1 1 0 1 3 2 1 1 1 1067 1 12 50 0 5 32 1 1 6 3 3 10 3 4 2 4 0 4 1 0 1 3 2 0 0 0 5 1 2 6 1 0 2 0 1068 0 13 49 0 2 37 0 0 6 3 0 16 2 0 0 3 1 0 0 3 0 6 3 0 0 0 2 1 0 2 3 0 1 2 1069 0 12 60 1 1 27 0 0 6 2 1 10 5 2 2 6 2 1 0 3 0 0 4 0 2 0 0 0 2 4 0 2 0 0 1070 1 26 42 1 2 29 0 9 9 3 2 6 2 3 1 3 0 0 0 3 1 1 1 0 1 0 2 0 0 1 0 1 1 0 1071 0 5 72 1 0 23 0 1 1 3 1 20 1 0 0 3 0 7 0 1 0 0 0 0 0 0 7 1 2 1 0 0 3 1 1073 0 2 68 0 0 29 0 0 0 1 0 22 0 1 0 2 0 1 0 0 0 8 1 0 0 0 2 2 0 3 1 0 3 1 1075 0 17 46 2 2 32 0 3 7 2 2 2 2 3 2 2 1 0 2 6 2 1 1 0 7 0 2 0 1 2 0 1 1 0 1078 1 21 50 0 1 28 0 2 10 3 0 6 8 3 1 2 0 0 3 4 2 0 1 0 3 0 1 0 2 2 0 0 1 0 1080 1 5 51 1 4 39 0 2 2 1 2 412 2 2 0 2 2 3 0 2 0 0 0 5 0 4 0 0 0 2 2 0 4 1082 0 13 51 0 0 36 0 5 6 0 0 618 4 0 3 2 0 1 2 2 1 0 0 5 0 6 0 1 2 2 1 3 0 1083 0 11 55 0 0 34 0 0 5 3 1 22 1 0 2 2 0 0 0 0 0 0 4 0 1 0 1 2 0 7 1 2 1 0 1085 0 7 55 0 0 37 0 0 3 2 2 611 4 0 0 0 1 2 3 2 1 1 0 0 0 2 0 2 2 0 1 4 2 1090 0 3 71 1 0 26 0 0 1 0 3 1213 1 2 0 2 3 1 0 5 0 0 0 0 0 1 1 1 0 0 0 3 2 1091 0 3 69 0 0 29 0 0 1 2 0 34 1 0 1 1 0 0 0 0 1 5 0 0 1 0 2 9 1 7 0 0 1 0 1093 0 4 63 0 2 32 0 1 0 2 2 915 4 0 1 2 2 0 1 4 0 0 0 0 0 4 0 0 2 0 1 6 2 1095 0 5 49 1 3 42 0 0 2 2 2 6 8 2 2 1 0 2 1 0 1 9 8 0 0 0 4 1 1 0 2 0 7 0 1097 0 16 49 0 1 34 0 5 8 3 3 514 3 0 0 1 0 1 1 1 0 1 0 5 0 4 0 1 0 1 0 1 1 1099 1 19 33 0 3 45 0 4 6 1 0 2 3 2 0 1 1 3 1 1 1 0 0 0 3 0 7 0 0 2 0 0 0 3 1101 1 9 54 0 2 35 1 1 1 3 2 6 7 2 0 1 0 2 2 8 0 1 0 0 2 0 8 0 0 0 0 0 0 1 1103 1 17 37 0 2 44 1 7 4 0 0 5 1 4 0 0 0 2 5 3 2 0 1 0 3 0 10 0 0 2 2 0 0 0 1103 0 13 52 0 3 32 0 5 0 2 4 7 3 2 0 2 1 1 3 5 1 0 1 0 2 0 3 0 0 1 2 1 0 5 1105 1 14 44 1 1 41 6 4 2 1 2 4 6 2 0 0 0 0 2 3 0 0 0 0 2 0 8 0 0 3 1 0 0 3 1109 0 3 75 0 2 19 0 0 0 2 2 1415 1 1 0 1 0 0 1 2 0 1 0 0 0 4 0 1 0 0 0 2 1 1111 0 9 65 0 1 26 0 3 2 2 2 13 8 2 7 0 2 2 1 0 2 0 0 0 1 0 2 2 2 2 1 0 2 2 1114 0 12 66 0 1 21 0 0 3 4 1 9 9 4 4 0 0 4 1 1 0 1 0 0 0 0 2 0 1 4 0 1 1 3 1116 0 6 62 1 1 31 0 3 1 0 2 17 1 2 2 2 0 5 1 4 0 3 2 0 1 0 6 0 2 3 0 2 0 1 1118 0 10 56 0 2 33 0 1 5 0 2 12 5 1 1 0 2 4 0 3 0 1 2 0 1 0 6 0 1 2 1 0 1 3 1121 0 7 54 1 1 36 1 1 1 1 3 8 4 2 0 0 1 5 1 2 3 0 1 0 0 0 2 0 0 2 1 0 2 7 1127 0 1 60 1 0 38 0 0 1 1 1 12 3 0 1 9 2 2 0 1 0 0 1 0 0 0 2 6 0 2 0 0 11 1 1128 0 4 70 0 1 25 0 0 0 1 2 26 3 0 1 0 0 5 1 0 0 0 0 0 0 0 9 0 1 2 0 0 1 2 1130 0 3 68 0 1 28 0 0 1 0 6 20 5 0 0 0 1 6 0 0 1 0 1 0 0 0 10 1 1 2 0 0 0 4 1131 0 7 55 0 1 38 0 0 2 2 2 13 3 2 0 1 1 3 1 4 2 2 2 0 0 0 3 2 1 5 0 0 3 2 1133 0 5 55 0 4 36 0 0 2 1 1 11 7 1 0 0 0 5 3 1 1 0 1 0 1 0 5 2 2 3 1 0 3 2 1134 0 23 48 1 4 24 0 7 7 2 1 5 1 4 1 2 1 1 0 1 0 0 1 0 1 0 0 0 0 3 0 0 1 0 1135 0 3 62 1 1 33 0 0 1 0 0 17 8 5 0 1 0 2 1 2 0 1 4 0 4 0 5 1 0 4 1 1 1 1 1137 0 9 41 0 4 47 0 1 2 1 1 316 1 1 1 2 1 4 1 5 0 0 0 2 0 5 0 0 0 0 1 2 2 1138 1 19 41 1 2 38 0 2 10 2 1 5 5 6 0 4 0 1 0 1 2 1 2 0 12 0 2 0 0 3 0 0 0 1 279 Journal of the Royal Society of Western Australia, 80(4), December 1997 Discussion The identified microfauna is composed of 242 species, including 236 benthonic taxa (comprising 20 agglutinated, 74 porcellaneous, and 142 hyaline types). Six planktonic species are also present. In terms of abundance (Table 1), the porcellaneous species dominate and, with the hyaline Rotaliida, they form the bulk of the foraminiferal assemblages. Most of the species are very rare and sporadic within gulf sediments. Four agglutinated species, ten porcellaneous species, and 14 hyaline rotaliid species are present at frequencies >5% of the total foraminiferal count in the 150 pm - 2 mm sediment fraction (Table 1). Because of lack of similar taxonomic databases for other areas along the western coast of Western Australia, only limited biogeographic comparisons can be drawn at present. In comparison with the shallow-marine microfauna from the Sahul Shelf recorded by Loeblich & Tappan (1994), unusual occurrences in Exmouth Gulf which may have biogeographic significance include; (1) the presence of Borelis schlumbergeri not recorded from the Sahul Shelf, and the absence of related Alveolinella cjuoyi (d'Orbigny) which was recorded from the Sahul Shelf; (2) the absence of Ccilcarina which is represented by several species on the Sahul Shelf; (3) the absence of Baculogypsinoides present on the Sahul Shelf; and (4) the presence of Neorotalia calcar not recorded from the Sahul Shelf, and the absence of Pararotalia domantayi McCulloch (often misidentified as Calcarina calcar d'Orbigny) which is present on the Sahul Shelf. A major foraminiferal habitat that is rare in Exmouth Gulf is the seagrass meadow. McCook et al. (1995) recorded low abundances of seagrasses in the Gulf, and attributed this to the lack of suitable substrates. The lack of seagrass beds is reflected in the foraminiferal assemblages by the very low abundance of the porcellaneous "larger" Foraminifera, Amphisorus hemprichii, which is an abundant epiphytic species elsewhere on the Western Australian coast. The paucity of Amphisorus and absence of related Marginopora in the Holocene sediment, suggest that seagrass stands may have been rare in the Gulf throughout the Holocene. Acktwiuledgements: Samples for this study were made available by the Australian Institute of Marine Science, and were collected by the ship¬ board scientists on AIMS cruises 669 and 672 during September-October 1994, Sedimentological and bathymetric data were provided by A Orpin and K Woolfe of the Department of F.arth Sciences, James Cook University of North Queensland. S Revets provided taxonomic advice on some of the species. References Albani A D & Yassini I 1993 Taxonomy and distribution of the Family Elphidiidae (Foraminiferida) from shallow Australian waters. In: Centre for Marine Science, Technical Contribution 5 (ed P Dixon). University of New South Wales, Kensington, 1-49. Betjeman K J 1969 Recent Foraminifera from the western conti¬ nental shelf of Western Australia. Contributions from the Cushman Foundation for Foraminiferal Research 20:119-138. Brady H B 1884 Report on the Foraminifera dredged by HMS Challenger during the years 1873-1876. Reports of the Scientific Results of the Voyage of HMS Challenger 9 (Zoology): 1-814. Brown R G 1988 Holocene sediments and environments, Exmouth Gulf, Western Australia. In: The North West Shelf, Australia: Proceedings of Petroleum Exploration Society Australia Symposium, Perth, 1988 (eds P G & R R Purcell). Petroleum Exploration Society of Australia, Perth, 85-93. Burdett, I.D.J., Hedley, R.H., Hornibrook, N. de B., Hurdle, C.M., 1963. Gaudryina convexa (Karrer) 1865 - Upper Eocene to Recent; an example of variation and synonymy among Foraminifera. New Zealand Journal of Science 6:513-530. Ellis B F & Messina A 1940 et seq Catalogue of Foraminifera. American Museum of Natural Flistory, New York. Haig D W 1988 Miliolid Foraminifera from inner neritic sand and mud facies of the Papuan Lagoon, New Guinea. Journal of Foraminiferal Research 18:203-236. Haig D W 1993 Buliminid Foraminifera from inner neritic sand and mud facies of the Papuan Lagoon, New Guinea. Journal of Foraminiferal Research 23:162-179. Hatta A & Ujiie H 1992 Benthic Foraminifera from Coral Seas between Ishigaki and Iriomote Islands, Southern Ryukyu Island Arc, Northwestern Pacific. Part 2. Systematic descriptions of Rotaliina. Bulletin of the College of Science, University of the Ryukyus 53:163-287. Hayward B W Hollis C J & Grenfell FI R 1997. Recent Elphidiidae (Foraminiferida) of the South-West Pacific and fossil Elphidiidae of New Zealand. New Zealand Institute of Geological & Nuclear Sciences Monograph 16:1-170. Heron-Alien E & Earland A 1915 The Foraminifera of the Kerimba Archipelago (Portuguese East Africa). Part II. Trans¬ actions of the Zoological Society of London 20:543-794. Hettinger L Halicz E & Reiss Z 1991 Architecture of Eponides and Poroeponides reexamined. Micropaleontology 37:60-75. Hottinger L Halicz E & Reiss Z 1993 Recent Foraminiferida from the Gulf of Aqaba, Red Sea. Slovenska Akademija Znanosti in Umetnosti, Ljubljana. Jones R W 1994 The Challenger Foraminifera. Oxford University Press, Oxford, New York & Tokyo. Loeblich A R Jr & Tappan H 1987 Foraminiferal Genera and their Classification. Van Nostrand Reinhold Company, New York. Loeblich A R Jr & Tappan H 1994 Foraminifera of the Sahul Shelf and Timor Sea. Cushman Foundation for Foraminiferal Research, Special Publication 31:1-661. McCook L J, Klumpp D W & McKinnon A D 1995 Seagrass communities in Exmouth Gulf, Western Australia: A preliminary survey. Journal of the Royal Society of Western Australia 78:81-87. McCulloch I 1977 Qualitative Observations on Recent Foramin¬ iferal Tests with Emphasis on the Eastern Pacific. University of Southern California, Los Angeles. McKenzie K G 1962 A record of Foraminifera from Oyster Harbour, near Albany, Western Australia. Journal of the Royal Society of Western Australia 45:117-132. Murray J W 1991 Ecology and Palaeoecology of Benthic Fora¬ minifera. Longman Scientific & Technical, Harlow. Quilty P C. 1977 Foraminifera of Hardy Inlet, southwestern Aus¬ tralia. Journal of the Royal Society of Western Australia 59:79-90. Revets S A 1990 The revision of Bulirninoides Cushman, 1911. Journal of Foraminiferal Research 20:50-55. Revets S A 1991 The generic revision of the reusellids (Foraminiferida). Journal of Micropalaeontology 10:1-15. Revets S A 1993 The revision of l he genus Elotigobula Finlay, 1939. Journal of Foraminiferal Research 23:254-266. Revets S A 1996 The generic revision of five families of rotaliine Foraminifera. Part I, The Bolivinitidae; Part II, The Anomalinidae, Alabaminidae, Cancrisidae & Gavelinellidae. Special Publication, Cushman Foundation for Foraminiferal Research 34:1-113. Ujiie H Shell structure of Japanese smaller Foraminifera Part 2. Pararotalia nipponica (Asano, 1936). Transactions of the Palaeontological Society of Japan, New Service 61:191-200. 280 Journal of the Royal Society of Western Australia, 80:281-286, 1997 Abundance of arthropods in tree canopies of Banksia woodland on the Swan Coastal Plain R A Tassone1'2 & J D Majer2 ’Entomology Section, Agriculture Western Australia, South Perth WA 6151 2School of Environmental Biology, Curtin University of Technology, Kent Street, Bentley, WA 6845 ema il: ima jerj@info. curtin .edu. a u Manuscript received January 1997, accepted May 1997. Abstract Canopy arthropods and foliar nutrients were quantified during summer for Banksia menziesii, B. altenuata, B. ilicifolia and Nui/tsia floribunda within a low woodland site at Jandakot Airport on the Swan Coastal Plain, Western Australia. Foliar nutrients were higher for N. floribunda than any of the Banksia species. B. ilicifolia differed from the other two banksias in having lower nitrogen and phosphorus but higher potassium. Invertebrate densities were also higher for N. floribunda than any of the Banksia species, while densities on the three Banksia species were relatively similar, albeit with a few predominantly phytophagous taxa having a higher density on B. ilicifolia. These trends provide evidence that, as in other ecosystems where this has been studied, differing arthropod loads between tree species are related to variation in foliar nutrient levels. Tree usage by predominantly insectivorous birds may reflect the relatively low difference in arthropod loads between the three Banksia species and an uneconomic strategy of seeking out low-density N. floribunda trees. Introduction Invertebrates are an ubiquitous component of most terrestrial ecosystems and a key element in the energy flow and nutrient turnover within a community. Arboreal invertebrates may function as major herbivores, pollinators, parasites and predators (Majer & Recher 1988), and are themselves an important food resource for vertebrates. The dynamics of bird communities are closely linked to the abundance and diversity of invertebrates, with insectivorous avifauna responding to variations in the abundance of prey (Recher et al. 1991). Studies in open eucalypt forests of New South Wales by Recher & Majer (1994) showed that the abundance and biomass of invertebrates on eucalypts selected by birds were greater than on less preferred species. Similar invertebrate-driven preferences in tree usage by birds exist in wandoo ( Eucalyptus wandoo Blakely) woodland (Majer & Recher 1988) and jarrah (£. marginata Donn ex Smith) forest (Recher et al. 1996) in Western Australia. In south-western Australia there is a distinct seasonality of climatic patterns and there are also considerable differences in weather patterns between years. These variations, coupled with soil nutrients, moisture and the effects of fire, result in the local vegetation having variable levels of invertebrates. Recher et al. (1996) described these forests and woodlands as fluctuating environments, which possess significant temporal and spatial variation in invertebrates. Another variable influencing invertebrates in tree canopies is vegetation type. This study provides data for invertebrate densities on tree canopies of Banksia woodland at Jandakot Airport on the Swan Coastal Plain. It complements data on canopy arthropods in other © Royal Society of Western Australia 1997 Western Australian ecosystems, ranging from mallee (Recher et al. 1993) to wandoo woodland (Majer & Recher 1988, Majer et al. 1996) to jarrah forests (Majer et al. 1990, 1994). At present, natural bushland covers approximately 237 ha of Jandakot Airport and most of this is in a relatively undisturbed state (Milewski & Davidge 1981). In 1993, a draft environmental impact statement for additional development of aviation facilities, extraction of silica sand, and construction of a commercial estate (Anon 1993) proposed that 96 ha (40.5%). of the airport bushland be conserved. The significance of urban bushland remnants to vertebrate fauna has been noted by How & Dell (1994) who have conducted surveys in this region. In view of this, it is important to gather information on how the area caters for the feeding requirements of insectivorous vertebrates, such as birds. Methods Study site The study was undertaken at Jandakot Airport (32 °6 'S, 115 °53 'E), 18 km south of the centre of Perth and 9.5 km east of the coast, in the City of Cockburn. The vegetation of this region of the Swan Coastal Plain is broadly classified as 'Tow woodland" over "heath" and the airport has mainly mixed woodlands of the Proteaceae, Banksia menziesii R Br, B. altenuata R Br and B. ilicifolia R Br (Milewski & Davidge 1981). These banksias can be co-dominant, but vary widely in relative abundance. The study site is located within the Bassendean Dune system in a swale that retains high soil moisture throughout winter. This grey soil is a heavily leached sesquioxide podsol and is acidic (McArthur & Bettenay 1960). The mean annual rainfall at Jandakot 281 Journal of the Royal Society of Western Australia, 80(4), December 1997 Airport is 802 mm, with most rain falling between May and August. Sampling was carried out within a woodland area measuring 100 m by 130 m. The site was 15 m from the edge of the woodland, and had a 1.5 m sand bank between this edge and an access road. The area was adjacent to the study site used by the Western Australian Museum (R A How & J Dell, pers comm) for determining the significance of remnant urban bushland to reptile and other fauna. We sampled the three Banksia species and also Nuytsia floribunda (Labill) R Br ex Fenzl (Loranthaceae), which was scarce and contributed less than 2% to the tree cover. Tree canopy cover was 47%, and averaged 6 - 7 m in height. The dense understorey was dominated by the 2 m tall shrub, Regelia ciliata Schauer (Myrtaceae), which had a patchy distribution throughout the study site, and also by grass trees, Xanthorrhoea preissii Endl (Xanthorrhoeaceae). The ground cover consisted of undershrubs and perennial herbs. An additional tree species. Eucalyptus todtiana F Muell (Myrtaceae), was also present, but was so sparsely distributed that it was not included in this study. All three Banksia species can be trees up to 10 m in height. The leaves of B. menziesii are dull green and very rigid, 10 - 40 mm wide and 80 - 250 mm long (Marchant et al 1987). The conspicuous flower heads (held above the foliage) are usually pink or red, occurring mainly in February to August. The leaves of B. attenuata are more crowded than B. menziesii and much narrower, being 5 - 15 mm wide, 40 - 270 mm long, and strongly serrated. The conspicuous flower heads are bright yellow, appearing from September to December. The common name holly-leaved banksia aptly describes B. ilicifolia , with its dark green leaves; they are 20 - 40 mm wide and 30 - 90 mm long, elliptical with an undulating edge, and the leaves are slightly serrated with each tooth possessing a spine. The flower heads are cream, turning deep pink with age. The flowers appear throughout the year, but peak in spring. N. floribunda is a tree up to 10 m high, that parasitises the roots of surrounding plants. The glabrous leaves are long (40 - 100 mm), narrow (3-8 mm) and thick. This tree has masses of orange flowers, which occur from October to January. To touch, the leaves of N. floribunda are much less sclerophyllous than the leathery Banksia leaves. When bent, B. menziesii and B. attenuata leaves crack and split, unlike B. ilicifolia whose leaves are more supple. Sampling The proportional composition of tree species within the canopy was assessed above 400 vertical sighting points taken at regular distances along transects throughout the study site. Arthropods were sampled from the canopy of the three Banksia and N. floribunda during early summer (December 1994). We sampled trees which were not in flower, since we were primarily interested in foliage- associated arthropods. The number of trees sampled was dictated by their availability in the study site. Consequently, 20 each of B. menziesii and B. attenuata, 10 of B. ilicifolia and four of N. floribunda were sampled. The height and crown diameter was recorded for each tree. Mature foliage was collected from 10 individuals of each species for subsequent analysis of nutrients. Leaf material was dried at 60 °C for 48 hours and ground to a fine powder prior to analysis. Cotton, funnel-shaped nets with a sampling area of 0.5 m2 were used to collect the chemical knockdown samples. Each net was fitted with a sleeve that held a 100 ml plastic tube. Within a given tree, five nets were suspended at different heights below the canopy foliage. Net positions were selected to equalize the amount of foliage (determined by visual inspection) in the column directly above the nets. Nets were positioned in the morning (0800 h), one hour prior to spraying, to allow disturbed invertebrates to return. The canopy above each net was sprayed with synthetic pyrethrin pesticide, synergised with piperonyl butoxide, using a motorized-knapsack mist-blower. Spraying was done only during dry and calm conditions. Two litres of diluted (0.2%) pesticide was used per tree, and trees were left for at least 30 minutes to allow silk- attached invertebrates to drop into nets. The canopy was then shaken to dislodge remaining invertebrates and specimens were brushed into the collecting tubes and preserved in 70% ethanol prior to sorting and counting to ordinal level. The numbers of invertebrates caught could be influ¬ enced by the amount of foliage above each net. To obtain a measure of this, we cut three 1 m long branches (measured from growing tip backwards) from each tree species and counted the number of leaves on each branch. Twenty leaves of each species were then removed and their surface area measured using a computer-based planimeter. This enabled a measure of the area of foliage on a 1 m branch to be calculated. Data analysis Following sorting, the number of animals within each taxon was summed for the five nets placed within each tree. Mean and standard error of the numbers of each invertebrate taxon on each tree species were then calcu¬ lated. Three comparisons were made; (1) between the three Banksia species for each invertebrate taxon; (2) between the three Banksia species for all taxa that were significantly different; and (3) between the four sampled tree species for each invertebrate taxon. Analyses were restricted to the common invertebrate taxa, those occur¬ ring in >65% of the samples. N. floribunda was excluded from comparisons (1) and (2) because of the low number of trees sampled. Comparison (1) was performed by univariate analysis of variance (ANOVA) and those means which differed significantly were detected using Tukey's pairwise comparisons (P < 0.05). Comparison of Banksia species for each invertebrate taxon required natural log transformations of the data due to the high degree of variability or the presence of zeros. Comparisons (2) and (3) involved ranking the invertebrate taxa and comparing these ranks using Kendall's coefficient of concordance. Foliar samples from each tree were analysed for total nitrogen, phosphorus and potassium (see Majer et al. 1992 for methodology). The variation in nutrient levels between the four sampled tree species was analysed by univariate analysis of variance (ANOVA) and those means which significantly differed were detected using Tukey's pairwise comparisons (P < 0.05). Standard 282 Journal of the Royal Society of Western Australia, 80(4), December 1997 deviations were relatively homogeneous, indicating that transformations were unnecessary. Results In terms of cover, B. attenuate i was the most dominant species (70% of projected foliage cover), followed by B. menziesii (20% of projected foliage cover) and B. ilicifolia with 7% of projected foliage cover (Table 1). Table 1 also indicates that B. ilicifolia was the smallest tree species sampled, having up to a 2 m crown diameter. The single main stem of this species had many branchlets that formed a cover of dense foliage around its stem. The large crown diameter of N. floribunda was in part associated with the tendency of this species to have more than one main stem. B. attenuata and B. menziesii had similar canopy structure, with the diameter (3.0 - 3.5 m) and height above the ground (6.5 - 7.0 m) of their crowns being similar. The area formed by the crown of N. floribunda (17.3 m2) was twice as large as B. menziesii (7.3 nr) and B. attenuata (9.5 m2), and approximately six times larger than B . ilicifolia (3.2 m2). N. floribunda supported by far the most leaves per metre of branch, followed by B. attenuata ; B. menzesii branches had the lowest number of leaves (Table 1). When leaf area was taken into account, N . floribunda L ranches supported the least foliage, B. attenuata the most, while the other two banksias were intermediate. There were generally significantly higher levels of ni¬ trogen, phosphorus and potassium within the leaves of N. floribunda than in the Banksia species (Table 2). Nitro¬ gen and phosphorus are generally mobile nutrients within the plant, and their within-plant concentrations are usually correlated (Majer et al. 1992). This is a probable explanation for the assayed leaves showing identical pat¬ terns of significance for nitrogen and phosphorus. B. ilicifolia had significantly lower levels of nitrogen and phos¬ phorus but higher levels of potassium than the two other Banksia species. B. menziesii and B. attenuata exhibited no significant difference between their levels of nitrogen, phosphorus or potassium. A total of 7 105 individual arthropods was obtained from the 54 tree canopies sampled. The invertebrate count was higher on N. floribunda than on any of the Banksia species (collected arthropods averaged 216 from N. floribunda , 182 from B. ilicifolia , and only 107 - 114 from B. menziesii and B. attenuata respectively). The F ratios for univariate ANOVA comparisons of the three Banksia species, together with means and standard errors for the data, are presented in Table 3. It is evident that the densities of only a few invertebrate taxa differ significantly between sampled Banksia species within woodland at Jandakot. These invertebrates are Acarina (mites), Blattodea (cockroaches), Homoptera (sucking bugs), Thysanoptera (thrips) and Neuroptera (lacewings) larvae. In almost all significant differences, numbers were significantly higher on B. ilicifolia than B. menziesii and B. attenuata. The exception was for Blattodea, where B. ilicifolia and B. attenuata had similar but higher densities than B. menziesii. Levels of the various invertebrate taxa generally did not significantly differ between B. attenuata or B. menziesii , although Acarina were significantly more abundant on B. menziesii nd, as mentioned, Blattodea were significantly more abundant on B. attenuata. Despite these abundance differences, almost the same variety of taxa were found on the four tree species. Over¬ all, arthropods from 20 orders of insects (Heteroptera and Table 1 Crown diameter and height (n = 10), the proportional composition of the overstorey, mean leaf area (n = 20), and the number and area of leaves along 1 m branches (n = 3) of sampled tree species in the Banksia woodland study site at Jandakot Airport during December 1994. Values are mean (± se). Species Crown diameter m Tree height m Percentage canopy (%) Leaf area cm2 Number of leaves on 1 m branch Leaf area on 1 m branch (cm2) Banksia attenuata 3.5 ± 0.2 7.0 ± 0.2 69.9 9.4 ± 0.4 979.7 ± 7.7 9256.9 Banksia menziesii 3.0 ± 0.2 6.7 ± 0.2 19.9 25.6 ± 0.6 274.3 ± 5.3 7021.9 Banksia ilicifolia 2.0 ± 0.3 6.4 ± 0.4 6.8 14.5 ± 0.4 490.7 ± 6.7 7124.0 Nuytsia floribunda 4.5 ± 0.4 6.7 ± 0.4 1.7 2.1 ± 0.2 2023.7 ± 15.8 4236.3 Others - - 1.7 - - - Table 2 Nutrient levels (pg g'1) within tree foliage (n = 10) from Banksia woodland at Jandakot Airport during December 1994. The means (± se) of are shown with level of significance (P) between tree species; different superscript letters indicate that means differ significantly using univariate analysis of variance (ANOVA) Total nutrient Species Nutrient comparison B. menziesii B. attenuata B. ilicifolia N. floribunda F Value P Nitrogen 5353.0 ± 6.6b 5540.0 ± 6.4b 3914.0 ± 8.3C 6746.0 ± 9.3a 34.5 < 0.05 Phosphorus 209.5 ± 1.7b 200.5 ± 1.7b 157.5 ± 1.8C 434.5 ± 2.0a 141.2 < 0.05 Potassium 1680.0 ± 6.2b 2065.5 ± 4.0b 2602.5 ± 7.2a 2471.0 ± 7.5a 9.2 < 0.05 283 Journal of the Royal Society of Western Australia, 80(4), December 1997 Table 3 Numbers of invertebrates per tree, sampled by pyrethrin knockdown of Banksia woodland canopy at Jandakot Airport during Decem¬ ber 1994. The number of invertebrates per tree was based on five 0.5 m2 nets within the canopy. Best ranks are assigned to all taxa and also the five taxa where significant differences occurred. Means (± se) for the three Banksia species were compared where possible using a univariate analysis of variance (ANOVA) and those means which differ significantly are indicated by different superscript letters. Taxa with significant differences between banksia species are indicated in bold. Ranks compared using Kendall's Coefficient of concordance. NS = not significant. Class Taxon B. menziesii Species B. attenuata B. ilicifolia N. floribunda Tree comparison n = 20 n = 20 n = 10 n = 4 F Value P Arachnida Pseudoscorpionida 0.2 ± 0.2 0.3 ± 0.2 0.1 ± 0.2 4.8 ± 0.9 - Acarina 23.7 ± l.lb 11.5 ± 0.7C 74.8 ± 2.5a 22.8 ± 2.2 15.4 < 0.05 Araneae 8.0 ± 0.5 12.1 ± 0.6 10.0 ± 0.8 22.0 ± 1.8 NS Crustacea Isopoda 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 - Collembola 0.5 ± 0.2 0.2 ± 0.1 1.1 ± 0.5 0.3 ± 0.4 - Insecta Thysanura 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 3.0 ± 0.6 - Blattodea 0.2 ± 0.1b 1.2 ± 0.3a 1.1 ± 0.3a 1.8 ± 0.6 5.1 < 0.05 Plecoptera 0.1 ± 0.1 0.3 ± 0.2 0.2 ± 0.2 1.0 ± 0.6 - Mantodea 0.2 ± 0.1 0.2 ± 0.1 0.1 ± 0.2 0.5 ± 0.4 - Isoptera 0.5 ± 0.2 0.4 ± 0.2 0.8 ± 0.4 2.5 ± 0.9 - Orthoptera 0.1 ± 0.1 0.0 ± 0.0 0.1 ± 0.2 0.3 ± 0.4 - Pscoptera 3.0 ± 0.4 2.2 ± 0.3 5.6 ± 0.7 54.0 ± 2.3 NS Hemiptera Homoptera 2.7 ± 0.4b 4.4 ± 0.4b 9.1 ± 1.1* 12.5 ± 1.6 4.2 < 0.05 Heteroptera 5.7 ± 0.6 4.7 ± 0.5 8.6 ± 0.8 3.3 ± 1.0 NS Thysanoptera 9.2 ± 0.5b 12.8 ± 0.7b 24.6 ± 1.4* 8.5 ± 1.7 3.3 < 0.05 Neuroptera adults 0.1 ± 0.1 0.5 ± 0.3 0.3 ± 0.3 0.5 ± 0.5 - larvae 1.0 ± 0.3b 1.0 ± 0.3b 2.6 ± 0.4a 7.3 ± 1.6 6.7 < 0.05 Coleoptera adults 12.1 ± 0.6 18.3 ± 0.7 14.5 ± 1.2 37.0 ± 3.6 NS larvae 12.3 ± 0.7 11.3 ± 0.6 6.8 ± 0.6 3.0 ± 1.1 NS Mecoptera 0.5 ± 0.3 0.6 ± 0.3 0.0 ± 0.0 0.0 ± 0.0 - Diptera adults 1.8 ± 0.3 1.8 ± 0.3 3.2 ± 0.6 6.5 ± 1.0 NS larvae 0.0 ± 0.0 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 - Lepidoptera adults 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.8 ± 0.5 - larvae 1.1 ± 0.3 1.0 ± 0.3 0.2 ± 0.2 0.0 ± 0.0 NS Hymenoptera ants 16.6 ± 0.9 17.9 ± 1.1 7.5 ± 1.2 7.8 ± 1.4 NS others 7.5 ± 0.4 11.9 ± 0.8 10.7 ± 0.8 16.5 ± 1.3 NS Total inverterbrates 106.7 114.4 182.0 216.3 Best rank (All taxa) 4 2 3 1 Arachnida Acarina 2 3 1 _ Insecta Blattodea 3 1.5 1.5 - Hemiptera Homoptera 2.5 2.5 1 - Thysanoptera 2.5 2.5 1 - Neuroptera larvae 2.5 2.5 1 - Best rank (Significant taxa) 3 2 1 - Homoptera counted as one order), arachnids, crustaceans and collembolans were collected, with 18 sampled on N. floribunda, 17 on B. attenuate i and B. ilicifolia, and 20 on B. menziesii. The most abundant invertebrate order on the four tree species was Acarina, with over 1540 individuals recovered (22% of total). Along with Acarina, Coleoptera (beetles) and Hymenoptera (ants, wasps, sawflies and bees) dominated the numbers of invertebrates collected from Jandakot. Of the 20 orders sampled, these three composed 61% of the total number of invertebrates. The numbers of invertebrates in each taxon were ranked across tree species for all taxa and also for those taxa which exhibited significant differences (Table 3). In the first of these rankings, N. floribunda ranked as having the most abundant invertebrate fauna, followed by B. attenuata, B. ilicifolia and finally B. menziesii. This ranking was not significant (Kendall's coefficient of concordance; P > 0.05), although it was in partial agreement with the ordering of total invertebrate counts in that N. floribunda was ranked the highest and B. menziesii the lowest; the 284 Journal of the Royal Society of Western Australia, 80(4), December 1997 other two Banksia sp. were ranked as intermediate. When only those invertebrates with significant differences on Banksia species were ranked (on the basis of their ordering in the ANOVA analysis), B. ilicifolia had the most individuals, while B. attenuata was only marginally ranked higher than B. menziesii. Discussion The higher abundance of invertebrates on the canopy of N. floribunda than on that of the Banksia species supports the conclusions from other studies that invertebrate densities are higher in canopies of tree species with higher foliar nutrient levels (Majer et al 1992), although it is not possible to test this statistically due to the low sample size. The possibility that higher invertebrate densities can be associated with greater amounts of foliage can here be eliminated as N. floribunda was the species with the lowest area of leaves above the sampling nets. As with N. floribunda, the densities of invertebrates on the three Banksia species were unrelated to foliage area above the nets. The densities of invertebrates in canopies of the three Banksia species were all relatively similar. B. menziesii and B. attenuata had almost identical foliar nutrient levels, so it is consistent that the density of invertebrates on these two species was similar. The position of B. ilicifolia in the invertebrate rankings was equivocal, since its foliar nutrient levels were lower in nitrogen and phosphorus, but higher in potassium than those of the other two banksias. Usually it is the levels of nitrogen which correlate with invertebrate levels, particularly herbivores, presumably because this reflects the protein content of the diet (see references in White 1969; Morrow 1983). Phosphorus and potassium levels may also reflect some aspects of the nutritional quality of the foliage, so possibly the effects of higher potassium cancels out the effects of lower levels of the other two nutrients. Certain taxa do exhibit significant differences between the Banksia species and four of these taxa (Acarina, Blattodea, Homoptera and Thysanoptera) contain species which derive their nutrients directly from live or decaying leaves. The fact that they reach greatest abundance on B. ilicifolia suggests that some taxa may be responding to the higher levels of potassium in leaves of this species. These trends therefore provide evidence supporting the fact that, as in other ecosystems where this has been studied, differing arthropod abundance's (Recher et al. 1996) and richness levels (Majer et al. 1994) between tree species may be related to foliar nutrient levels. An alternative possibility is that the higher foliage density on B. ilicifolia may create a more favourable microclimate for arthropods. The other studies on canopy arthropods elsewhere in Australia (see Introduction) used identical procedures, with the exception that 10 nets were used to sample the overstorey trees. How do the levels of invertebrates compare with other trees in Western Australia? If values are expressed as mean number of invertebrates per net, then the Banksia species support similar levels of arthropods to E. marginata and E. calophylla Lindley (marri) trees sampled during the same season in the nearby jarrah forest (Recher et al. 1996). Levels of invertebrates on N. floribunda were considerably higher than those normally present on local Eucalyptus species. Arthropod levels on all tree species in the present study were higher than on trees in the more arid mallee (Recher et al. 1993) and wheatbelt woodland sites (Majer & Recher 1988, Majer et al. 1996), suggesting the possible existence of a gradient of canopy arthropod densities with increasing rainfall. Tree usage by insectivorous birds was recorded in the same site during the period when invertebrates were sampled through until September 1995 (P Kirkpatrick, School of Environment Biology, Curtin University, unpublished data). When the proportion of feeding observations was assigned to the various tree species, it was found that values reflected the proportional composition of the overstorey for each tree species. Thus, insectivorous birds at Jandakot do not appear to be selecting particular tree species as foraging substrates, but rather use the various species of tree in a random manner. The absence of any apparent selection of tree species by insectivorous birds is consistent with the Banksia species supporting relatively similar invertebrate loads. The absence of any selection for N. floribunda, despite its higher nutrient loads and invertebrate numbers, is surprising. It is possible that its low density within this ecosystem means that birds do not find it economical to seek out this tree species. Another possibility is that birds may find it difficult to probe for invertebrates amongst the dense foliage of this species. These findings suggest that, on a unit area basis, Banksia woodland provides similar amounts of food for canopy-feeding insectivorous birds as does the jarrah forest in the Darling Ranges. Indications are that, in contrast to the jarrah forest, insectivorous bird richness is low at Jandakot Airport and that some species are no longer present (How & Dell 1993). Probably the fragmentation of the woodlands of the Swan Coastal Plain, and the reduction in area of this habitat, has reduced the available feeding areas for tree-dependent birds and this has resulted in the local extinction of some species. The presence of this array of invertebrates in the canopies of Banksia species and N. floribunda presents a compelling case for maintaining this sizeable remnant of native woodland at Jandakot Airport. Acknowledgments : We would like to thank K Flecknell and the Federal Airports Corporation for permission to use the Jandakot Airport site and also P Kirkpatrick, R P Tassone and E Kostas who assisted with field work. N Keals and M J de Sousa Majer helped with sorting of arthropod material and K McGregor, J O'Neil and M Costigan of the Department of Ecosystem Management, University of New England, performed the nutrient analyses. P Groom assisted with the leaf area measurements and R How, B Lamont, J Scott and an anonymous referee kindly commented on an earlier draft of this paper. References Anon 1993 Proposed additional developments at Jandakot Air¬ port. Environmental impact statement. Dames & Moore, Perth. How R A & Dell J 1993 Vertebrate fauna of the Perth Metropolitan Region: Consequences of a changed environment. In: Urban Bushland Management (ed M Hipkins). Australian Institute of Urban Studies, Perth, 24-47. 285 Journal of the Royal Society of Western Australia, 80(4), December 1997 How R A & Dell J 1994 The zoogeographic significance of urban bushland remnants to reptiles in the Perth region, Western Australia. Pacific Conservation Biology 1:132-140. Majer J D & Recher H F 1988 Invertebrate communities on West¬ ern Australian eucalypts: A comparison of branch clipping and chemical knockdown procedures. Australian Journal of Ecology 13:269-278. Majer J D, Recher H F & Ganeshanandam S 1992 Variation in foliar nutrients in Eucalyptus trees in eastern and Western Australia. Australian Journal of Ecology 17:383-393. Majer J D, Recher H F & Reals N 1996 Branchlet shaking: A method for sampling tree canopy arthropods under windy conditions. Australian Journal of Ecology 21:229-234. Majer J D, Recher H F, Perriman W S & Achuthan N 1990 Spatial variation of invertebrate abundance within the canopies of two Australian eucalypt forests. Studies in Avian Biology 13:65-72. Majer J D, Recher H F & Postle A C 1994 Comparison of arthropod species richness in eastern and western Australian canopies: A contribution to the species number debate. Memoirs of the Queensland Museum 36:121-131. Marchant N G, Wheeler J R, Rye B L, Bennett E M, Lander N S & MacFarlane T D 1987 Flora of the Perth region. Depart¬ ment of Agriculture, Perth. McArthur W M & Bettenay E 1960 The development and distri¬ bution of the soils of the Swan Coastal Plain. CSIRO (Austra¬ lia) Soil Publication 16. CSIRO, Canberra. Milewski A V & Davidge C 1981 The physical environment, floristics and phenology of a Banksia woodland near Perth, Western Australia. Western Australian Herbarium Research Notes 5:29-48. Morrow P A 1983 The role of sclerophyllous leaves in determin¬ ing insect grazing damage. In: Mediterranean-Type Ecosystems (eds F J Kruger, D T Michael & J U Jarvis). Sp ringer- Verlag, Berlin, 509-524. Recher H F & Majer J D 1994 On the selection of tree species by Acanthizidae in open-forest near Sydney, New South Wales. Emu 94:1-7. Recher H F, Majer J D & Ford H A 1991 Temporal and spatial variation in the abundance of eucalypt canopy invertebrates: The response of forest birds. Acta XX Congressus Intemationalis Omithologicii 20:1568-1575. Recher H F, Majer J D & Ganesh S 1996 Eucalypts, arthropods and birds: On the relation between foliar nutrients and species richness. Forest Ecology and Management 85: 177-196. Recher H F, Majer J D, Gowing G & Sarti N L 1993 Canopy invertebrate communities in woodlands - A comparison of morning and afternoon samples by chemical knockdown. Mulga Research Centre Journal 11:27-30. White T C R 1969 Food and outbreaks of phytophagous insects, with special reference to Cardiaspina densitexta Taylor (Psyllidae, Homoptera) on Eucalyptus fasciculosa F.v.M. (Myrtaceae) in South Australia. PhD Thesis. University of Adelaide, Adelaide. 286 Journal of the Royal Society of Western Australia, 80:287, 1997 The Royal Society of Western Australia Medallists, 1997 Dr Ernest Hodgkin, OAM Dr Arthur McComb Ernest Hodgkin was born in Madagascar in 1908, and subse¬ quently grew up in the United King¬ dom. He graduated in 1930 with a BSc (Hons) in Zoology from the Victoria University of Manchester. Following his graduation, he was em¬ ployed as an entomologist for the Institute of Medical Research for the Federated Malay States, where he worked on arthropod vectors of malaria, filariasis and rickettsial disease until 1942, when he was interned in Singapore as a civilian Prisoner of War. After his release from internment in 1945, he was re¬ united with his family, who had moved to Perth. Ernest was appointed a lecturer in the Department of Zoology at the University of Western Australia in 1946, where he remained until 1973 having been promoted to Associate Professor. In 1950, he gained a DSc from UWA on "The transmission of malaria in Malaya", and has published 14 papers on this subject. From the mid-1950s he carried out research on the ecology of invertebrate fauna of freshwater, estuaries and rocky shores, and coastal geo¬ morphology. He supervised the research and training of a series of research students and published 28 papers in this area up to 1974. Many of his students have gone on to positions in Universities, Government Departments or private consultancies. From 1955 to 1958 Ernest was also a Trustee of the Western Australian Museum, being Vice-Chairman from 1960-1981 and Chairman from 1982-1983. On retiring from his University position in 1974 at the age of 65, Ernest has continued a remarkable contribu¬ tion to science in Western Australia. From 1974 to 1996 he was Environmental Consultant at the Department of Conservation and Environment, and the Environmental Protection Agency, when he co-ordinated team research studies of the Blackwater River and on eutrophication in the Peel-Ha rvey estuarine system. He published a further 25 papers on the Blackwood Estuary, the Peel-Harvey and other Estuaries, including the "Estuarine Studies Series" published by the EPA, This series sets out geo- morphological and biological characteristics of the 21 estuaries from Augusta to Esperance. From 1989-1996 he was environmetal advisor to the Department of Environ¬ mental Protection, when he provided advice on the manage¬ ment of Culham Inlet and other estuarine systems, in 1996 he co-authored the book "Estuaries of the South Coast". In recognition of his outstanding work, Ernest was awarded the Order of Australia Medal (OAM) in 1983, and the Royal Society of Western Australia is pleased to acknowledge his remarkable contributions to various aspects of knowledge of estuaries in the south-west of the State both before and since that time, by the award off the Royal Society of Western Australia Medal for 1997. Arthur McComb was born in Melbourne in 1936. He attended the University of Melbourne where he graduated in 1956 with a joint major in Botany and Zoology, and an Exhi¬ bition and First Class Honours in Botany. This was followed by an MSc from the University of melbourne in 1959. With the help of a Royal Com¬ mission for the Exhibition of 1851 Overseas Scholarship, he undertook his PhD research at the University of Cambridge , which was completed in 1962. Arthur returned to Australia with a lectureship at the University of Western Australia (1963-1967), and was promoted to Senior lecturer in 1967. He spent two years carrying out research abroad, first in 1969 as a Research Associate at the Atomic Energy Commission, Michigan State University, and then in 1976 as an Honorary Research Fellow at the University of Leicester. Promo¬ tions followed, first to Associate Professor in 1977 and Head of the Botany Department ay at UWA (1981-1986), then to Professor of Environmental Sciences at UWA (1981-1986), and Professor of Environmental Sciences at Murdoch University (1989-present). From 1982 (until present), Arthur has been Joint Director of the Centre for Water Research. His work has been in two major areas, the control of plant growth by internal factors, and the control of plant growth by the environment. The first area involved plant growth regulators (gibberelins). As a logical extension to this work, Arthur's interests began to move towards ecological aspects of plant physiology, particularly in relation to annual Western Australian plants, but also studies on resins produced by native species, that have similar structures to gibberelins. The major part of Arthur's work has been on control of growth in aquatic plants, particularly understanding the fundamental processes that control plant biomass in aquatic systems. This has been of considerable management significance, especially in relation to the accession of nutrients from catchments and their effects in receiving waters. Publications include more than 120 scientific papers has written or edited six books. Arthur has supervised some 30 PhD students. His former students are spread throughout Australian science, in Universities, State Government Departments and in private consulting firms. Service to the scientific community includes President of the Royal Society of Western Australia (1978), on the Scientific Advisory panel of the World Wildlife Fund (1984-1986), Deputy Chair then Chair of the WA National Parks and Nature Conservancy Authority (1985-1995), Chair of the WA Lands and Forests Commission (1988- 1998), and President of the Ecological Society of Australia (1987-1988). Arthur McComb has had a seminal influence on a generation of researchers. The Royal Society of Western Australia is pleased to honour his contribution to Western Australian and world science by the award of the Royal Society of Western Australia Medal for 1997. [MGK Jones, President, RSWA, July 1997] 287 Journal of the Royal Society of Western Australia, 80:289-290, 1997 OBITUARIES William Harold Cleverly William Harold Cleverly, known to everyone as 'Bill', died on 19 April 1997, shortly after his 80th birthday. Born in Guildford, Western Australia, on 25 January 1917, he joined the Royal Society of Western Australia in 1938 and was later to become a regular contributor to the Society's Journal. In recognition of his contribution to science in Western Australia, Bill was made an Honorary Member of the Society in 1982. Bill graduated from the University of Western Australia in 1939 with a Bachelor of Science degree in Geology. An arduous correspondence course in mathematics with the University of Western Australia eventually saw him awarded a BA degree in 1951. Known throughout the world for his work on Australian meteorites and tektites. Bill's early work was more geological. In 1939, he joined the North Australia Survey. During the early years of World War II, he carried out geological surveys of Queensland, mainly on the Atherton Tableland. In 1941 he returned to Western Australia to take up a short-lived appointment as Science Assistant at the Western Australian School of Mines (WASM, now part of Curtin University of Technology) in Kalgoorlie, before war service as an ammunition examiner with the AIF between 1942-46 on Borneo and other islands. Following the war. Bill returned to Kalgoorlie to resume his post at the School of Mines and to settle down with his wife Evelyn. In 1947 he was appointed Lecturer and Head of the Geology Department at the WASM, the position he held until his retirement in 1977. It was during this period that his interest in meteorites and tektites increased. The 1960's saw the recognition of the arid zone of Western Australia, particularly the Nullarbor Region, as a potential source of meteorite finds. Together with his great friend, the late Keith Quartermaine, Bill made many forays into the Nullarbor to search for meteorites and tektites, and investigated many reports of meteorite falls and finds throughout the Goldfields. He also sparked an interest in meteorites in the late John Carlisle who subsequently collected many meteorites from the Nullarbor. Consequently, during the period 1960-1974 the number of meteorites known from Australia increased by 57, with the largest contribution coming from the Western Australian Nullarbor. During the 1960's alone, nearly a tonne of meteoritic material passed through the WASM into its collection, or via the School into other collections. This material included 37 new meteorites representing an addition of about 2% to all meteorites known in the world at that time. Together with colleagues such as Brian Mason, Joe McCall and John de Laeter, and aided by Evelyn, Bill documented and described this wealth of new meteoritic and tektite material. In addition to his work at the WASM, Bill was one of the longest serving members (1966-1987) of the Meteorite Advisory Committee of the Western Australian Museum. As Curator of the Museum attached to the WASM from 1972-1977, Bill also had the job of caring for the collection, including meteorites and tektites, that had been so enriched by his work. During twenty years of retirement, as an Honorary Research Fellow (1977-97) at the WASM, and as an Honorary Associate of both the Western Australian (1966-97) and South Australian (1977-97) Museums, Bill's scientific output never waned, and he published many papers on meteorites and tektites. Overall, he published more than 40 scientific papers and popular articles, including twelve as first author and one as a co-author in the Journal of the Royal Society. His meticulous and systematic work on the shapes and size distribution of tektites in Australia remains as a valuable contribution to science. Last year, while on one of his regular tektite hunting expeditions with Evelyn, Bill suffered a stroke and died in hospital shortly after. On a personal note, all who knew him will remember his warm, friendly nature, wonderful turn of phrase, and great sense of humour. His tall, gentlemanly figure will be greatly missed in the Goldfields. [Alex Bevan , Department of Earth and Planetary Sciences, Western Australian Museum] Brian John Grieve Professor Brian Grieve, long-time member of the Royal Society of Western Australia, passed away on 5 September 1997, aged just over 90. Although a native Victorian, Brian Grieve became so much part of the botanical scene in Western Australia that he was readily accepted as a Sandgroper. Born in Allans Flat on 15 August 1907, he was educated at Williamstown High School and the University of Melbourne. In 1929 he gained First Class Honours in Botany, and the following year an MSc. Receiving an 1851 Exhibition, he studied for a PhD at the University of London in 1930-31, then returned to take up a lectureship in Botany at the University of Melbourne. He again spent two years in Britain in 1938-39, studying mycology at Cambridge. Early in the Second World War he served in the Royal Australian Naval Reserve but then returned to the University to investigate fungal contamination of field glasses in New Guinea. Brian moved to Western Australia at the start of 1947 to become head of the Botany Department at the University of Western Australia. At the time there was one other lecturer and two graduate assistants. He set about building up both the Department and the courses offered. By 1957 the Department had grown to the extent that it was given a Chair, and Brian became the Foundation Professor. As staff and student numbers increased, the old wooden buildings (originally part of 'Tin Pot Alley' in Irwin Street in the City where the University was housed until 1930) became quite inadequate, and Brian planned the move to a new building which opened in 1969. The greatest strength of his contribution was the breadth of botanical subjects offered — general botany, anatomy, physiology, genetics, biosystematics, ecology, mycology and other cryptogams, systematics. Hundreds of students who passed through the Department have gone on to make significant contributions in their fields. 289 Journal of the Royal Society of Western Australia, 80(4), December 1997 His own research turned towards the physiology of native plants, especially water relationships, laying a firm foundation for others to build on. He joined the Royal Society in 1948, and at the end of his first term as President in 1953, his address was on 'The physiology of sclerophyll plants' (J Royal Soc Western Australia 39:31-45, 1955). He was President again in 1970-71, closing this term with 'Botany in Western Australia: a survey of progress 1900-1971 (J Royal Soc Western Australia 58:33- 53, 1975). In 1975 he was made an Honorary Life Member, and in 1979 was awarded the Society's Medal. In 1983 he delivered the Nancy Burbidge Memorial Lecture to the Australian Systematic Botany Society. He served on many committees both within and outside the University, including the Kings Park Board from 1959 to 1978. Among the public and systematists, Brian is best known for his painstaking work on the project conceived by William Blackall — the illustrated keys published as 'How to Know Western Australian Wildflowers'. Blackall prepared considerable manuscript material before his death in 1941. His family asked the University to complete the work, and the task fell to Brian Grieve. Part 1 was published with little amendment in 1954 (the first work published by the University's Press), but all later parts required increasing input, both to incorporate existing taxa and to interpolate new discoveries. Brian worked towards this goal for almost 50 years, always in addition to his other duties, and often during vacation. Fittingly, his role was recognised in the authorship of later parts as 'Grieve and Blackall' rather than 'Blackall and Grieve'. The latest revision is currently in press. Brian never published a formal taxonomic paper, yet his contribution to systematics has been immeasurable, providing a means for identifying wildflowers and hence encouraging the interest of students and users alike. Never one to push himself forward, Brian was respected with esteem and affection by colleagues and students for his gentleness, wisdom and compassion. [A George, RSWA Council] Clee F H Jenkins Mr Clee Francis Howard Jenkins, naturalist, broad¬ caster, entomologist and supporter of the Royal Society of Western Australia, died on July 13th aged 89. He was born in Adelaide in 1908 where he received his education at St Peter's College, and joined the Western Australian Museum as a cadet in 1929. He was appointed as an entomologist in the Western Australian Department of Agriculture in 1933, and as Government Entomologist in 1939 then Chief of the Division of Biological Services (in¬ corporating Entomology, Botany, Plant Pathology and Weeds and Seeds) of the Department of Agriculture. He graduated with a BA from the School of Science, Univer¬ sity of Western Australia, in 1935 then a MSc in 1939. His obituary entitled 'Death of a Gentleman', written by Keith McDonald and published in the West Australian (21st July 1997), well sums up Clee Jenkins and his life. He joined the Royal Society of Western Australia in 1929 and played a prominent role in all aspects of the Society for many years. From the humble beginning as Assistant Librarian, he served in Executive or on Council for more than 30 years; his various roles included Honorary Sec¬ retary, Vice-president (4 years). President of the Society (three times over a span of 34 years, in 1945, 1963 and 1979) and member of Council. He was the tenth recipient to be awarded the Kelvin Medal of the Royal Society of Western Australia in 1966 for exceptional service to the Society, for advancement of science in all of its branches, and for his personal achievment in his own field. He worked for the Department of Agriculture, mainly in the field of insect control, and his sound recommendations were responsible for eradication of many important insect pests; this work greatly benefited Western Australia. Clee Jenkins also made a distinguished contribution to Sci¬ ence by his enthusiastic encouragement and publicity of natural history and ornithology, where he played an exceptionally important role in promoting interest in the natural environment and its conservation. Clee Jenkins began writing for the West Australian in 1937, and his articles on an amazing range of conserva¬ tion and natural history subjects appeared each week from then up to the present time, he also wrote more than 1 000 articles for radio and TV broadcasting. The Council and members of the Royal Society of Western Australia recognise the outstanding contribu¬ tions of Clee Jenkins to promoting Science in Western Australia, and extend their condolences and best wishes to his family. [MGK Jones, President, RSWA] 290 5 702Z-