(ISSN 0161-8202) I Journal or ARACHNOLOGY PUBLISHED BY THE AMERICAN ARACHNOLOGICAL SOCIETY VOLUME 38 2010 NUMBERS THE JOURNAL OF ARACHNOLOGY EDITOR-IN-CHIEF: James E. Carrel, University of Missouri-Columbia MANAGING EDITOR: Douglass H. Morse, Brown University SUBJECT EDITORS: Ecology — StaBO Pekar, Masaryk University; Systematics — Mark Harvey, Western Aus- tralian Museum and Ingi Agnarsson, University of Puerto Rico; Behavior — Linden Higgins, University of Vermont; Morphology and Physiology — ^Jason Bond, East Carolina University EDITORIAL BOARD: Alan Cady, Miami University (Ohio); Jonathan Coddington, Smithsonian Institution; William Eberhard, Universidad de Costa Rica; Rosemary Gillespie, University of California, Berkeley; Charles Griswold, California Academy of Sciences; Marshal Hedin, San Diego State University; Herbert Levi, Harvard University; Brent Opell, Virginia Polytechnic Institute & State University; Norman Platnick, American Museum of Natural History; Ann Rypstra, Miami University (Ohio); Paul Selden, University of Kansas; Matthias Schae- fer, Universit*t Goettingen (Germany); William Shear, Hampden-Sydney College; Petra Sierwald, Field Mu- seum; I-Min Tso, Tunghai University (Taiwan). The Journal of Arachnology (ISSN 0161-8202), a publication devoted to the study of Arachnida, is published three times each year by The American Arachnological Society. Memberships (yearly): Membership is open to all those interested in Arachnida. Subscriptions to The Journal of Arachnology and American Arachnology (the newsletter), and annual meeting notices, are included with membership in the Society. Regular, $55; Students, $30; Institutional, $125. Inquiries should be directed to the Membership Secretary (see below). Back Issues: James Carrel, 209 Tucker Hall, Missouri University, Columbia, Missouri 65211-7400 USA. Telephone: (573) 882-3037. Undelivered Issues: Allen Press, Inc., 810 E. 10th Street, P.O. Box 368, Lawrence, Kansas 66044 USA. THE AMERICAN ARACHNOLOGICAL SOCIETY PRESIDENT: Rosemary Gillespie (2009-2011), Environmental Science, Policy & Management, Division of Organisms and Environment, University of California, Berkeley, CA 94720-3114 USA. PRESIDENTELECT: Jonathan Coddington (2009-2011), Smithsonian Institution, Washington, DC 20013-7012 USA. MEMBERSHIP SECRETARY: Jeffrey W. Shultz (appointed), Department of Entomology, University of Maryland, College Park, MD 20742 USA. TREASURER: Karen CangialosI, Department of Biology, Keene State College, Keene, NH 03435-2001 USA. SECRETARY: Alan Cady, Dept, of Zoology, Miami University, Middletown, Ohio 45042 USA. ARCHIVIST: Lenny Vincent, Fullerton College, Fullerton, California 92634 USA. DIRECTORS: Paula Cushing (2009-2011), Todd BlacMedge (2009-201 1), James Harwood (2010-2012) PAST DIRECTOR AND PARLIAMENTARIAN: H. Don Cameron (appointed), Ann Arbor, Michigan 48105 USA. HONORARY MEMBERS: C.D. Dondale, H.W. Levi, A,R Millidge. Cover photo: Eukoenenia maquinemis (Palpigradi: Eukoeneniidae), a new species from a Brazillian cave. Photo by Rodrigo Lopes Ferreira Publication date: 24 November 2010 @This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanenee of Paper). 2010. The Journal of Arachnology 38:393-397 The effects of predation risk on female silk deposition and male response to predator-cued conspecifics in the wolf spider, Pardosa milvina (Araneae: Lycosidae) Alexander L. Sweger, Michael Marr and Adam Satteson: Department of Biology, Susquehanna University, Selinsgrove, Pennsylvania 17870, USA Ann L. Rypstra: Department of Zoology, Miami University, Hamilton, Ohio 45011, USA Matthew H. Persons': Department of Biology, Susquehanna University, Selinsgrove, Pennsylvania 17870, USA Abstract. Female Pardosa milvina (Hentz 1844) wolf spiders advertise sexual receptivity toward males via silk draglines, and male P. milvina exhibit conspicuous courtship behavior when encountering silk from females. Previous studies suggest that female P. milvina may benefit by limiting silk advertisements and excreta deposition when encountering silk from the predator, H. helluo, and male P. milvina may exhibit corresponding reductions in courtship when encountering silk from conspecific females previously exposed to H. helluo silk. We tested these predictions by comparing the amount of silk and excreta deposited by unmated female P. milvina exposed or not exposed to predator cues (silk and excreta) from H. helluo. We also measured and compared male P. milvina courtship latency and intensity in the presence of silk from females previously exposed or not exposed to predator silk from H. helluo. Contrary to predictions, we found a significant increase in excreta, dragline, and attachment disk deposition after females were exposed to H. helluo cues. Male P. milvina courtship latency did not vary among treatments despite increases in female silk deposition, but males significantly decreased courtship intensity when exposed to silk from females under predation risk. Vertical climbing to escape the predator cues may cause an increase in female silk deposition. Keywords: Dragline, kairomone, pheromone, chemical cue, courtship Conspicuous ornamentation and courtship displays of male wolf spiders increase mating success by attracting females and provide information about species identity (Stratton & Uetz 1986; McClintock & Uetz 1996; Scheffer et al. 1996; Parri et al. 1997; Hebets & Uetz 1999, 2000). However, elaborate ornaments and complex displays may incur considerable fitness-related costs because of their potential to draw the attention of nearby predators (Kotiaho et al. 1998; Pruden & Uetz 2004; Roberts et al. 2007; Hoefler et al. 2008). Consequently, males may benefit by reducing their overall activity level, including courtship intensity and duration, when under predation risk (Kotiaho et al. 1998; Taylor et al. 2005; Hoefler et al. 2008). Female wolf spider dragline silk is an important medium for sexual communication (Kaston 1936; Hegdekar & Dondale 1969; Richter et al. 1971; Dondale & Hegdekar 1973; Tietjen 1979; Hebets & Uetz 1999; Rypstra et al. 2003; Schultz 2004; Gaskett 2007) and often serves as an advertisement to males. Although the predation costs associated with conspicuous male displays have been studied in a number of wolf spider species (Kotiaho 1998; Hebets 2005; Roberts et al. 2007; Hoefler et al. 2008), what remains less appreciated is that female silk may be chemically conspicuous to predators and increase predation risk as well, particularly among predators that use silk as a mode of communication. Consequently, females may benefit by reducing or modifying their silk advertisements toward males when aware of the presence of a predator. A number of studies demonstrate that the wolf spider Pardosa milvina (Hentz 1844) can detect fine differences in predation risk through silk and excreta cues left by the larger ' Corresponding author. E-mail: persons(^susqu.edu predatory wolf spider, Hogna helluo (Walckenaer 1837), and respond with graded reductions in activity proportional to the perceived risk. Pardosa milvina can determine how recently H. helluo was in the area (Barnes et al. 2002), its diet (Persons et al. 2001), its hunger level (Bell et al. 2006), and the quantity of silk and excreta present (Persons & Rypstra 2001). Such discrim- inatory facilities suggest that P. milvina has sophisticated chemoreceptive abilities and that there is strong selection for recognition and evaluation of varying levels of predation risk. Female Pardosa milvina exposed to PI. helluo silk and excreta reduce activity (Persons & Rypstra 2001) and increase vertical orientation and climbing behavior (Persons et al. 2002; Folz et al. 2006). Exposure to these same cues can change foraging patterns to such an extreme that it directly affects body condition, prey capture behavior, and egg sac produc- tion of females, ultimately affecting the direct fitness of the spider (Persons et al. 2002). When in the presence of conspecific adult females, male P. milvina respond to H. helluo cues by delaying courtship (Taylor et al. 2005). In addition, H. helluo is attracted to silk and excreta produced by female P. milvina when they have been fed a diet of P. milvina (Persons & Rypstra 2000). Given the potential predation costs associated with female silk advertisements, we addressed two questions: 1) Do female P. milvina alter the quantity and type of silk they deposit while in the presence of silk cues from the predatory wolf spider, H. helluol; 2) Do male P. milvina respond differently to silk cues from female P. milvina that were under predation risk, without being exposed to the predator cues themselves? We predicted that female P. milvina would reduce their silk deposition or possibly pheromones associated with the silk when detecting chemical cues produced from the predator, H. helluo. If females reduce chemical advertisements toward males when 393 394 THE JOURNAL OF ARACHNOLOGY under predation risk, then males should also show corre- sponding reductions in courtship behavior when detecting female silk deposited while the female was detecting a predator in the area, even when the male has no direct access to the predator’s cues. METHODS Spider collection. — Subadult female H. helluo and P. milvina were collected in agricultural fields near the campus of Susquehanna University, Selinsgrove, Pennsylvania, Snyder County, USA. To assure the virginity of test spiders, we allowed all spiders to mature in the laboratory. The P. milvina were maintained in 0.074 1 translucent containers (8 cm diameter, 5 cm height) and H. helluo were maintained in 0.473 1 (9.8 cm diameter, 8.5 cm height) plastic deli dishes. Each spider was given water ad libitum and fed a diet of juvenile (ca 0.25 cm) domestic house crickets (Aclieta domesticus Lin- naeus) for P. milvina, and adult house crickets for H. helluo. All spiders were fed 2-3 crickets every 3^ days, and then fed to satiety 1-2 h before testing to minimize the effects of body condition and hunger level on silk production and male courtship. We only used spiders that had finished feeding at the time of testing. General experimental protocol. — We collected data relevant to each question with a separate protocol and a different set of males and females. Males in the second protocol and females in both protocols were tested twice, once under conditions of predation risk and once under conditions of no predation risk. We randomized the order of exposure such that half of the test spiders were subjected to the predation risk treatment first while the other half were exposed to the “no predator cue” treatment first. This was done to minimize any confounding effect of experience or other associated sequence effect. We analyzed female silk and male courtship behavior using paired /-tests, with predator risk and no risk as independent variables. We conducted tests in September-December, 2006. Quantifying silk deposition. — We randomly chose 30 adult, laboratory-reared, virgin female H. helluo and used them in both parts of the experiment to deposit predator cues for multiple treatments. We placed each female H. helluo in a clean, 0.473 1 white plastic testing container (9.8 cm diameter, 8.5 cm height) for four hours on a black paper substrate (9.8 cm diameter) printed with a grid to allow us to quantify silk production. Each substrate was divided into a grid of 3 mm X 3 mm sub-squares (approximately 766 squares). In a second control treatment, we used no predator stimulus in the container but the container setup was otherwise identical. Once the H. helluo was removed, we examined each paper substrate for the presence and amount of silk under a dissecting stereomicroscope. Silk types included attachment disks, which were discrete butterfly or u-shaped strands of silk produced by the piriform glands, and thin linear dragline silk produced by the ampullate silk glands (Richter et al.l971). Each grid sub-square was then scored from 0^ based on the percentage of the area covered by silk (0 = 0%, 1 = 1-25%, 2 = 26-50%, 3 = 51-75%, 4 = 76-100%). We also counted the total number of drops of excreta and attachment disks across the entire grid surface. Hogna helluo silk served as a predator stimulus for female P. milvina. We quantified H. helluo silk and excreta prior to the introduction of the female P. milvina in order to calculate the amount of P. milvina dragline silk, excreta, and attachment disks. For control trials, we quantified P. milvina dragline silk, excreta, and attachment disks alone, with no predator stimulus. After each grid was scored, we placed a laboratory-reared, virgin, female P. milvina in the container for four hours to deposit cues on top of the H. helluo silk and excreta or on the blank substrate in control trials. Hogna helluo cues were scored within one-two hours of deposition before a female P. milvina was placed on the cues. After scoring the first two sets of 15 replicates per treatment {n = 30), we reversed the treatments and ran them a second time. After running female P. milvina, we scored the grids a second time, subtracting the initial H. helluo silk score to estimate female P. milvina silk deposition. Quantifying male Pardosa milvina courtship behavior. — To measure the effects of the context in which female silk was deposited on male courtship behavior, it was necessary to separate the cues of the predator H. helluo from those of the female P. milvina exposed to the predator cues (Fig. 1 A, B). In a clean, 0.473 1 container, we placed a single adult, laboratory-reared, virgin female H. helluo on a blank white paper substrate for four hours to deposit cues. We then removed the female H. helluo, and inserted another 0.473 1 dish with the bottom completely removed (a circle of 8 cm diameter), save for an approximately 9 mm rim around the bottom, into the original container. We placed an unmated laboratory-reared adult female P. milvina on the substrate in the presence of predator cues with the inserted dish for four hours. No food or water was provided during this time period. We then removed the inserted dish and wiped the bottom with a small amount of ethanol to deactivate and remove any remaining predator cues. After the ethanol had evaporated, we placed the insert in a clean dish with a male P. milvina and documented courtship behavior for 30 minutes while the male was exposed to only the female conspecific cues on the 9 mm rim. We ran an identical treatment without the initial presence of H. helluo cues (Fig. 1 B). In both treatments, the male was responding solely to the conspecific female silk. As in the protocol for females, both treatments were conducted with the same set of 24 male subjects, assigned randomly to first experience cues from either females with or females without exposure to a predator. We used paired /-tests to compare female silk deposition with and without predator cues. We also used paired /-tests to compare male courtship latency and courtship intensity across treatments. We defined courtship in P. milvina as a combination of two separate and distinct behaviors, leg raises and body shakes. A leg raise is raising the first pair of legs in unison above the cephalothorax and bringing them down abruptly. A body shake is a set of rapid oscillations of the abdomen and cephalothorax, often in conjunction with a leg raise. These behaviors have been described elsewhere and are known to influence female mate choice (Montgomery 1903; Kaston 1936; Brautigam & Persons 2003). Courtship latency consti- tutes the time elapsed between when a male was first placed on the substrate and when he began courtship (either with a leg raise or a body shake). We measured courtship intensity as the sum of body shakes and leg raises divided by courtship duration. SWEGER ET AL.— PREDATOR EFFECT ON LYCOSID SILK AND COURTSHIP 395 Figure 1. — Experimental containers for measuring behavioral response in male P. milvina with (A) and without (B) H. helliio silk present during female P. milvina silk deposition. A. An adult female H. helliio is allowed to deposit silk and excreta on the bottom of a container for 4 hours. A second clean container with the bottom removed (except for a shallow lip along the edge) is nested within the predator-cued container. A single unmated female is then allowed to deposit silk for four hours on top of the predator cues within these two nested containers. The inner container is then removed, the bottom cleaned, and nestled into a clean container devoid of predator silk. A single male P. milvina is then introduced into the container and its courtship behavior is quantified. Note that males have access only to silk that females deposited while they detected H. helliio chemical cues, but that the males have no silk cues available from H. helliio directly. B. The same procedure as in A is followed, except female P. milvina do not perceive predator cues during silk deposition. RESULTS We found that when exposed to cues (silk and excreta) from the larger predatory wolf spider Hogna helluo, female Pardosa milvina showed a significant increase in deposition of both silk and excreta. When we exposed male P. milvina to conspecific cues from females exposed to predator cues, we found that males significantly decreased courtship behaviors and courted significantly less intensely. Effects of predator cues on female silk deposition. — There was a large increase in total dragline silk (paired t = 2.449, P = 0.0206), attachment disks (paired t = 2.708, P = 0.0112) and excreta (paired t = 2.574, P - 0.0154) from female P. milvina when exposed to H. helluo cues (Fig. 2 A-C). Effects of predator-cued female silk on male courtship behavior. — There were significant decreases in male courtship behaviors when exposed to predator-cued, conspecific females. During courtship, males dramatically decreased the total number of body shakes, the intensity of leg raises and the intensity of body shakes when exposed to cues from females under predation risk (Table 1 for all behaviors). Total courtship intensity (leg raises and body shakes/courtship duration) also decreased among males exposed to predator- cued conspecifics (Table 1). Total leg raises showed a qualitatively similar decrease, but the decrease was not statistically significant (Table 1). DISCUSSION Our results clearly show that female P. milvina produce greater quantities of silk, excreta and attachment disks in the presence of predator cues from PI. helluo and that in some A Total Silk B Attachment Disks Figure 2. — Female Pardosa milvina silk and excreta deposition with and without silk from a predator present. A. Total silk score was calculated as the cumulative sum of all the sub-squares of each grid. Each grid square was scored from 0-4 based on an approximate percentage of the area of the square that was covered by silk. The total possible score if all grid squares were completely covered with silk would be 3064 (see text for more details). B. Deposition of female P. milvina attachment disks and C. Excreta. Unlike with the scoring method for total silk, each drop of excreta and attachment disk were discrete and were therefore counted directly. manner, a potential predator can indirectly affect the behavior of a male even though the male has no direct access to predator cues. Several studies have examined the effects of conspicuous courtship displays of wolf spiders on the probability of attacks from nearby predators (Kotiaho et al. 1998; Pruden & Uetz, 2004; Roberts et al. 2007), but to our 396 THE JOURNAL OF ARACHNOLOGY Table 1. — Measures of male courtship behavior (mean ± S.E.) in response to female silk deposited while the female was exposed to H. helliio silk and excreta or control treatments without H. Iielluo cues. R-values are based on paired t tests. See text for definitions of courtship behaviors. Behavior Treatment w/ Hogna cues Control / P Leg raise number 21.33 ± 5.2250 46.79 ± 15.005 1.917 0.0677 Body shake number 2.54 ± 0.8470 12.08 ± 4.125 2.718 0.0123 Leg raise intensity 0.0123 ± 0.0030 0.0279 ± 0.008 2.135 0.0436 Body shake intensity 0.00148 ± 0.0005 0.00699 ± 0.002 2.770 0.0109 Total courtship intensity 0.0137 ± 0.0030 0.0349 ± 0.011 2.318 0.0297 knowledge, this is the first study documenting that male courtship behavior can be modified in direct response to the predation context of female silk deposition. Female silk deposition increased while under predation risk, yet male courtship intensity significantly decreased when encountering silk from a conspecific female that had previously detected predator silk. This surprising result runs counter to several of our predictions. When adult H. helliio have been reared on a diet of P. milvina, they preferentially choose to forage in areas where adult female P. milvina silk and excreta cues are present (Persons & Rypstra 2000). This suggests that females that deposit greater quantities of silk may attract H. helluo. If a female is under perceived predation risk, then it may be adaptive to limit the amount of silk deposited to reduce attracting predatory wolf spiders. We might then expect that reductions in silk would lead to an impaired ability of males to detect females, which would translate into increased courtship latency and less intense courtship. However, our results showed that females signifi- cantly increased total silk deposited, attachment disks, as well as excreta when under perceived predation risk. The females that were exposed to a potential predator’s cues could be engaging in a number of different anti-predator tactics. Female P. milvina increase climbing behavior and attempt to move up the side of a container when the horizontal substrate contains silk and excreta from H. helluo (Persons et al. 2002; Folz et al. 2006). This behavior would likely require a greater number of attachment disks as a means of anchoring the spider while climbing. If so, climbing females may release significantly more silk in an effort to escape the container. Alternatively, increased deposition of silk may simply be a non-adaptive reflexive response to an acute stressor such as the presence of predator cues. Excretion increased in the presence of predator silk as well. Excretion in the presence of a predation threat is known to be an effective anti-predator response for a number of arthropods by either increasing locomotor efficiency or repelling potential predators (Weiss 2006). It is possible that H. helluo silk caused P. milvina to avoid the center of the container, resulting in proportionally more time spent exhibiting thigmotaxis (wall- hugging). We suggest this is unlikely, however, since numerous studies have shown that female P. milvina spend significantly more time on H. helluo-cuQd substrates due to freezing responses. Anecdotal observations of spider dragline deposi- tion suggest that deposition rate is a function of time spent moving. Given that numerous studies have shown that female P. milvina locomotion decreases in the presence of H. helluo cues (e.g.. Persons et al. 2001; Persons & Rypstra 2001), it suggests that the deposition rate must dramatically increase as a proportion of time spent moving. The fact that adult males reduced courtship when detecting larger quantities of female silk suggests that there is not a simple positive relationship between the quantity of female silk and male courtship intensity. If anything, our results indicate a negative relationship between these variables. The relative importance of the tactile component of silk versus sex pheromones perfused on silk remains unknown. At least some wolf spider species respond to silk cues that are devoid of pheromones (Tietjen 1977; Tietjen & Rovner 1982), while other wolf spiders, such as P. milvina, are able to produce airborne pheromones that induce responses in males indepen- dent of male contact with silk (Searcy et al. 1999). Our results suggest that females could have reduced sex pheromone deposition on the silk even while increasing the amount of silk. If H. helluo attraction to female P. milvina is based on these pheromones rather than silk, increased silk deposition would have little consequence in attracting H. helluo. Also, the strength of H. helluo preference for cues associated with P. milvina may not necessarily be dependent on the quantity of P. milvina silk or excreta, but only on the presence or absence of these cues. Regardless of the proximate cue used to elicit courtship behavior in males, our study provides indirect evidence that qualitative differences in the silk may be more important than quantitative differences in mediating male courtship displays. Pardosa milvina appears to produce at least two discrete types of dragline silk, heavy gauge cord silk and fine gauge dragline silk, which has a diameter approximately ten times smaller than heavy gauge silk. The ratio or quantity of these two silk types could vary markedly with and without predator cues, and each silk type may convey different information to males. One hypothesis that could explain the significant decrease in several male courtship behaviors is that there is a change in the cues deposited among the females when under predation risk. Regardless of the specific source of the female stimulus, males appear capable of appropriately reducing courtship levels which, in turn, likely decreases the probability of predation. Since we know almost nothing about the quantitative relationship between sex pheromone production and silk production in lycosids, this explanation must remain specu- lative. However, given the sophisticated ability of P. milvina to extract a variety of information about predators from silk and excreta cues alone, we believe their chemoreceptive capabilities may be sufficiently advanced that an indirect transfer of SWEGER ET AL.— PREDATOR EFFECT ON LYCOSID SILK AND COURTSHIP 397 information about the presence of a predator is possible through conspecific female P. milvina. Additional studies that examine differences in predation risk based on the quantity of female silk deposited may prove fruitful. ACKNOWLEDGMENTS We thank Chris Latanich and Becky Holler for their help collecting and maintaining spiders used for this study, as well as technical assistance. This research was funded in part through NSF grant C-RUI DBI-02 16776 (for M. Persons) and C-RUI DBI 0216947 (for A. Rypstra). We also thank members of the Rypstra Laboratory for their helpful comments of an earlier draft of this paper. LITERATURE CITED Bell, R.D., A.L. Rypstra & M.H. Persons. 2006. The effect of predator hunger on chemically mediated antipredator responses and survival in the wolf spider Pardosa milvina (Araneae: Lycosidae). Ethology 112:903-910. Barnes, M.C., M.H. Persons & A.L. Rypstra. 2002. The effect of predator chemical cue age on antipredator behavior in the wolf spider Pardosa milvina (Araneae: Lycosidae). Journal of Insect Behavior 15:269-281. Brautigam, S.E. & M.H. Persons. 2003. The effect of limb loss on the courtship and mating behavior of the wolf spider Pardosa milvina (Araneae: Lycosidae). Journal of Insect Behavior 16:571-587. Dondale, C.D. & B.M. Hegdekar. 1973. The contact sex pheromone of Pardosa lapidicina Emerton (Araneae: Lycosidae). Canadian Journal of Zoology 51:400^01. Folz, H.C., S.M. Wilder, M.H. Persons & A.L. Rypstra. 2006. Effects of predation risk on vertical habitat use and foraging of Pardosa milvina. Ethology 112:152-1158. Gaskett, A.C. 2007. Spider sex pheromones: emission, reception, structures, and functions. Biological Reviews 82:27^8. Hebets, E.A. 2005. Attention-altering signal interactions in the multimodal courtship display of the wolf spider Schizocosa iietzi. Behavioral Ecology 16:75-82. Hebets, E.A. & G.W. Uetz. 1999. Female responses to isolated signals from multimodal male courtship displays in the wolf spider genus Schizocosa (Araneae: Lycosidae). Animal Behaviour 57:865-872. Hebets, E.A. & G.W. Uetz. 2000. Leg ornamentation and the efficacy of courtship display in four species of wolf spider (Araneae: Lycosidae). Behavioral Ecology and Sociobiology 47:280-286. Hegdekar, B.M. & C.D. Dondale. 1969. A contact sex pheromone and some response parameters in lycosid spiders. Canadian Journal of Zoology 47:1-4. Hoefler, C.D., M.H. Persons & A.L. Rypstra. 2008. Evolutionarily costly courtship displays in a wolf spider: a test of viability indicator theory. Behavioral Ecology 10:1-6. Kaston, B.J. 1936. The sense involved in the courtship of some vagabond spiders. Entomology Americana 16:97-167. Kotiaho, J., R.V. Alatalo, J. Mappes, S. Parri & A. Rivero. 1998. Male mating success and risk of predation in a wolf spider: a balance between sexual and natural selection? Journal of Animal Ecology 67:287-291. McClintock, W.J. & G.W. Uetz. 1996. Female choice and pre-existing bias: visual cues during courtship in two Schizocosa wolf spiders (Araneae: Lycosidae). Animal Behaviour 52:167-181. Montgomery, T.H. 1903. Studies on the habits of spiders particularly those of the mating period. Proceedings of the Academy of Natural Sciences of Philadelphia 55:59-149. Parri, S., R.V. Alatalo, J. Kotiaho & J. Mappes. 1997. Female choice for drumming in the wolf spider Hygrolycosa ruhrofasciata. Animal Behaviour 53:305-312. Persons, M.H. & A.L. Rypstra. 2000. Preference for chemical cues associated with recent prey in the wolf spider Hogna helliio (Araneae: Lycosidae). Ethology 106:27-35. Persons, M.H. & A.L. Rypstra. 2001. Wolf spiders show graded antipredator behavior in the presence of chemical cues from different sized predators. Journal of Chemical Ecology 27: 2493-2504. Persons, M.H., S.E. Walker, A.L. Rypstra & S.D. Marshall. 2001. Wolf spider predator avoidance tactics and survival in the presence of diet-associated predator cues (Araneae: Lycosidae). Animal Behaviour 61:43-51. Persons, M.H., S.E. Walker & A.L. Rypstra. 2002. Fitness costs and benefits of antipredator behavior mediated by chemotactile cues in the wolf spider Pardosa milvina (Araneae: Lycosidae). Behavioral Ecology 13:38^392. Pruden, A.J. & G.W. Uetz. 2004. Assessment of potential predation costs of male decoration and courtship display in wolf spiders using video digitization and playback. Journal of Insect Behavior 17:67-80. Richter, C.J.J., H.C.J. Stolting & L. Vlum. 1971. Silk production in adult females of the wolf spider Pardosa amentata (Lycosidae, Araneae). Journal of Zoology 165:285-290. Roberts, J.A., P.W. Taylor & G.W. Uetz. 2007. Consequences of complex signaling: Predator detection of multimodal cues. Behavioral Ecology 18:236-240. Rypstra, A.L., C. Wieg, S.E. Walker & M.H. Persons. 2003. Mutual mate assessment in wolf spiders: differences in the cues used by males and females. Ethology 109:315-325. Scheffer, S.J., G.W. Uetz & G.E. Stratton. 1996. Sexual selection, male morphology, and the efficacy of courtship signaling in two wolf spiders (Araneae: Lycosidae). Behavioral Ecology and Sociobiology 38:17-23. Schultz, S. 2004. Semiochemistry in Spiders. Pp. 110-150. In Advances in Insect Chemical Ecology. (R.T. Carde & J.G. Millar, eds.). Cambridge University Press, Cambridge, UK. Searcy, L.E., A.L. Rypstra & M.H. Persons. 1999. Airborne chemical communication in the wolf spider Pardosa milvina. Journal of Chemical Ecology 25:2527-2533. Stratton, G.E. & G.W. Uetz. 1986. The inheritance of courtship behavior and its role as a reproductive isolating mechanism in two species of Schizocosa wolf spiders (Araneae: Lycosidae). Evolution 40:129-141. Taylor, A.R., M.H. Persons &. A.L. Rypstra. 2005. The effect of perceived predation risk on male courtship and copulatory behavior in the wolf spider Pardosa milvina (Araneae: Lycosidae). Journal of Arachnology 33:76-81. Tietjen, W.J. 1977. Dragline-following by male lycosid spiders. Psyche 84:165-178. Tietjen, W.J. 1979. Is the sex pheromone of Lycosa rabida (Araneae: Lycosidae) deposited on a substratum? Journal of Arachnology 6:207-212. Tietjen, W.J. & J.S. Rovner. 1982. Chemical communication in lycosids and other spiders. Pp. 249-280. In Spider Communication: Mechanisms and Ecological Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Weiss, M.R. 2006. Defecation behavior and ecology of insects. Annual Review of Entomology 51:635-661. Manuscript received I October 2009, revised 15 May 2010. 2010. The Journal of Arachnology 38:398^06 The visual system of the Australian wolf spider Lycosa leuckartii (Araneae: Lycosidae): visual acuity and the functional role of the eyes Christofer J. Clemente’, Kellie A. McMaster-, Elizabeth Fox^, Lisa Meldrum'', Tom Steward and Barbara York Main"*: 'The Rowland Institute at Harvard, Harvard University, Cambridge, MA 02142, USA. E-mail: clemente@rowland.harvard.edu; -Outback Ecology, 1/71 Troy Terrace Jollimont, Western Australia 6014, Australia; ^Ecologia Environment, 1025 Wellington Street, West Perth, Western Australia 6005, Australia; ‘‘School of Animal Biology M092, University of Western Australia, 35 Stirling Highway, Crawley, Western Australia, 6009, Australia Abstract. Ocular arrangement and visual acuity were examined in Lycosa leuckartii Thorell 1870 (Araneae: Lycosidae), using histological techniques. Major structural and functional features of the visual system, including external and internal ocular organizations, resolution, sensitivity, focal lengths and the field of view, were characterized for each eye. Lycosa leuckartii had a large developmental investment in a specialized visual system with high visual acuity. The field of view extended 360° and displayed the potential for good depth perception. Anterior eyes showed average focal lengths (AL eyes 230.88 pm, AM eyes 276.84 pm), while the posterior eyes far exceeded them (PL eyes 499.26 pm, PM eyes 675.35 pm). Resolution of the anterior eyes was comparable to records in the literature for other lycosids (inter-receptor angle AL eyes 2.45°, AM eyes 1.85°), while the resolution of the posterior eyes was higher (PL eyes 0.78°, PM eyes 0.67°). Sensitivity of the lens (/-numbers) was highest in the secondary eyes and was close to some of the highest reported for Araneae (/-numbers PM eyes 0.58), but when receptor diameters were included in estimates, S-numbers were similar or lower than closely related species (PL eyes 17.5 pm", PM eyes 17.6 pm"). There is a clear distinction in organization and function between the posterior and anterior eyes of L. leuckartii. The posterior eyes suit long- range predator and prey detection, while the anterior eyes are best for distance judgment and prey capture. Keywords: Field of view, focal length, resolution, sensitivity The modern arachnids are the only group of arthropods in which the eyes are camera-type, similar to our own, rather than compound eyes (Land 1985). Despite this, the structure and function of the visual system of spiders has been inadequately studied in many families of spiders when compared to chemo- and mechanoreception. This is likely based on the assumption that most spiders are nocturnal and vision may be of limited use (Foelix 1982). However, for some species, vision can be very important. Members from at least one family, the Salticidae, have been shown to hunt exclusively using vision (Jackson 1977), and members of five other families of spiders, Lycosidae, Pisauridae, Thomisidae, Oxyo- pidae and Deinopidae, show visually guided behaviors during locomotion, homing, prey capture, and courtship (Bristowe & Locket 1926; Kaston 1936; Whitcomb & Eason 1965; Robinson & Robinson 1971; Forster 1982; Uetz & Stratton 1982; Rovner 1996). The visual acuity of a species is determined by character- istics of the eyes such as field of view, focal length, resolution, and sensitivity. The external placement and internal arrange- ment determine the field of view of each eye (Land 1985). The ancestral eye arrangement, as hypothesized by Homann (1971), consists of two transverse rows, each containing four eyes. The first row consists of the anterior median (AM) eyes in the middle and the anterior lateral (AL) eyes on the periphery. Similarly, the posterior eyes are grouped into posterior median (PM) eyes and posterior lateral (PL) eyes. The visual angle of the field of view for each eye can vary greatly, from relatively narrow pinpoint views of only 24° in the PL eyes of Baditmna insignis Koch 1872 (Clemente et al. 2005) to 182° wide-angle views in the AM eyes of Octonoba sinensis Simon 1880 (Opell 1988). Forward-facing binocular vision is a product of overlapping visual fields, and is necessary for good distance judgment. The distance over which an eye can focus upon an object is determined by the focal length of its lens (Homann 1971). This ranges from 38 pm in the AL eyes of the uloborid Hyptiotes cavatus Hentz 1847 (Opell & Ware 1987), to 448 pm in the PM eyes of Cupiennius salei Keyserling 1877 (Land & Barth 1992), up to 1980 pm in the AM eyes of the jumping spider Portia (Williams & McIntyre 1980). The ability of the eye to resolve detail depends on the fineness of the retinal mosaic, usually expressed as the inter- receptor angle. The finer this angle, the better the resolution of the eye. The finest inter-receptor angles reported in the literature are for members of Salticidae, the diurnal jumping spiders, for which the inter-receptor angle in the AM eyes can be less than 1°. The largest angle reported for hunting spiders is 7° in the AL eyes of a lycosid species (Homann 1931; Land 1969, 1985). Sensitivity, or the ability to see in low light levels, is a combination of the physical properties of an optical system and the physiological sensitivity of photoreceptors. The physical ability of the lens to admit light is often expressed in the literature as an /^number, which decreases with increasing sensitivity (Opell & Ware 1987). This ranges from 2.68-5.90 in diurnally active jumping spiders (Land 1969; Foelix 1982) to 0.58 in the PM eyes of the wholly nocturnal ogre-faced spider Deinopis suhriifa, Koch 1879 (Blest & Land 1977). However, a more complete estimate of sensitivity is given by an S-number (Land 1981), which is a product of the relative aperture of the eye, determining the light flux passing through to the retina, the cross sectional area of the receptor, and the proportion of light entering a receptor that is actually 398 CLEMENTE ET AL.— VISUAL SYSTEM OF A WOLF SPIDER 399 Figure 1 . — External measurements taken on L. leuckartii. A = anterior view, B = dorsal view. AM, anterior median eyes; AL, anterior lateral eyes; PM, posterior median eyes; PL, posterior lateral eyes; TED, total eye diameter; TEW, total eye width; IS, interocular space. absorbed by it. This has the advantage of increasing as sensitivity increases, and can range from 0.09 in the AM eyes of jumping spiders to 387 pm^ in the PM eyes of Deinopis (Land 1985). Both resolution and sensitivity vary in relation to the light conditions under which species operate (Opell & Ware 1987). While resolution improves as the ratio of receptor diameter to focal length decreases, sensitivity improves as the same ratio increases. Therefore, in the structure of the eyes, there is a trade-off between resolution and sensitivity. The only solution to the trade-off is to increase the total size of the eye; therefore, total eye size can be an indicator of the relative importance of vision. Many studies link various eyes to their probable role or relative use in prey acquisition (Uehara et al. 1978; Forster 1979; Kovoor et al. 1992; Rovner 1993; Schmid 1998; Ortega- Escobar & Munoz-Cuevas 1999). There are two important components of prey acquisition. The first is prey detection, which incorporates initial detection of an item, specifically distinguishing the item as prey, predator or conspecific, and some orientation and movement toward the prey item. Therefore, prey detection requires long-distance detection and image clarity. The second component is prey capture. This may involve some identification of prey or non-prey items and orientation toward the prey, but mainly comprises judgment of distance for accurate lunging and striking (Lizotte & Rovner 1988). Previous studies have demonstrated that the visual system of lycosid spiders is particularly complex. The eyes of lycosids have intricate visual fields (Homann 1931), high resolution and sensitivity (Lizotte & Rovner 1988; Kovoor et al. 1992; Rovner 1993; Ortega-Escobar & Munoz-Cuevas 1999) and the ability to detect polarized light (Kovoor et al. 1993; Dacke et al. 2001; Ortega-Escobar 2006). However, much of this work has been performed on relatively few species of lycosids, and it is unclear how much variation exists within the family. A recent molecular phylogeny of lycosids has suggested that the family consists of several clades (Murphy et al. 2006). While much of the work on vision in lycosids has focused on Palaeartic and Nearctic species, the Australian species form a separate distinct clade. Almost no information is available on variation among clades. As a taxonomic aside, it is noted that the generic position of Lycosa leuckartii is currently under review by V.W. Framenau with a proposed reallocation of the species pending. We present details of the visual acuity of Lycosa leuckartii and compare them to other lycosids and other families of spiders. METHODS External ocular organization. — Twenty individual Lycosa leuckartii were used to record external measurements. These included total eye width (TEW), total eye depth (TED) and eye diameters (Figs. lA & B). We took measurements under a binocular dissecting microscope with an eyepiece micrometer. The values were then standardized for the animals’ size by dividing each measurement by the carapace length. A repeated-measures ANOVA with one within-subject factor and no between-subject factors, and Student-Newman-Keuls post-hoc test, were used to determine significant differences in the relative eye diameters. Internal ocular organization. — Two specimens of L. leuck- artii were killed, using CO2 gas, trimmed to a small block of tissue and fixed in Karnovsky’s fixative for at least 72 h. We then washed and further trimmed down specimens in spider saline (scorpion saline excluding the CaCL: Zwicky 1968) and placed them in phosphate buffer prior to their being embedded in araldite/procure. Longitudinal and transverse (frontal plane) sections (1 pm thin) were cut using an LKB ultratome and a diamond knife. We mounted sections on slides and stained them with toluidine blue. As well as determining the internal ocular organization, we also used these sections in measurements of resolution and sensitivity. Focal length. — The focal length (/) of each lens was determined using the ‘hanging drop’ method described in Homann (1928) and Land (1985). The lens, along with a small proportion of the surrounding cuticle, was dissected from the head and stored in spider saline. After being cleared of excess tissue in warm, dilute, sodium hydroxide, the lens was suspended in a drop of spider saline from the underside of a cover slip. Using a microscope, we then viewed the image through the lens, targeting an object of known size (o). The distance between the slide and the object was then measured using calipers (u). The size of the image (i) was ascertained, using calibrated digital images, and the focal length was 400 THE JOURNAL OF ARACHNOLOGY calculated, using the formula (1) described by Opell & Ware (1987): /=(i/o)(u) (1) For each lens type, we determined an average of the values measured. Repeated-measures ANOVA, with one within- subject factor (eye) and no between-subject factors, with a Tukey-Kramer post-hoc test, were used to determine differ- ences between the focal lengths of the different eyes. Sensitivity. — Sensitivity (/’-number), or the eye’s ability to admit light, was calculated using values for focal length if) and the diameter of the retina (</), measured from the extremities of the rhabdomeres in each species (Opell & Ware 1987). We determined focal lengths by the above methods and ascer- tained retinal diameters by taking measurements from slides obtained using methods described for internal ocular organi- zation. These values were then entered into the sensitivity equation (2) outlined in Opell & Ware (1987): /-number =//<7 (2) The sensitivity of the eye can also be given in terms of an S- number, as described by Land (1985). Here sensitivity is defined, not only by the relative aperture of the eye, but also by the cross sectional area of the receptor and the amount of light absorbed by the receptor, as shown in equation (3): s=(jr(^)V;)(i -<•-*') (3) where D is the diameter of the lens, /is the focal length, dr is the receptor diameter (assumed to equal the center-to-center spacing of the receptors), / is the length of the receptor and k is the extinction coefficient of the photopigment in the receptors. Following Land (1981, 1985) k was approximated to be 0.0067 for rhabdomeric photoreceptors, and / was multiplied by 2 in the secondary eyes as the reflective tectum may allow photons to bounce back past the receptor, effectively lengthening it. Resolution. — We counted the numbers of axons exiting each eye from sections cut using the same methods as for internal ocular organization. Resolution is dependent upon the number of photoreceptors, or rhabdomeric cells per eye. The higher the density of cells, the finer the resolution of an image (Land 1985). The number of nerve axons exiting a spider’s eye is in a 1:1 ratio with the number of photoreceptors (Uehara & Uehara 1996). To compare the density of visual cells per eye, we figured the inter-receptor angle (de) based upon Land (1985) as given by equation (4): where d^c is the center-to-center spacing of retinal receptors and / is the focal length. The inter-receptor angle was calculated by measuring the pigment ring diameter of the retinal mosaic from histological sections and using the average maximum visual angle from the field of view (see below) to calculate the total area of the retinal mosaic. This was divided by the total numbers of photore- ceptors per eye to give an estimate of receptor area, and hence diameter. Retinal diameter was assumed to equal dec based on Land (1985). Table 1. — Optical data and sensitivities of the eyes of Lycosa leiickartii. Where multiple measurements were taken, mean plus standard error is shown. D = diameter of the lens, / = focal length, /c = center to center spacing of the photoreceptors, / = number as calculated in Equation 3. S-numbers, as calculated in Equation 3, are shown based upon the length of the tapetum. S-numbers based on 2x length of the tapetum are shown in parentheses. Eyes D |im / |xm Jec |im /-number S-number pm^ AM 292 ± 10.4 276.84 ± 16.89 8.93 1.08 - AL 230 ± 8.4 230.88 ± 5.74 9.87 0.69 - PM 752 ± 21.9 675.35 ± 56.19 7.96 0.56 9.82 (17.6) PL 647 ± 18.7 499.26 ± 13.40 6.81 0.75 9.72 (17.5) Field of view. — We used a Welch Allyn medical ophthal- moscope, along with an Aimark perimeter arc (153 mm diameter) constructed for use on spiders, to determine the extent of visual fields for L. leuckartii. A freshly killed L. leuckartii was utilized, and measurements were taken in a darkened room, using the light reflecting from the tapetum to determine limits of the fields of view of each eye. We moved the ophthalmoscope around the perimeter arc, viewing the bright green reflection from the tapetum of the secondary eyes, or the dull red reflection from the rhabdomeres of the primary eyes. Readings were taken on the perimeter arc at every 10° from horizontal, and were plotted directly onto a Geological Stereonet by rotating the stereonet 10° between each point. A stereonet is typically used to plot three-dimensional angles in two dimensions and, similar to the Aitoff equal area projection (utilized by Land & Barth 1992), the resultant plot represents 180° of a globe at infinity with the spider at the center. Average maximal visual angle (AMVA) was calculated by averaging the maximum horizontal and vertical arcs. Representative specimens from the study population along with slide preparations are held in the Zoology Building, School of Animal Biology, University of Western Australia. RESULTS External ocular organization. — The average carapace length of L. leuckartii was 826 pm. Lycosa leuckartii’ s eyes appear to form three rows. The first of these is similar to the ancestral arrangement, a single row consisting of the AL and AM eyes. With subtle evolutionary modifications, the PL eyes have moved around and back to form a row separate from the central row of the PM eyes. These modifications result in the total eye width being similar to total eye depth (TEW = 247 ± 4.8 pm; TED = 221 ± 4.4 pm). For the Lycosa leuckartii specimen, all pairs of eyes showed significant differences in diameters from one another (Fi9,3 = 406.9, P < 0.001), the forward facing PM eyes had the largest diameter, closely followed by the obliquely oriented PL eyes, and then the AM and AL eyes (Table 1). Internal ocular organization. — The AM eyes display a typical bi-convex lens formed by a visible thickening of the cuticular layer. The lens is separated from the retina by a layer of columnar vitreous cells. The retina is composed of visual cells and pigment cells. The most anterior portion of the visual cell, which contains the rhabdomeres, borders the vitreous CLEMENTE ET AL.— VISUAL SYSTEM OF A WOLF SPIDER 401 Figure 2. — Structure of the posterior lateral eye (PL) and the posterior medial eye (PM) of L. leuckartii, in a transverse section of the prosoma. layer, and the nuclei lie below. The secondary eyes of L. leuckartii have a distinct ‘grid-shaped’ tapetum (Fig. 2). A thin dark layer of visual cells containing the rhabdomeres separates the vitreous cells and the tapetum. The length of the tapetum measured directly from cross-sections averaged 31.75 and 32.22 pm long for the PL eyes and the PM eyes, respectively. In the anterior eyes of L. leuckartii, one discrete nerve bundle was found to emerge from each eye, but the posterior eyes produced multiple bundles of nerve axons. These emerged from various points around the perimeter of the eyecups and moved through the middle of the prosoma toward the supra- esophageal ganglion (Fig. 3). The PL eyes together gave rise to 18 bundles of axons. Each contained from 92 to 987 nerve axons per bundle. The PM eyes together generated 35 bundles. Each of these contained from 30 to 728 nerve axons per bundle. Focal length. — There were significant differences in the focal lengths of the four pairs of eyes in L. leuckartii (F/ j = 42.4, P = 0.006; Table 1). A Tukey-Kramer post-hoc test revealed that there were significant differences between the anterior and posterior pairs of eyes. Both pairs of anterior eyes had similar focal lengths (average = 253.86 pm). The posterior eyes were similar to each other, but showed much longer focal lengths than the anterior eyes (average = 587.30 pm). Direct inspection of the image produced by the lens showed that it was of good quality and therefore not subject to spherical aberration (Fig. 4). Sensitivity. — Based upon /-numbers, the highest sensitivity for L. leuckartii was found in the PM eyes. The other secondary eyes of L. leuckartii (PL and AL) also displayed similarly low /-numbers, indicating high sensitivity (Table 1). The highest /-number in L. leuckartii was found in the AM eyes, suggesting it has lower light sensitivity than the secondary eyes. S-numbers calculated for the PL eyes were similar to estimates for S-number of the PM eyes (Table 1). Resolution. — The optic nerves from all four pairs of eyes of L. leuckartii were found grouped together, surrounded by muscle along the midline of the prosoma. Identification of which nerve bundles originated from the left and right anterior eyes was possible by observing the arrival sequence (within the slides) of each nerve bundle and the direction from which it originated. The multiple nerve bundles originating from the posterior eyes of L. leuckartii could be distinguished as belonging to the PM or PL eyes, but could not be further separated into left and right eye nerve bundles; thus, total counts were made and half of this attributed to each eye. The posterior eyes of L. leuckartii gave rise to approxi- mately 12 times the number of nerve axons seen in the anterior eyes (Table 2). Receptor diameters were larger for the anterior eyes (average = 9.4 pm) when compared to the posterior eyes (average = 7.4 pm; Table 1). Accounting for eye size and visual fields, we concluded that this translated into greater resolution, the inter-receptor angle for the posterior eyes being less than a third that of the anterior eyes (Table 2). This appears to be below the potential resolution of the eyes. Following Land & Barth (1992), we discovered that the ultimate limit to resolution is determined by the blur resulting from diffraction at the aperture, given by 57.3ii7D deg, where w is the wavelength of light (assumed to be 0.5 pm) and D is the diameter of the lens. This results in potential grating 402 THE JOURNAL OF ARACHNOLOGY Figure 3. — Multiple nerve bundles of the posterior eyes of L. leiickcirtii. Upper panel = lOX magnification in transverse section; Lower panel = magnified (lOOX) view of the enclosed area above showing the discrete nerve bundles. CLEMENTE ET AL.— VISUAL SYSTEM OF A WOLF SPIDER 403 Figure 4. — View through the lens of L. leuckartii, suspended from a drop of saline underneath a glass cover-slip. The image is focused on two parallel lines. periods between 0.12° (AL eyes) and 0.04° (PM eyes), which are an order of magnitude smaller than the inter-receptor angles measured, suggesting diffraction does not limit resolution. Field of view. — Visual fields for L. leuckartii (Figs. 5A & B) were found to extend 360° around the animal, with a large overlap within 70° of the center. The PL eyes provided 220° of peripheral vision along the horizontal meridian, extending around the animal’s sides and overlapping behind the animal. AMVA of each individual eye was greater than 80° (Table 1) and up to 107.8° in the PL eyes. The other three pairs of eyes Table 2. — Optical data and resolution of the eyes of Lycosa leuckartii. d = diameter of the pigment ring, AMVA = average maximal visual angle, Ar = calculated area of the retina based upon AMVA, A0 = inter-receptor angle based upon Equation 4. Eyes d (pm) AMVA Nerve no. axons Ar pm^ Aa AM 255.7 82.4 476 29,797 1.85 AL 335.8 80.2 257 19,670 2.45 PM 1194.7 96.9 4848 241,208 0.68 PL 662.5 107.8 4423 160,871 0.78 (AM, AL and PM) all had some degree of forward facing overlap, indicating the potential for good overall binocular vision. The PM eyes, while covering 120° vertical maximum and 140° along the horizontal meridian, did not greatly overlap (maximum overlap 7° between 0 to 10° above the horizontal). The AL eyes provided a large degree of overlap (maximum 57°), but their visual axis was directed well below the horizontal (maximum overlap at 42° below horizontal). The AM eyes had a potentially large amount of binocular vision, with a maximum overlap of 42° at 24 to 30° above the horizontal, and with overlap extending from 20° below to 60° above horizontal. DISCUSSION The greatest differentiation of the eyes of L. leuckartii occurs between the posterior eyes and the anterior eyes, suggesting that these eyes play different roles in the visual system. This is most evident in the overlapping visual fields of not only the AL eyes with the PM eyes, as was shown by Homann (1931), but also of the AM and PM eyes (Fig. 5). The visual fields of the posterior eyes have large visual angles and cover an almost 360° view, while the anterior eyes appear to be focused forward. 404 THE JOURNAL OF ARACHNOLOGY Qone eye's field of view HFour eyes □two eyes BFive eyes Hxhree eyes Hsix eyes Figure 5. — Fields of view for L. leuckartii: A = frontal view; B = overhead view. Differentiation in focal length is also evident between anterior and posterior eyes. These estimates of focal length can also be used to calculate the eye’s depth of focus, which is the animal’s nearest distance of clear vision. Following Land (1981), we determined that the nearest distance of clear vision (U) is given by (7 = fDI2dcc, where /is the focal length, D is the diameter of the lens, and dec is the center-to-center spacing of the photoreceptors. This results in values for U of 4.5 mm for the AM eyes and 2.7 mm for the AL eyes. This is much less than the length of the legs, which suggests even the closest objects appear in focus. The posterior eyes, in contrast, have much larger U values of 32 mm (PM eyes) and 24 mm (PL eyes) and may therefore be of limited use at close range. The posterior eyes also show potential for superior performance in both resolution and sensitivity, when com- pared to the anterior eyes. This implies that the posterior eyes may be best suited for long-range, wide-angle recognition of objects in low light conditions and would therefore be ideal for tasks such as prey detection. Similar predictions were reported based upon behavioral experiments on the lycosid spider Rahidosa rabida Walckenaer 1837 (Rovner 1993). When different combinations of eyes were occluded in R. rabida, spiders with usable PL eyes were able to perform sizable orientations (up to 160°) toward a stimulus, while the PM eyes were found essential for mediating long-range approaches toward the stimulus (Rovner 1993). This suggests the PL and PM eyes could determine the outer limits of the spider’s visual perception and foraging patch. In contrast, the short focal length and considerable binocular vision of the anterior eyes indicate good potential depth perception at short range, which may be an important component for short-range orientation and approaches during activities such as prey capture or courtship. This has been supported behaviorally for the AL eyes by Rovner (1993), but not for the AM eyes. However, a comparison of the roles of anterior eyes between L. leuckartii and R. rabida may be difficult, since the AM eyes are smaller than the AL eyes in R. rabida, while the opposite is true for L. leuckartii (AM eyes larger than AL eyes). This may signify a greater role in orientation, or more likely, approach, toward stimuli, for the AM eyes in L. leuckartii. Alternatively, the AM eyes have also been shown to play a role in orientation via polarized light (Magni et al. 1964, 1965; Magni 1966; Ortega-Escobar & Munoz-Cuevas 1999). Further, the AM eyes of L. tarantula Linnaeus 1758 also differ from the other eyes in having muscle attachments, and therefore better mobility (Ortega-Escobar & Munoz-Cuevas 1999). This led Land & Barth (1992) to conclude that one function of the AM eyes may be to analyze stationary objects, since small movements prevent the neural image from adapting, as occurs in the secondary eyes. We observed a further distinction between the posterior and anterior eyes of L. leuckartii in the organization of the optic nerves. While each of the anterior eyes of L. leuckartii connects to one discrete nerve bundle, the posterior eyes exhibit multiple bundles (up to 35) exiting each eye. Multiple nerve bundles have previously been reported in another lycosid species. Researchers have found Lycosa tarentula fasciventris to have 20 nerve bundles exiting the PL eyes and 30 from the PM eyes (Kovoor et al. 1992). The function of multiple nerve bundles in the posterior eyes of L. leuckartii is not known. The presence of these discrete nerve bundles may be the result of developmental or functional differences and remains to be investigated. The posterior eyes appear to have partially overcome the trade-off between resolution and sensitivity by increasing in size relative to the carapace. A comparison of the inter- receptor angle of the PL with other species of lycosid suggests that L. leuckartii has a much better resolution than L. horrida Keyserling 1877 (1.5-2. 5°; Homann 1931), L. singoriensis Laxmann 1770 (= Trochosa singoriensis) (1. 7-2. 6°; Homann 1931) and two other species of Lycosa published in Homann (1931) (both 1.8°). Further, it appears that resolution for each eye of L. leuckartii is better than the corresponding eyes of that found in the closely related, nocturnal ctenid spider Cupiennius salei (Land & Barth 1992). The sensitivity of the lens was also high for L. leuckartii when compared with other species. The least sensitive of the CLEMENTE ET AL.— VISUAL SYSTEM OF A WOLF SPIDER eyes of L. leuckartii, the AM eyes, are comparable in /- numbers to nocturnally active web-building uloborids, whose y-number’s range from 0.88-1.70 (Opell & Ware 1987). The sensitivity of the secondary eyes of L. leuckartii even closely approximates that of the PM eyes of the nocturnal ogre-faced spider Deinopis siihrufa (/-number of AM eyes = 0.58: Blest «fe Land 1977). When estimates of receptor size are included (S-numbers), the sensitivity for the PL eyes of L. leuckartii appeared twice as good as that of the PL eyes of a lycosid species reported in Homann (1931) and Land (1985); however, the S-number for the PL eyes of L. leuckartii was about ten times less than the PL eyes of Cupieimius salei (S-number 147 pm^). This suggests that the visual system of L. leuckartii may be biased toward providing better resolution rather than sensitivity, though whether this translates into differences in performance remains to be investigated. There appear to be important differences in the visual systems between Australian lycosids and Palaeartic lycosids. One possible source of this variation may be prey capture strategies. While most of the Australian lycosid species studied appear to be burrowing, the Palaeartic species are vagrant, or build temporary webs (Murphy et al. 2006). The quality of vision in one particular genus, Pardosa, should be examined, since this group appears to be predominately vagrant and may therefore show more reliance on vision. Nevertheless, there are a number of vagrant and web-weaving lycosids in Australia. Also a distinctive granite-rock-inhabiting genus in southern Western Australia, which shelters under exfoliated slabs on the rock surface and hunts in a vagrant fashion, is currently being described by Framenau et al. (in press) as a new genus. It certainly invites further research regarding its ocular capability. ACKNOWLEDGMENTS We are grateful to the following people from the School of Animal Biology, University of Western Australia: Professor Lyn Beazley, for specialist information on optical nerves; Wally Gibb, for advice on collection of specimens; and Phil Runham, for help with experimental methods and general advice. We would like to dedicate this work in memory of Lisa Meldrum. She made a great contribution in terms of data, analysis, and writing, but it is her friendship that will be most sorely missed. Without Lisa there would have been a lot less laughter in the production of this work. LITERATURE CITED Blest, A.D. & M.F. Land. 1977. The physiological optics of Dinopis suhrufus L. Koch: a fish-lens in a spider. Proceedings of the Royal Society of London Series B 196:197-222. Bristowe, W.S. & G.H. Locket. 1926. The courtship of British lycosid spiders, and its probable significance. Journal of Zoology 96:317-347. Clemente, C.J., K.A. McMaster, L. Fox, L. Meldrum, B.Y. Main & T. Stewart. 2005. Visual acuity of the sheet-web building spider Badumna insignis (Araneae, Desidae). Journal of Arachnology 33:72^734. Dacke, M., T.A. Doan & D.C. O’Carroll. 2001. Polarized light detection in spiders. Journal of Experimental Biology 204: 2481-2490. Foelix, R.F. 1982. Biology of Spiders. Harvard University Press, Cambridge, Massachusetts. 405 Forster, L.M. 1979. Visual mechanisms of hunting behaviour in Trite planiceps, a jumping spider (Araneae: Salticidac). New Zealand Journal of Zoology 6:79-93. Forster, L.M. 1982. Visual communication in jumping spiders (Salticidae). Pp. 161-212. lu Spider Communication: Mechanisms and Ecological Significance (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Homann, H. 1928. Beitrage zur Physiologie der Spinnenaugen. 1. Untersuchungsmethoden. 11. Das Sehvermbgen der Salticiden. Zeitschrift fiir Vergleichende Physiologie 7:201-269. Homann, H. 1931. Beitrage zur Physiologie der Spinnenaugen. III. Das Sehvermogen der Lycosiden. Zeitschrift fur Vergleichende Physiologie 14:40-67. Homann, H. 1971. Die Augen der Araneae, Ontogenie und Bedeutung fur die Systematik (Chelicerata, Arachnida). Zoomor- phology 69:201-272. Jackson, R.R. 1977. Prey of the jumping spider Phidippus johnsoni (Araneae, Salticidae). Journal of Arachnology 5:145-149. Kaston, B.J. 1936. The senses involved in the courtship of some vagabond spiders. Entomologia Americana 16:97-167. Kovoor, P.J., A. Munoz Cuevas & J. Ortega Escobar. 1992. Le systeme visuel de Lycosa tarentida fasciiventris (Araneae, Lycosi- dae). I: Organisation des nerfs et des premiers ganglions optiques. Annales des Sciences Naturelles 13:25-36. Kovoor, P.J., A. Munoz Cuevas & J. Ortega Escobar. 1993. Microanatomy of the anterior median eye and its possible relation to polarized light reception in Lycosa tarentida (Araneae, Lycosidae). Bollettino di Zoologia 60:367-375. Land, M.F. 1969. Structure of the retinae of the principal eyes of jumping spiders (Salticidae: Dendryphantinae) in relation to visual optics. Journal of Experimental Biology 51:443-470. Land, M.F. 1981. Optics and vision in invertebrates. Pp. 471-592. In Handbook of Sensory Physiology Volume VII/6B. Vision in Invertebrates. B: Invertebrate Visual Centers and Behaviour 1 (H. Antrum, ed.). Springer Verlag, Berlin. Land, M.F. 1985. The morphology and optics of spider eyes. Pp. 53-78. In Neurobiology of Arachnids (F.G. Barth, ed.). Springer-Verlag, Berlin. Land, M.F. «& F.G. Barth. 1992. The quality of vision in the ctenid spider Cupieimius salei. Journal of Experimental Biology 164:227-242. Lizotte, R.S. & J.S. Rovner. 1988. Nocturnal capture of firellies by lycosid spiders: visual versus vibratory stimuli. Animal Behaviour 36:1809-1815. Magni, F. 1966. Analysis of polarized light in wolf-spiders. Pp. 171-186. In The functional organization of the compound eye (C.G. Bernhard, ed.). Pergamon Press, Oxford, UK. Magni, F., F. Papi, H.E. Savely & P. Tongiorgi. 1964. Research on the structure and physiology of the eyes of a lycosid spider. 11. The role of different pairs of eyes in astronomical orientation. Archives Italiennes de Biologic 102:123-136. Magni, F., F. Papi, H.E. Savely & P. Tongiorgi. 1965. Research on the structure and physiology of the eyes of a lycosid spider. 3. Electroretinographic responses to polarised light. Archives Ita- liennes de Biologic 103:146-158. Murphy, N.P., V.W. Framenau, S.C. Donnellan, M.S. Harvey, Y.C. Park & A.D. Austin. 2006. Phylogenetic reconstruction of the wolf spiders (Araneae; Lycosidae) using sequences from the 12S rRNA, 28S rRNA, and NADHl genes; Implications for classification, biogeography, and the evolution of web building behavior. Molecular Phylogenetics and Evolution 38:583-602. Opell, B.D. 1988. Ocular changes accompanying eye loss in the spider family Uloboridae. Journal of Morphology 196:1 19-126. Opell, B.D. & A.D. Ware. 1987. Changes in visual fields associated with web reduction in the spider family Uloboridae. Journal of Morphology 192:87-100. 406 THE JOURNAL OF ARACHNOLOGY Ortega-Escobar, J. 2006. Role of the anterior lateral eyes of the wolf spider Lycosa tarentula (Araneae, Lycosidae) during path integra- tion. Journal of Arachnology 34:51-61. Ortega-Escobar, J. & A. Munoz-Cuevas. 1999. Anterior median eyes of Lycosa tarentula (Araneae, Lycosidae) detect polarized light: behavioral experiments and electroretinographic analysis. Journal of Arachnology 27:663-671. Robinson, M.H. & B. Robinson. 1971. The predatory behavior of the ogre-faced spider Dinopis longipes F. Cambridge (Araneae: Dinopidae). American Midland Naturalist 85:85-96. Rovner, J.S. 1993. Visually mediated responses in the lycosid spider Rahidosa rahida: the roles of different pairs of eyes. Memoirs of the Queensland Museum 33:635-638. Rovner, J.S. 1996. Conspecific interactions in the lycosid spider Rahidosa rahida: the roles of different senses. Journal of Arachnology 24:16-23. Schmid, A. 1998. Different functions of different eye types in the spider Cupienniiis salei. Journal of Experimental Biology 201:221-225. Uehara, A., Y. Toh & H. Tateda. 1978. Fine structure of the eyes of orb-weavers, Argiope amoena L. Koch (Aranea: Argiopidae). Cell and Tissue Research 186:435^52. Uehara, A. & K. Uehara. 1996. Efferent fibers and the posteromedial eye of the liphistiid spider Heptathela kimurai (Araneae: Liphis- tiomorphae). Journal of Experimental Zoology 275:331-338. Uetz, G.W. & G.E. Stratton. 1982. Acoustic communication and reproductive isolation in spiders. Pp. 123-158. In Spider Commu- nication: Mechanisms and Ecological Significance (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Whitcomb, W.H. & R. Eason. 1965. The mating behavior of Peucetia viridans (Araneida: Oxyopidae). Florida Entomologist 48:163-167. Williams, D.S. & P. McIntyre. 1980. The principal eyes of a jumping spider have a telephoto component. Nature 288:578-580. Zwicky, K.T. 1968. Innervation and pharmacology of the heart of Urodacus, a scorpion. Comparative Biochemistry and Physiology 24:799-808. Manuscript received 30 October 2009, revised 21 May 2010. 2010. The Journal of Arachnology 38:407^14 Possible functional significance of spigot placement on the spinnerets of spiders William G. Eberhard: Smithsonian Tropical Research Institute, and Escuela de Biologia, Universidad de Costa Rica, Ciudad Universitaria, Costa Rica. E-mail: william.eberhard@gmail.com Abstract. This paper discusses the possible functional significance of the locations of the spigots of different types of silk gland on the different spinnerets of spiders. Deductions are based on recognition that some types of line are initiated by being attached to the dragline, that there is an anterior-posterior asymmetry in how such lines can be initiated, and that spigot location also affects the possibility of attaching lines to the substrate. Possible explanations are given for several morphological details, including the anterior location of the dragline, piriform and cribellum spigots, planar arrays of piriform and cribellum spigots, and posterior location of aciniform spigots. I argue that piriform gland products are not used to attach egg sac lines to each other, that sticky wrapping lines are initiated in theridiids and pholcids by attaching them to draglines and that lines from both aciniform and cylindriform glands are laid along with liquid that renders them sticky. The possible role of phylogenetic inertia in determining spigot locations is discussed. Further work is needed to determine whether termination of lines and accessibility of spigots for cleaning also influence their positions. Keywords: Silk, silk gland, phylogenetics inertia Spinneret morphology provides many useful taxonomic characters in spiders, and the distributions and forms of the spigots of different silk glands have been described for many species (summaries in Coddington 1989; Platnick et al. 1991; Agnarsson 2004; Griswold et al. 2005). Surprisingly however, there has been little discussion of the possible functional significance of the locations of spigots on spinnerets. Perhaps the most striking exception is the onchyroceratid Ochyrocera cachote Hormiga, Alvarez-Padilia & Benjamin 2007, which builds small domed sheet webs containing sectors with large numbers of precisely parallel lines (Hormiga et al. (2007). The posterior lateral spinnerets of O. chachote have an unusual row of tightly spaced aciniform spigots, and the similarity between the length of this row and the number of spigots with the width of the arrays of parallel lines and the numbers of parallel lines in the swaths (about 20) leave little doubt that each swath is produced during a single pass of the spider’s spinnerets (Hormiga et al. 2007). Similar arrays of many parallel lines also occur in the webs of another Ochyrocera species in Costa Rica (G. Barrantes & W. Eberhard unpubl. results). A second case in which the possible functional significance of spigot positions on spinnerets has been discussed is the tight physical association between the spigots of the aggregate and the flagelliform glands in araneoid orb weavers, allowing the spider to coat the flagelliform line with sticky material from the aggregate gland as the line emerges from its spigot (Coddington 1987; Blackledge et al. 2009). The present note combines direct observations and video recordings (30 frames/s) of the behavior of mature females of large araneoid species, data from the extensive literature on spinneret morphology, and data from the much less extensive literatures on how different types of lines are initiated and on their morphology (especially in the SEM) to propose possible functional explanations for the positions of a number of types of spigots in spiders. I argue that the sites of different spigots relative to each other influence the spider’s ability to initiate and to coordinate the production of different types of lines, and that selection on these abilities may explain why particular spigots are located at particular sites. The arguments are not all complete, and the aim is to initiate discussions rather than to provide exhaustive, final answers. General considerations regarding how lines are initiated and fastened. — Spigot placement probably affects how a spider can initiate lines. As is well known, spiders cannot eject silk lines, but must have them pulled from their bodies (Witt et al. 1968; Foelix 1996). Three different mechanisms have been proposed for line initiation. The best known, “direct contact” initiation technique occurs when the spider presses its spinnerets against a substrate and then pulls away (e.g., Kullmann 1975). If the spinnerets are held in appropriate positions, if the spigots are at appropriate sites on the spinnerets to bring their tips into contact the substrate, and if there is liquid silk at the tips of the spigots (they are “wet”) (presumably due to abdominal pressure - Wilson 1962a), then this liquid will adhere to the substrate and the spider can initiate a new line when it pulls its spinnerets away. Similarly, lines could be initiated by direct contact between different spinnerets, with or without another line between them (see below). A second way of initiating lines, “dragline initiation”, was observed when Nephila clavipes (Linnaeus 1767) and Argiope argentata (Fabricius 1775) wrapped prey. For instance, when a large female N. clavipes begins to wrap a prey, the lines in the swath of wrapping lines (presumably from her aciniform gland spigots) are attached to her dragline (Fig. 1). The spider wraps the prey by snagging the wrapping lines with her legs IV and pulling out more silk by extending them toward the prey to press the wrapping lines against it. A second context in which dragline initiation occurs in a variety of species is initiation of airborne lines, when the distal ends of airborne lines are attached to the spider’s dragline as it descends (Eberhard 1987) (in this case, the glandular origins of the lines are unknown). Initiation of these wrapping and airborne lines presumably occurs when the spider applies the “wet” tips of the spigots that produce these lines to a dragline while the dragline is being pulled as the spider moves. This dragline initiation technique is only feasible if the spigots of different glands are aligned on the spinnerets in 407 408 THE JOURNAL OF ARACHNOLOGY Figure 1. — A swath of aciniform lines is attached to the dragline of a mature female Nephilci clavipes as she descends, just before she initiates prey wrapping (drawn from a video recording). certain ways. Spiders almost always move forward rather than backward while spinning draglines, so lines from spigots on the “downstream”, posterior median and posterior lateral spinnerets (e.g., the aciniform spigots in araneoids) can be initiated using lines from spigots on the “upstream”, anterior lateral spinnerets (e.g, dragline or major ampullate spigots). Movements of more posterior spinnerets in an anterior direction, and manipulations of lines with the legs can relax these constraints somewhat (see description of Nephila behavior below). The short lengths of most spinnerets, however, suggest the general rule that spigots that are located farther posterior on the spider’s body (more downstream) can initiate lines by touching their tips to lines from more anterior (upstream) spigots, but not vice versa. A final possible mechanism for initiating lines, “clapping initiation”, involves spreading movements of the spinnerets with respect to each other (Blackwell cited in McCook 1889; Nielsen 1931; Eberhard 1987). For example, the spider could clap or rub pairs of spinnerets together, and then pull them apart. The spider could then either pull further silk with its legs, or use friction with the air to elongate the short lines between the spinnerets. Although there are reports that spinnerets are spread widely when some lines (e.g., airborne lines) are being initiated (Blackwell in McCook 1889; Eberhard 1987), I know of no confirmed demonstration of clapping initiation. One further aspect of “upstream - downstream” locations concerns the use of one type of silk to fasten lines of another type of silk to the substrate or to other lines. The spigots of the fastening silk need to be either at the same level or downstream of the spigots that produce the lines that are being fastened. Possible functional significance of spigot placements. — The placements of spigots on the spinnerets could alter the feasibility of all three possible types of initiation, and their observed positions may possibly be explained in terms of functional consequences. I will discuss possible functions according to the probable glandular origin of the lines. Ampullate gland spigots: As far as I know, the only technique that has been observed in dragline initiation (which is relatively rarely observed in the many species that produce draglines more or less continuously) is direct contact (e.g., Kullmann 1975 on an unspecified species; W. Eberhard unpubl. results on Micrathena duodecimspinosa (O. Pickard- Cambridge 1890) when they initiate the dragline after finishing construction of the sticky spiral). The rather basal position of the major ampullate spigots on the anterior lateral spinnerets of many species (Coddington 1989) (Fig. 2) probably obliges the spider to either spread these spinnerets widely, or to insert the substrate (e.g., another line, the spider’s tarsus) deep between them. I know of no direct observations, however, of this detail. Kullmann (1975) noted that dragline initiation can occur by contact even when the spider is anesthetized. Aciniform gland spigots: As noted above, initiation of aciniform wrapping lines in araneoids apparently occurs using the dragline initiation technique (Fig. 1). The downstream placement of the aciniform spigots (on the posterior median and lateral spinnerets) (Fig. 2) in araneomorph spiders in general (Griswold et al. 2005) with respect to the major ampullate spigots may function to allow initiation of acini- form wrapping lines by attachment to draglines. The downstream placement of aciniform spigots in mygalomorph spiders and presumably on the common ancestor of all spiders must have a different explanation, however, as these spiders lack dragline silk (Palmer et al. 1982; Palmer 1985; Blackledge et al. 2009). The arrangements of the aciniform spigots of some mygalomorphs in a line along the long axis of the long posterior lateral spinneret (Peters 1967) is apparently an ancient trait (Vollrath & Selden 2007). Similar linear arrays of aciniform spigots also occur in other groups with long posterior lateral spinnerets, such as agelenids and hersiliids (Peters 1967). They probably facilitate both the production of sheets of fine lines, rather than of cables of lines, and the placement of these swaths of lines using movements of the highly mobile spinnerets (Fig. 3). Sheets rather than cables are useful to these spiders, which line their retreats with sheets of silk, build sheet webs to capture prey and wrap their prey in wide bands of silk (Coyle 1986; Barrantes & Eberhard 2007). The aciniform lines of both mygalomorphs and araneo- morph spiders adhere at least weakly both to each other and to the substrate, and SEM and light microscope images indicate that they have small amounts of liquid associated with them when they are produced (Kullmann 1975 on ctenizid and diplurid mygalomorphs; Weng et al. 2006 on an uloborid and an araneid; W. Eberhard unpublished on the araneid A. argentata). This stickiness of aciniform lines helps make functional sense of some aspects of spinneret morphology. Mygalomorphs are able to glue lines together and to the EBERHARD— FUNCTIONAL SIGNIFICANCE OF SPIGOT LOCATIONS 409 Posterior median spinneret Figure 2. — Basic diagram of the locations of spigots on the spinnerets of araneoid spiders (after Coddington 1989). Figure 3. — The complex muscles associated with the spinnerets of an agelenid (above) and an araneid (below) indicate that these spiders are able to make diverse, finely controlled spinneret movements (above from Peters 1967; below from Wilson 1962b). 410 THE JOURNAL OF ARACHNOLOGY substrate, despite that fact that they lack piriform glands, whose products are used to fasten lines to other objects in araneomorph spiders (Schiitt 1996; Foelix 1996; below). The aciniform spigots of araneomorphs are downstream from the piriform spigots, so their aciniform lines probably cannot be easily attached using piriform silk. The aciniform lines also vary substantially in diameter, in both mygalomorphs (Kullmann 1975 on a diplurid; Stern & Kullmann 1981 on a ctenizid), and araneomorphs (Kullmann 1975 on an araneid; Weng et al. 2006 on an uloborid and an araneid; W. Eberhard unpubl. results on the araneid A. argentata). Presumably these differences are associated with the differences in aciniform spigot diameters seen in many species (Griswold et al. 2005), but nothing appears to be known regarding the possible functional significance of these differences, or the possibility that silk composition varies in lines with different diameters. In Ulobonis there are two kinds of cells in all aciniform glands (Kovoor 1977). Piriform gland spigots: Small masses (“attachment discs”) of piriform gland products are used to attach draglines to the substrate or to other lines. They are apparently always initiated using the direct contact technique. Piriform lines are initiated either when the spider places a silk line against her spinnerets or presses her spinnerets against a substrate such as a leaf Production of an attachment disc generally involves a short pause in spinning other lines, and in at least some cases the spinnerets rub rapidly against each other or against the substrate when an attachment is made (Schiitt 1996; W. Eberhard unpubl. results on various araneid, nephilid, tetragnathid, and theridiid species). The looped forms of individual piriform lines in an attachment disc (e.g., Foelix 1996) probably result from these spinneret movements, but this behavior has never been studied in detail. Piriform lines are generally terminated almost immediately, when the spider pulls her abdomen away from contact with the substrate. Use of the contact technique for initiation means that the piriform gland spigots do not need to be “downstream” of other spigots. In accord with this, piriform spigots consistently occur at relatively “upstream” sites, on the anterior lateral spinnerets (Fig. 2). Two additional details regarding the positions of piriform spigots on the anterior lateral spinnerets may be functionally important. The spigots are consistently at or near the tips rather than along the sides or at the bases of the spinnerets; and the surface on which the spigots are present is often relatively flat and tends to slope downward medially (e.g., Platnick et al. 1991; Agnarsson 2004; Griswold et al. 2005). Both the placement at the tip and the slanting flat field may serve to facilitate simultaneous contact between numer- ous piriform spigots and a flat substrate when the spinnerets are spread. Perhaps the different degrees of slant in different species (e.g., Griswold et al. 2005) are associated with different degrees to which the spinnerets are spread when producing attachment discs. The piriform spigots are also generally well placed to apply their products to the dragline emerging from the major ampullate spigots on these same spinnerets, thus allowing the spider to attach its dragline to the substrate. In some species the piriform spigots physically surround the major ampullate spigots (Fig. 2). Microscopic examination of piriform attachment discs shows that they include liquid components, and also possibly more than one type of line (Kullmann 1975 and Stern & Kullmann 1981 on an oecobiid and an araneid; Schiitt 1996 on a linyphiid and an araneid; Foelix 1996 on a thomisid). It is apparently not known whether different piriform spigots produce different products. Judging from the relatively uniform coating of apparent liquid on the piriform lines of the linyphiid Drapestica socialis (Sundevall 1833) (Schiitt 1996), piriform lines in this species may emerge from their spigots with a liquid coating, rather than having liquid applied later. The description of Griswold et al. (2005: 59) “These spigots extrude the glue that attaches silken lines” is correct in a general sense, but does not include the thread-like products of these spigots, which are drawn from rather than “extruded” by these spigots. The behavior of the spinnerets associated with the production of piriform silk was studied in video recordings of a large, slow-moving spider (an adult female N. clavipes) as she attached draglines to other non-sticky lines at the hub, and sticky spiral lines to radii. Spinneret behavior was highly stereotyped. In both types of attachment, the spider held the segment of the non-sticky line to which the attachment was to be made between her tarsi III and IV and placed it (using movements of these legs and of her abdomen) deep in the cleft between her anterior lateral spinnerets (the radius was approximately halfway down the length of her anterior lateral spinnerets). The radius often, though not always, passed over the tips of her posterior lateral spinnerets. Both the basal and the distal segments of her anterior lateral spinnerets were flexed medially, and the tips of the spinnerets rubbed rapidly back and forth against each other briefly, with one spinneret moving anteriorly while the other moved posteriorly. Probably contact occurred on the surfaces where the piriform spigots are located (Kuntner et al. 2008). When attaching the dragline, each anterior lateral spinneret rubbed several times against the other (up to three complete cycles in about 0.77 s). Such movements may serve both to initiate piriform lines and to surround the dragline and the other line with piriform lines and glue; this could result in an increase in the surface area of these lines that is contacted by piriform products, presumably increasing the strength of the attachment that is formed. When attaching the sticky spiral, the spider made only a single forward-backward rubbing movement with each spin- neret (lasting about 0.06-0.09 s). Her posterior lateral spinnerets, which bear the spigots for the sticky spiral line and the glue that covers it, were flexed forward. This flexion occurred 0.03-0.06 s before the radius was brought into contact with her spinnerets and was maintained while the anterior lateral spinnerets rubbed against each other. Mean- while her leg ilV pushed the sticky line ventrally and somewhat anteriorly, and the line appeared to make contact with the anterior lateral spinnerets during their rubbing movements. The anterior movements of the posterior lateral spinnerets, in combination with the fact that the spider pushed the new sticky line ventrally and anteriorly with one of her legs IV, thus brought the sticky line far enough forward that the piriform spigots were able to apply their silk to it during the rubbing movements. Finally the anterior lateral spinnerets spread apart and (often about 0.03 s later) the posterior lateral spinnerets spread and moved posteriorly as the spider pulled EBERHARD— FUNCTIONAL SIGNIFICANCE OF SPIGOT LOCATIONS 411 her abdomen away from the radius, leaving the finished attachment. It is possible that these piriform products differ from others, because there are other data which suggest that the attachments of sticky spiral lines to radii in araneoid orbs may not be made from “typical” piriform gland products (Kullmann 1975 on differences in silk morphology in SEM images; Eberhard 1976 on differences in physical properties; Tillinghast et al. 1981 on differences in chemical properties). Cribellum gland spigots: A second type of line that is also both initiated and then pulled by an external agent is cribellum silk; multiple fibers are pulled from the cribellum with brushing movements of the comb of setae (the calamistrum) on the spider’s leg IV. There is no need for the spigots of lines that are pulled by the spider’s legs to be downstream of other spigots, and the cribellum is indeed upstream of all other spigots. In addition, the spinnerets on which cribellum spigots occur (the anterior medians) have been modified to form flat plates whose angles can apparently be modified by special muscles (Peters 1967), a design that is appropriate to allow the linear calamistrum to snag lines from the entire array with a single pass. Pseudo jlagelliform gland spigots: The placement of the spigots of the pseudoflagelliform glands is also functionally logical. These glands are thought to produce the straight baselines associated with the mat of cribellum lines in the sticky spiral of uloborids. The pseudoflagelliform spigots are on the tips of the posterior lateral spinnerets (Coddington 1987), appropriately distant from the cribellum to avoid being snagged by the calamistrum (the pseudocribellar line, which is much shorter than the cribellum lines, is presumably pulled by the walking movement of the spider rather than by the combing action of the calamistrum). The sticky spiral of the uloborid Uloborus diversus is initiated when the spider touches her spinnerets to a radius (Eberhard 1972), so the direct contact technique may be used to initiate the pseudoflagelli- form lines. Cylindriform gland spigots: Cylindriform gland lines are used to make egg sacs. They seem to be slightly sticky when they emerge, because some lines in egg sacs are covered with liquid (Kullmann 1975 on an araneid, a mimetid and a deinopid; Stern & Kullmann 1981 on a theridiid), and they very frequently adhere at least weakly to each other, even though the spider only briefly dabs her abdomen to the egg sac during construction; the dabs are not accompanied by short pauses, as when piriform discs are produced (Kullmann 1961; Gheysens et al. 2005; Moya et al. 2010). It thus appears that piriform silk is not used to hold sticky cylindrical gland lines together; this is consistent with morphology, as the cylindrical gland spigots are downstream with respect to the piriform gland spigots (Fig. 2) Flagelliform and aggregate gland spigots in theridiids: The upstream placement of dragline spigots probably has still another important functional consequence in theridiid spiders, which use flagelliform and aggregate gland products to wrap and subdue prey. Initiation of these wrapping lines occurs as the spider moves toward the prey (Barrantes & Eberhard 2007), and it seems very likely (though initiation has never to my knowledge been directly observed) that these wrapping lines are initiated by being attached to the emerging trail line. The downstream placement of the flagelliform and aggregate gland spigots in theridiids (on the posterior lateral spinnerets) with respect to the ampullate gland spigots (on the anterior lateral spinnerets) (Coddington 1989; Agnarsson 2004) makes dragline initiation feasible. The silk used in the later stages of theridiid wrapping attacks tends to be dry rather than wet (Barrantes & Eberhard 2007), presumably as aciniform lines replace wet sticky lines. Initiation of these presumed aciniform lines does not involve any obvious change in the wrapping movements of the spider’s legs or abdomen. The downstream position of aciniform spigots (posterior median and lateral spinnerets) with respect to major and minor ampullate spigots, and the fact that they are also downstream or parallel in position with respect to the flagelliform spigots on the basal portions of the posterior lateral spinnerets (Agnarsson 2004), also makes it feasible that the aciniform lines are initiated by being attached to either the dragline (if it is still being produced) or to the flagelliform line. In some of the orbicularion groups that no longer spin orbs, the tightness of the physical association between the flagelli- form and aggregate spigots on the posterior lateral spinnerets is less consistent (e.g., Griswold et al. 1998, Agnarsson 2004). The possible the reasons for this dissociation are not clear. Pholcid spigots: The spinnerets of pholcids have only small numbers of spigots compared with those of araneoids (as few as 8, with a maximum of 20 - B. Huber, pers. comm.; contrast this with >1300 spigots in Araneus diadematus Clerck 1757- Foelix 1996), and some glands and their spigots have been difficult to homologize with those of other spiders (Kovoor 1986; Platnick et al. 1991). The enormously enlarged tip of one spigot (the spigot for gland “B” of Millot) at the tip of the anterior lateral spinneret (Platnick et al. 1991; Huber 2000) appears to produce sticky liquid that is used both in wrapping prey (Kirchner & Opderbeck 1990; Eberhard 1992), and in making the puddle of liquid that attaches lines to the substrate (Schiitt 1996). This spigot lies immediately adjacent to the spigot that probably produces the dragline (major ampullate gland; gland “C” of Millot) (Kovoor 1986; Platnick et al. 1991). Pholcids resemble theridiids in rapidly initiating sticky silk production at the beginning of wrapping attacks and then later using non-viscous silk to wrap the prey (Kirchner & Opderbeck 1990; Eberhard 1992; Barrantes & Eberhard 2007). As in theridiids, there is no perceptible pause associated with initiation of sticky silk production as the spider runs to the prey and begins to wrap it. This rapid initiation of sticky wrapping silk probably occurs when spigot B material is pulled out by (or poured onto?) the dragline while the dragline is produced as the spider moves toward the prey, in much the same way that aggregate gland silk is added to flagelliform gland silk in araneids (above). In Pholcus phalangioides (Fuesslin 1775) the spider applies up to four lines of silk at once to the prey (Kirchner & Opderbeck 1990), so lines from at least one additional set of spigots in addition to the major ampullates must be involved, perhaps the spigots for gland A on the posterior median spinnerets (these are thought to correspond to the minor ampullate glands) (Kovoor 1986; Platnick et al. 1991). It would be feasible (though certainly not demonstrated at the moment) for the spider to use dragline initiation for these additional minor ampullate lines during the latter stages of prey wrapping by touching the A spigots on the 412 THE JOURNAL OF ARACHNOLOGY median posterior spinnerets to the dragline after it emerges from spigots C on the anterior lateral spinnerets. Pholcids use small pools of what is probably this same adhesive silk from spigot B, combined with other lines (presumably from other, smaller wide-mouthed spigots on the anterior lateral spinnerets), to fasten lines to each other and to the substrate (Kirchner 1986; Schiitt 1996) (e.g., the typical function of piriform glands). The positions of the B spigots at the tips of the anterior lateral spinnerets are also appropriate for this function. It seems likely that the wrapping function for these glands in pholcids was derived from the attachment function, and indeed B glands have chemical characteristics reminiscent of piriform glands (Kovoor 1986; Platnick et al. 1991). If this derivation is correct, the position of the piriform spigots adjacent to the major ampullate gland spigots may have represented a preadaptation for the subsequent use of this sticky silk for wrapping prey. Presumably one piriform spigot became oversized as the gland product became more liquid, possibly in association with its use in prey wrapping. Fine airborne lines: The glandular origins of very fine bridge or balloon lines are not known, but the arguments made here suggests that their spigots are located downstream of the dragline spigots, and perhaps upstream of the spigots of the thicker lines that the spider apparently attaches to these fine lines while the fine lines are being produced (Eberhard 1987). Direct observations of the spinnerets of the tetragnathid Leucauge mariana (Taczanowski 1881) suggested that the fine airborne lines originate on the posterior median or posterior lateral spinnerets (Eberhard 1987). DISCUSSION Implications. — The data above show that the consistent sites at which the spigots of different glands are placed on different spinnerets have functionally reasonable “upstream-down- stream” explanations. It is also possible, however, that these explanations for spigot placement are simply an example of historical constraint. It makes functional sense, for instance, that the major ampullate spigots are not downstream of others, due to their possibly central role in initiating lines from other glands. But the major ampullate spigots apparently evolved only once (Griswold et al. 2005), and the lack of variation of major ampullate spigot placement across spiders indicates that their placements in different species probably represents (at least in some senses) only a single evolutionary event. This skepticism regarding the functional interpretations proposed here must be tempered, however, by the fact that the “upstream-downstream” arguments can explain not only why these and other spigots are located where they are in the first place (clearly, spigots are not randomly scattered over the spinnerets); they can also explain why they have not moved subsequently. The possibility of evolutionary flexibility in spigot placement is demonstrated by the fact that aciniform gland spigots were apparently regained twice independently after having been lost from the posterior median spinneret in Pimoa (Hormiga 1994). Additional predictions derived from the arguments presented here could offer further tests. For instance, there should be a correlation between the positions and movements of the anterior lateral spinnerets when attachments are made and the site and slant of the planar field on which the piriform spigots are located. It is also important to note that the kinds of functional arguments made here cannot yet explain some spigot placements and thus also represent possible areas of future research. I do not understand, for instance, how the calamistrum on the spider’s leg IV can snag lines from the paracribellar spigots on the posterior median spinnerets of a species like Filistata imidiatrix (Forsska! 1775). In this species the paracribellar lines are highly curled and presumably pulled out by the brushing action of the calamistrum (Griswold et al. 2005), but the posterior median spinnerets are shorter than the anterior lateral spinnerets, thus apparently making it difficult for the calamistrum to pull lines from the paraflagelliform spigots (Griswold et al. 2005) (perhaps the spider can protrude the posterior median spinnerets?). The functional significance of the substantial differences in the forms of the calamistra of different species has also, to my knowledge, never been discussed or related to the differences in the form of the cribellum. Many other smaller morphological details are still myste- rious, such as the “T” shaped array of piriform spigots on the anterior lateral spinnerets of Mecynogea lemniscata (Walck- enaer 1842) and Cyrtophora citricola (Forsskal 1775) (Cod- dington 1987), the nearly complete loss of piriform spigots in immature but not adult Eriauchenius {=Archaea) workmani O. Pickard-Cambridge 1881 (Griswold et al. 2005), and the loss and subsequent independent recovery of aciniform spigots on the posterior median spinnerets of some species of Pimoa (Hormiga 1994). It is not clear why there is variability in the tightness of the physical association between the flagelliform and aggregate spigots in some orbicularions that have lost orb webs (e.g., Agnarsson 2004). The function of the “small gland” spigots on the anterior lateral spinnerets of pholcids (Kovoor 1986), which were presumed by Platnick et al. (1991) and in the discussion above to represent piriform spigots that are used to fasten lines to the substrate and other lines, needs to be confirmed. The origin of the screw lines (Kirchner & Opderbeck 1990) in pholcid webs is apparently unknown. Future directions. — The focus here has been on morphology, but understanding the functional significance of morphology depends on combining it with behavioral data. The behavioral capabilities of spider spinnerets, however, are as yet nearly completely unstudied. The spinnerets of many species are relatively short and thus difficult to observe, but they are segmented, and are equipped with well-developed musculature (Fig. 3), suggesting that they may be capable of substantial subtlety in their movements. The observations of Nephila described above document brisk, highly coordinated spinneret movements. Some mygalomorph and labidognath spiders can move at least their posterior lateral spinnerets with certain dexterity, including asymmetrically raising one and lowering the other during prey wrapping) (Barrantes & Eberhard 2007). Foelix (1996) mentions lifting, lowering, twisting and spread- ing movements (species not specified). Perhaps, even though the context is somewhat artificial, study of spinneret movements during forcible silking can help establish the behavioral capabilities of spinnerets. Craig (2003) stated (though with no evidence) that spinneret movements have EBERHARD— FUNCTIONAL SIGNIFICANCE OF SPIGOT LOCATIONS 413 played an important role in giving spiders great flexibility with respect to the types and character of the threads they spin. Perhaps further study of spinneret movements will help to resolve some of these problems. Do spinnerets actually “clap” together medially to initiate airborne lines, as proposed by Blackwall (in McCook 1889), Nielsen (1931) and Eberhard (1987)? Are such claps always symmetrical (as appeared to be the case in N. ckmpes), or is it possible for the piriform field of one anterior lateral spinneret to move medially to press on the lower interior face of the other anterior lateral spinneret and thus press on the line emerging from the major ampullate spigot? When spiders evolved to attach their draglines to other lines with piriform discs, their anterior lateral spinnerets may have already been capable of opposable movements that permitted them to “grasp” the other line to which the attachment was to be made, and thus apply piriform silk precisely to the other line. Perhaps araneomorphs evolved their relatively shorter spinnerets in association with the evolution of more types of silk glands, due to the advantage of having the spigots of different glands close to each other to facilitate coordination of production of different types of lines (B. Huber, pers. comm.). It will be important in future work to keep in mind that the spinneret positions in taxonomic papers are somewhat unnatural, because taxonomists routinely spread the spinner- ets to make the locations of different spigots easier to observe (Coddington 1989). In general, the spigots on different spinnerets are probably closer together in life than would be suggested by figures in taxonomic works. In attempting to think about “upstream - downstream” positions, it is also necessary to take into account cases in which the spider uses her leg to push a line anteriorly, as in sticky spiral production by N. clavipes. It will also be important to avoid typology, and keep in mind the possible consequences for spinneret use of variations in behavior; for instance, the different ways the spider grasps lines to which they are going to attach in the theridiids. Achaearanea tesselata (Keyserling 1884) (Jorger & Eberhard 2006) and Tidarren sisyphoides (Walckenaer 1842) (Madrigal-Brenes & Barrantes 2009) seem likely to affect how their spinnerets contact these lines. Another topic that needs further work and that may provide understanding of spigot placement concerns the mechanisms that spiders use to terminate lines. Piriform initiation and termination is sometimes repeated literally hundreds of times very rapidly during the construction of a single orb. For instance, one mature female M. diiodecimspinosa made 832 attachments of her hub spiral and temporary spiral to radii in the space of 428 s, or an average of 1.9 attachments/s; the time during which the spider’s spinnerets were in contact with web lines (and thus the time during which piriform silk could be deposited at each attachment) was on the order of one or two tenths of a second/attachment. In other words, M. duodecim- spinosa turns on pyriform silk production for only about a tenth of a second about twice a second, making hundreds of attachments in rapid succession. This is truly an impressive feat of coordination! What little is known concerning termination of silk lines suggests (at least in the ampullate glands, which have long ducts) that termination generally occurs when there is a lack of liquid silk in the duct leading to the spigot, and the line breaks there as it is being pulled (Wilson 1962b; Work 1977). Presumably liquid silk then needs to be pushed to the tip of the spigot for the line to be reinitiated. It is not clear whether this termination mechanism also occurs in other glands such as the piriform and aciniform glands, which have much shorter ducts. So little is known at present that it is desirable to keep an open mind regarding whether spigot placement is also functionally related to termination. Spigot placement could also possibly influence the spider’s ability to clean its spinnerets after lines are terminated. For instance, when an adult female N. clavipes finishes wrapping a prey, she often “scrubs” her spinnerets together, scrapes their surfaces with the tarsus of her leg IV and then pulls the tarsus away, sometimes repeatedly. ACKNOWLEDGMENTS I thank H.W. Levi for identifying spiders, Ingi Agnarsson, Gilbert Barrantes, Bernhard Huber and two anonymous referees for helpful comments on preliminary drafts, and the Smithsonian Tropical Research Institute and the Vicerrectoria de Investigacion of the Universidad de Costa Rica for financial support. LITERATURE CITED Agnarsson, 1. 2004. Morphological phylogeny of cobweb spiders and their relatives (Araneae, Araneoidea, Theridiidae). Zoological Journal of the Linnaean Society 141:447-626. Barrantes, G. & W.G. Eberhard. 2007. Evolution of prey wrapping behavior in spiders. Journal of Natural History 41:1631-1658. Blackledge, T., N. Scharff, J.A. Coddington, T. Szuts, J.W. Wenzel, C. Hayashi & 1. Agnarsson. 2009. Reconstructing web evolution and spider diversification in the molecular era. Proceedings of the National Academy of Sciences, USA 106:5229-5234. Coddington, J.A. 1989. Spinneret silk spigot morphology: evidence for the monophyly of orbweaving spiders, Cyrtophorinae (Aranei- dae), and the group Theridiidae plus Nesticidae. Journal of Arachnology 17:71-95. Coyle, F.A. 1986. The role of silk in prey capture by nonaraneomorph spiders. Pp. 269-305. In Spiders: Webs, Behavior and Evolution. (W.A. Shear, ed.). Stanford University Press, Stanford, California. Craig, C.L. 2003. Spiderwebs and Silk. Oxford University Press, New York. Eberhard, W.G. 1972. The web of Ulohorus diversus (Araneae: Uloboridae). Journal of Zoology, London 166:417-465. Eberhard, W.G. 1976. Physical properties of sticky spirals and their connections: sliding connections in orb webs. Journal of Natural History 10:48 1M88. Eberhard, W.G. 1987. How spiders initiate airborne lines. Journal of Arachnology 15:1-10. Eberhard, W.G. 1992. Notes on the ecology and behavior of Physocyclus glohosus (Araneae, Pholcidae). Bulletin of the British Arachnological Society 9:38-42. Foelix, R. 1996. Biology of Spiders. Second edition. Oxford University Press, New York. Gheysens, T., L. Beladjal, K. Gellynck, E. Van Nimmen, L. Van Langenhove & J. Mertens. 2005. Egg sac structures of Zygiella x- notata (Arachnida, Araneidae). Journal of Arachnology 33: 549-557. Griswold, C.E., J.A. Coddington, G. Hormiga & N. Scharff 1998. Phylogeny of the orb-web building spiders (Araneae, Orbiculariae: Deinopoidea, Araneoidea). Zoological Journal of the Linnean Society 123.T-99. Griswold, C.E., M.J. Ramirez, J.A. Coddington & N.I. Platnick. 2005. Atlas of phylogenetic data for entelegyne spiders (Araneae: 414 THE JOURNAL OF ARACHNOLOGY Araneomorphae: Entelegynae) with comments on their phylogeny. Proceedings of the California Academy of Sciences Series 4, 56(Supplement II); 1-324. Hormiga, G., F. Alvarez-Padilla & S.P. Benjamin. 2007. First records of extant Hispaniolan spiders of the families Mysmenidae, Symphytognathidae, and Ochyroceratidae (Araneae), including a new species of Ochyrocera. American Museum Novitates 3577:1-21. Huber, B. 2000. New World pholcid spiders (Araneae; Pholcidae): a revision at generic level. Bulletin of the American Museum of Natural History 254:1-348. Jorger, K. & W.G. Eberhard. 2006. Web construction and modification by Achaearanea tessekita (Araneae: Theridiidae). Journal of Arachnology 34:511-523. Kirchner, W. 1986. Das Netz der Zitterspinne (Pholcus phalcmgioides Fuesslin) (Araneae: Pholcidae). Zoologischer Anzeiger 216: 151-169. Kirchner, W. & M. Opderbeck. 1990. Beuteerwerb, Giftwirkung und Nahrungsaufnahme bei der Zitterspinne Pholcus phcilangioides (Araneae, Pholcidae). Abhand Verhandlungen des Naturwis- senschaftlichen Vereins in Hamburg (NF) 31/32:15^5. Kovoor, J. 1977. L’appareil sericigene dans le genre Ulohonis Latr. (Araneae, Uloboridae). I. Anatomie. Revue Arachnologique 1:89-102. Kovoor, J. 1986. Affinites de quelques Pholcidae (Araneae) decelables d’apres les caracteres de I’appareil sericigene. Memoires de la Societe Entomologique de Belgique 33:111-118. Kullmann, E. 1961. Der Eierkokonbau von Cyrtophora citricola Forskal (Araneae, Araneidae). Zoologische Jahrbuch Systematics 89:369^06. Kullmann, E. 1975. The production and function of spider threads and spider webs. Pp. 318-378. In Nets in Nature and Technics. (K. Bach, B. Burkhardt, R. Graefe & R. Raccanello, eds.). Druckerei Heinrich Fink KG, Stuttgart. Kuntner, M., J.A. Coddington & G. Hormiga. 2008. Phylogeny of extant nephilid orb-weaving spiders (Araneae, Nephilidae): testing morphological and ethological homologies. Cladistics 24:147-217. Madrigal-Brenes, R. & G. Barrantes. 2009. Construction and function of the web of Tidarren sisyphoides (Araneae: Theridiidae). Journal of Arachnology 37:306-311. McCook, H.C. 1889. American Spiders and their Spinning Work. Volume 1. Published by the author, Philadelphia. Moya, J., R. Quesada, A. Arias, G. Barrantes, W.G. Eberhard, I. Escalante, C. Esquivel, A. Rojas & E. Triana. 2010. Egg sac construction behavior of by folding dead leaves in Pozonia nigroventris and Micrathena sp. (Araneae: Araneidae). Journal of Arachnology 38:371-373. Nielsen, E. 1931. The Biology of Spiders. Levin & Munksgaard, Copenhagen. Palmer, J.M. 1985. The silk and silk production system of the funnel- web mygalomorph spider Euagriis (Araneae, Dipluridae). Journal of Morphology 186:195-207. Palmer, J.M., F.A. Coyle & F.W. Harrison. 1982. Structure and cytochemistry of the silk glands of the mygalomorph spider Antrodiaetus unicolor (Araneae, Antrodiaetidae). Journal of Morphology 174:269-274. Peters, R. 1967. Vergleichende Untersuchungen iiber Bau und Funktion der Spinnwarzen und Spinnwarzenmuskulatur einiger Araneen. Zoologische Beitrage NF 13:29-120. Platnick, N.L, J.A. Coddington, R.R. Forster & C.E. Griswold. 1991. Spinneret morphology and the phylogeny of haplogyne spiders (Araneae, Araneomorphae). American Museum Novitates 3016: 1-73. Schiitt, K. 1996. Wie Spinnen ihre Netze befestigen. Mikrokosmos 84:274-278. Stern, H. & E. Kullmann. 1981. Leben am seidenen Faden. Kindler Verlag, Munchen. Tillinghast, E.K., E.J. Kavanagh & P.H. Kolbjornsen. 1981. Carbohydrates in the webs of Argiope spiders. Journal of Morphology 169:141-148. Vollrath, F. & P. Selden. 2007. The role of behavior in the evolution of spiders, silks, and webs. Annual Review of Ecology, Evolution and Systematics 38:819-846. Weng, J.-L., G. Barrantes & W.G. Eberhard. 2006. Feeding by Philoponella vicina (Araneae, Uloboridae) and how uloborid spiders lost their venom glands. Canadian Journal of Zoology 84:1752-1762. Wilson, R.S. 1962a. The control of dragline spinning in the garden spider. Quarterly Journal of Microscopical Science 104:557-571. Wilson, R.S. 1962b. The structure of the dragline control valves in the garden spider. Quarterly Journal of Microscopical Science 103:549-555. Witt, P.N., C. Reed & D.B. Peakall. 1968. A Spider’s Web. Springer Verlag, New York. Work, R. 1977. Mechanisms of major ampullate silk fiber formation by orb-web-spinning spiders. Transactions of the American Microscopical Society 96:170-189. Manuscript received 2 November 2009, revised 15 March 2010. 2010. The Journal of Arachnology 38:415^24 Eukoenenia (Palpigradi: Eukoeneniidae) in Brazilian caves with the first troglobiotic palpigrade from South America Maysa Fernanda V. R. Souza and Rodrigo Lopes Ferreira': Laboratorio de Ecologia Subterranea, Setor de Zoologia, Departamento de Biologia, Universidade Federal de Lavras, Lavras, MG. CEP 37200-000, Brazil Abstract. Reports of Palpigradi from South American caves are rare, and no troglobiotic species have yet been described. This apparent deficiency, however, reflects merely a lack of reporting. Ten years of biospeleological surveys of 603 caves in 16 of the 26 Brazilian states, in bedrocks including limestone, quartzite, iron ore, granite, and gneiss, have led to the capture of 494 palpigrades, and specimens with troglomorphic characteristics have been found in Minas Gerais, Bahia, and Espirito Santo. Palpigrades have been found to be relatively more common in iron ore caves, and troglomorphic species apparently occupy cave habitats different from those occupied by edaphomorphic species. The description of the first troglobite species from South America is presented here. Eukoenenia maquinensis, new species, collected in the Maquine Cave Minas Gerais, Brazil, has six blades in the lateral organs, seven pairs of setae on the propeltidium, six setae on the basitarsus IV (a single proximal sternal seta) and a singular chaetotaxy of opisthosomal sternites. Keywords: Eukoenenia maquinensis. Neotropics, troglomorphic The order Palpigradi Thorell 1888 includes small arachnids found in soil, leaf litter, caves, and semi-aquatic interstitial environments (Barranco & Harvey 2008). Since their original discovery in Catania (Grassi and Calundruccio 1885), researchers have found palpigrades in many locations around the world, including northern Africa, Europe, Madagascar, Australia, southeastern Asia, and both North and South America (e.g., Conde 1956, 1974b, 1984a, 1987, 1996; Peyerimhoff 1902; Harvey 2003). In South America, only nine palpigrade species are known: Prokoenenia chilensis (Hansen 1901), Eukoenenia cf. grassii (Hansen 1901), and a member of the E. mirahilis-berlesei complex from Chile; Eukoenenia florenciae (Rucker 1903) from Paraguay, Colombia and Argentina; E. grassii, from Paraguay; E. improvisa Conde 1979, AUokoenenia afra Silvestri 1913 and Koeneniodes notabilis Siivestri 1913 from French Guiana; E. janetscheki Conde 1993 and E. roquettei (Mello- Leitao & Arle 1935) from Brazil (Conde 1979, 1984a, 1986, 1993, 1996; Mello-Leitao & Arle 1935). Except for Leptokoenenia Conde 1965 (with two marine interstitial species) and Triadokoenenia Conde 1991 (mono- typic, associated with rain forests of northeastern Madagas- car), all genera of the order have species endemic to caves. However, species of the genus Eukoenenia Borner 1901 are by far the most highly morphologically modified and most numerous in the underground environment. Of the 66 species of palpigrades included in this genus, 27 are restricted to caves (21 in Europe, one in Cuba, and at least five in tropical Asia) (Conde 1996). The species E. orghidani Conde «fe Juberthie 1981, found in the Bellamar cave in Cuba, is the only recognized troglobiotic species of the order Palpigradi to have been discovered in the Americas (Conde 1996), and no records of troglobiotic palpigrades have been reported from South America. In the present paper, a new troglobiotic species of the genus Eukoenenia, the first from South America, is described, and ' Corresponding author. E-mail: drops@ufla.br various aspects related to the geographical distribution of the genus Eukoenenia in Brazilian caves are noted. METHODS Various inventories of cave fauna using the methodology proposed by Ferreira (2004) have been compiled in 603 Brazilian caves over the past ten years, including caves in 16 of the 26 Brazilian states (Minas Gerais, Goias, Espirito Santo, Bahia, Rio de Janeiro, Mato Grosso, Sao Paulo, Parana, Santa Catarina, Rio Grande do Sul, Sergipe, Ceara, Rio Grande do Norte, Alagoas, Pernambuco and Tocantins). Caves in various lithologies were included in the inventories: carbonate rocks (limestone, dolomite, calcarenite, marble, and calcoschist), conglomerates, iron ore (hematite and itabirite), siliclastic rocks (quartzite and sandstone), and magmatic rocks (gneiss and granite). In all the caves we conducted a thorough search for palpigrades in all potential microhabitats: on the floor of the cave, on the speleothems, and under rocks, from regions near the entrance to those deeper inside the cave. They were caught with fine brushes and fixed in 70% alcohol. The area outside the caves was also inspected when the environmental conditions appeared to be favorable for the occurrence of palpigrades. The specimens examined for this study are lodged in the Cole^ao de Invertebrados Subterraneos do Laboratorio de Ecologia Subterranea do Departamento de Biologia da Universidade Federal de Lavras (UFLA), Lavras, MG, Brazil, and the Instituto Butantan, Sao Paulo, SP, Brazil (IBSP). A chi-square test was used to evaluate the abundance of palpigrades in different lithologies. In order to obtain the local average yield of non-quantitative palpigrade sampling (by municipality), the total number of palpigrades collected in the caves of a municipality was divided by the number of palpigrade-inhabited caves. The material was examined by clearing in Nesbit’s solution and mounted in Hoyer’s medium on 7.6 X 2.5 cm glass slides using standard procedures for mites (Krantz & Walter 2009). We made drawings under phase contrast microscopy, with measurements reported in micrometers (pm). Body length was 415 416 THE JOURNAL OF ARACHNOLOGY Table 1. — Brazilian municipalities with occurrence of underground palpigrades. Municipality Caves State Lithology Total abundance Relative abundance luiu 2 BA Limestone 2 1 Sao Desiderio 1 BA Limestone 3 3 Itaete 1 BA Limestone 2 2 Afonso Claudio 1 ES Gneiss 2 2 Arcos 11 MG Limestone 13 1.18 Bambui 1 MG Limestone 1 1 Brejo da Madre de Deus 1 PE Granite 1 1 Corsdisburgo 1 MG Limestone 7 7 Doresdpolis 10 MG Limestone 10 ! Felipe Guerra 2 RN Limestone 7 3.5 Iguatama 2 MG Limestone 6 3 Itabirito 5 MG Iron ore 17 3.4 Itumirim 1 MG Limestone 1 1 Lagoa da Prata 7 MG Limestone 20 2.85 Lagoa Santa 1 MG Limestone 7 7 Lima Duarte 3 MG Quartzite 7 2.33 Mambai 1 GO Limestone 1 1 Moeda 4 MG Iron ore 140 35 Nova Lima 10 MG Iron ore 77 7.7 Padre Paraiso 1 MG Granite 2 2 Pains 36 MG Limestone 65 1.8 Paracatu 1 MG Limestone 1 1 Prudente de Morais 3 MG Limestone 3 1 Santa Tereza 1 ES Granite 3 3 Sete Lagoas 1 MG Limestone 3 3 Vargem Alta 1 ES Marble 1 1 Varre e Sai 1 RJ Granite 5 5 Curvelo 1 MG Limestone 1 1 Januaria/Itacarambi 1 MG Limestone 1 1 Vazante 3 MG Dolomite 21 7 measured from the apex of the propeltidium to the posterior margin of the opisthosoma. The following abbreviations were used, based on Barranco & Mayoral (2007): L, total length of body (without flagellum); B, length of dorsal shield; P, pedipalpus; I and IV, legs I and IV; ti, tibia; btal, basitarsus 1; bta2, basitarsus 2; bta3, basitarsus 3; bta4, basitarsus 4; tal, tarsus I; ta2, tarsus 2; ta3, tarsus 3; a, width of basitarsus IV at level of seta r; er, distance between base of basitarsus IV and insertion of seta r; grt, length of tergal seta; gla, length of lateral seta; r, length of stiff seta; t/r, ratio between length of basitarsus IV and stiff seta length; t/er, ratio between length of basitarsus IV and distance to insertion of stiff seta; gla/grt, ratio between lengths of lateral and tergal setae; B/bta, relation between lengths of prosomal shield and basitarsus IV; bta/ti, ratio between lengths of basitarsus IV and tibia IV; F, flagellar segments. Setal nomenclature follows that of Conde (1974a, 1974b, 1981, 1984a, 1988, 1989, 1992, 1993, 1994). RESULTS Distribution of species of genus Eukoenenia in Brazilian caves. — Individuals of the order Palpigradi were found in 131 of the 603 caves surveyed (21.7% of the total). We discovered these palpigrades in various municipalities in the states of Minas Gerais, Goias, Espirito Santo, Rio de Janeiro, Pernam- buco, Rio Grande do Norte, and Bahia. The lithologies of the caves in which palpigrades were found included carbonate rocks (limestone), iron ore, magmatic rocks (granite and gneiss), and siliclastic rocks (quartzite) (Table 1). Palpigrades were especially prevalent in iron ore caves, and a chi-square test showed that the abundance of palpigrades varied as a function of host rock type, with the number found in iron ore caves double the number that would have been expected. A total of 494 individuals was found in the various habitats inside the different caves, normally under rocks or in the soil on the floor of the cave; rarely, they were discovered crawling along the walls or on speleothems. The average number of palpigrades per cave per municipality shows that these invertebrates are most abundant in the state of Minas Gerais, especially in caves in iron ore, although many are also found in limestone caves (Fig. 1). In the epigean environment in the vicinity of the caves, we only found palpigrades in the state of Minas Gerais (municipalities of Iguatama, Lavras, Pedro Leopoldo, Novo Oriente de Minas, Arcos, and Pains). They were especially prevalent in the latter two municipalities, where dozens of individuals were found under the rocks in the forest; however, all sightings were made during rainy periods. It seems probable that during the dry seasons they penetrate the interstices of the soil to protect themselves from the dry conditions of the surface. All exemplars collected in Brazilian caves were identified as Eukoenenia. In addition to the troglobiotic species found in the Maquine cave, described in this study, individuals col- lected in six other caves in the states of Espirito Santo, Minas SOUZA & FERREIRA— EUKOENENIA IN BRAZILIAN CAVES 417 Figure 1. — Distribution of the municipalities with occurrence of subterranean palpigrades in Brazil and its respective average abundances. 3 Figures 2, 3. — Eitkoenenia maqumensis new species (holotype): 2. Frontal organ, dorsal view; 3. Lateral organ, dorsal view. Scale bars: 32.5 pm. 418 THE JOURNAL OF ARACHNOLOGY Figures 4-6. — Eukoenenia maquinensis new species (holotype): 4. Propeltidial chaetotaxy; 5. Metapeltidial setae; 6. Deuto-tritosternal setae. Scale bars: 150 pm (Fig. 4), 100 pm (Fig. 5), 50 pm (Fig. 6). Gerais and Bahia presented troglomorphic traits, such as lengthening of appendices and the flagellar segments and an increase in body size. These other troglomorphic species are under description. The study of the edaphomorphic species captured has revealed the presence of at least four undescribed species. We observed these troglomorphic species moving along the walls and speleothems, whereas the edaphomorphic species were generally hidden under rocks in the soil on the cave floor, although two were found walking on top of water accumu- lated in travertine dams in the Lapa Nova cave in Vazante and Figures 7-10. — Eukoenenia maquinensis new species (paratype): 7. Coxa I; 8. Coxa II; 9. Coxa III; 10. Coxa IV. Scale bar: 100 pm. Figures 11-13. — Eiikoenenia maqumemis new species (holotype): 11. Chelicerae; 12. Basitarsus 3^ of leg I; 13. Basitarsus IV. Scale bar: 100 pm. another in the cave of Santuario in Pains, all in the state of Minas Gerais. TAXONOMY Family Eukoeneniidae Petrunkevitch 1955 Genus Eiikoenenia Borner 1901 Koenenia Grass! & Calandruccio 1885:165 [junior primary homonym of Koenenia Beushausen 1884 (Mollusca: Bival- via)]. Koenenia {Eiikoenenia) Borner 1901:551. Type species. — Koenenia mirabilis Grass! & Calandruccio 1885, by monotypy. Eiikoenenia maqtiinensis new species (Figs. 2-20) Material examined. — Brazil; Minas Gerais: Holotype adult female, Maquine cave (collected from a speleothem), Cordis- burgo (19°11T5"S, 44°18'45"W), 16 December 2007, R.L. Ferreira (IBSP, IBSP05). Paratype: 1 adult female, Maquine cave, Cordisburgo (19°11T5"S, 44°18'45"W), 16 December 2007, R. L. Ferreira (UFLA, ISLA 502). Diagnosis. — Eiikoenenia maqiiinensis differs from all other species of the genus by the following combination of characters: presence of 6 blades on prosomal lateral organs; propeltidium with 7 + 7 setae; six setae on basitarsus IV with a single proximal sternal seta; opisthosomal sternites IV-VI with 14, 13, and 11 setae, respectively, in an irregular row. Description. — Prosoma. Frontal organ with two branches, blunt apically, each 5.5 times longer than wide (55 pm/lO pm) (Fig. 2). Lateral organ with six pointed blades, each 6.7 times longer than wide (33.75 pm/5 pm) (Fig. 3); Fig. 3 only shows five blades due to the hidden position of the sixth blade. Propeltidium with 7 + 7 short setae, first pair on either side of frontal organ longer than others (Fig. 4). Metapeltidium with 3 + 3 setae (c, c, is), each of a different length; outer setae shortest (157.5 pm, 125 pm and 115 pm) (Fig. 5). Deutotritosternum with nine or eight setae in U-shaped arrangement (Fig. 6). Chaetotaxy of coxae I-IV: 11, 15, 15 and 9 (Figs. 7-10). The holes in Figures 7 and 10 represent the insertions of each of the respective setae, which are positioned just near each hole. Chelicerae with 9 teeth on each finger; 4 dorsal setae and 1 ventral seta inserted near the third segment, and 1 seta inserted near the row of teeth of the second segment (Fig. 1 1). 420 THE JOURNAL OF ARACHNOLOGY Figures 14, 15. — Eiikoenenia maquinensis new species (holotype): I 300 pm. Basitarsus 3 of leg I slender, 4.23 times longer than wide, with 3 setae {grt 140 pm; r 125 pm). Seta r shorter than segment (137.5 pm/125 pm, tir — 1.1), inserted in proximal half and surpassing hind edge (42.5 pm/125 pm, sler = 0.34) (Fig. 12). Basitarsus of leg IV long, 9.1 times longer than wide, with 6 setae (2 esd, esp, gla, grt and r) (Fig. 13), bta/ti 1.07. Stiff seta r 2.75 times shorter than tergal edge of article (275 pm/100 pm, t/r = 2.75) and inserted in distal third (275/ 185, tier = 1.48). Seta esp proximally inserted, followed by grt and gki, more or less at the same level, all of them in proximal half. Opisthosoma: Tergites II-VI with 3 + 3 setae each, 2 pairs of tergal setae (t) (/;, 0) between both slender setae (s). Tergites VII-VIII with 2 + 2 setae, only t setae present and without s (Fig. 14). Sternite III with 2 + 2 setae. One seta on left side of this structure was not represented in the drawing, but was visible unattached near the sternite. Its insertion was not represented due to the difficulty in finding it precisely. Sternites IV-VI with 14, 13 and 11 setae in slightly irregular transverse row, all of them of similar shape and with length varying between 65-75 pm, 52.5-77.5 pm and 72.5-82.5 pm respectively. Sternites IV-V each with two glandular pores. Sternites VII-VIII with 2 + 2 setae. Segments IX-XI each with 6 setae (Fig. 15). Genitalia: With 2 lobes, first lobe (Fig. 16) with 11 + 12 setae in 5 transverse rows, 4 sternal 2 + 2, 2 + 3 (asymmetry caused by dislocations due to the lack of regular and/or the presence of additional setae), 2 + 2, 1 + 1 and distal 4 + 4, of which a,, a2, as, a4 measure 35 pm, 30 pm, 37,5 pm and 50 pm. L Opisthosoma, dorsal view; 15. Opisthosoma, ventral view. Scale bar: respectively. Second lobe (Fig. 17) with 3 + 3 setae (seta y was possibly lost during the collection or the specimen presents an asymmetry) (x, y, z), measuring 32.5 pm, 40 pm, and 30 pm, respectively; six glandular orifices. Spermathecae triangular, with the base measuring 15 pm and the height measuring 10 pm. Flagellum: Longer than opisthosoma, with 14 long, slender articles (Fig. 18). First, thirteenth and fourteenth segments considerably shorter than others. First 3 segments have structures similar to a crown of spines around the extremity, with 1 1 long setae inserted in the distal half. Fourth, sixth and eighth segments without the crown of spines and with 9 long setae inserted in distal third. Fifth, seventh and ninth segments also with spines on their extremities and with 8 long setae inserted in distal half (Fig. 19). In the other segments, the crown of spines is lacking. Tenth segment with 9 long setae inserted in the distal half. Eleventh segment with 8 long setae in proximal half. Twelfth and thirteenth segments with 8 long setae inserted in proximal third. Last segment with 6 long setae inserted in middle of segment and 4 setae inserted apically (Fig. 20). Dimensions (\im): See Table 2. Habitus: See Fig. 21 Etymology. — The specific name refers to the cave, Maquine, where the specimens were found. Habitat. — Maquine cave is the only known habitat of E. maquinensis. This cave is the oldest tourist cave in Brazil and has been exploited for this purpose since 1908. In 1967, the infrastructure was modernized to include not only stairs and SOUZA & FERREIRA— EUKOENENIA IN BRAZILIAN CAVES Figures 16, 17. — Eiikoenenia mciquinensis new species (holotype): 16. Female genitalia, first lobe; 17. Female genitalia, second lobe. Scale bars: 100 pm (Fig. 16), 50 pm (Fig. 17). topographical alterations to the floor of the cave, but also electric lights. At the far end of the cave, the relative humidity is approximately 91%, and the average temperature is about 24° C. The trophic resources consist of organic material left during the installation of the tourist infrastructure, such as wood scraps, as well as scraps of food left by tourists during their visit to the cave and small plants growing near the lights. The specimens described here were collected in an area not subjected to visitation by tourists. DISCUSSION The environments suitable for populations of palpigrades include caves and other subterranean environments, and these invertebrates have been reported from such systems in many locations around the world, including Europe, Asia, Central America and Africa (e.g., Conde 1984a, 1987, 1996; Barranco & Harvey 2008). Caves in sixteen of the Brazilian states were surveyed, and palpigrades were found in seven of these states. Moreover, since relatively few surveys of cave fauna have been conducted in the country, the geographical distribution of the palpigrades in Brazilian caves is probably much greater than that 421 Table 2. — Measurements (pm) of selected body parts of the two type specimens of Eukoenenia maqiiinensis. Body part Female (Holotype) Female (Paratype) L 1490 2140 B 435 462.5 Pti 235 252.5 PBtal 90 95 PBta2 105 112.5 Ptal 55 60 Pta2 75 75 Pta3 80 112.5 Iti 320 320 IBtal+2 245 260 IBta3 137.5 147.5 IBta4 115 112.5 Ital 70 70 Ita2 80 72.5 Ita3 215 220 IVTi 255 252.5 IVBta 275 270 IVTal 102.5 100 IVTa2 125 130 A 30 25 Er 185 187.5 Grt 137.5 130 Gla 127.5 135 R 100 100 t/r 2.75 2.7 t/er 1.48 1.44 Gla/grt 0.92 1.03 B/bta 1.58 1.7 Bta/ti 1.078 1.07 Flagelo 3865 - FI 197.5 - FII 240 Fill 275 300 FIV 290 325 FV 250 - FVI 320 305 FVII 265 275 FVIII 332.5 325 FIX 260 260 FX 377.5 385 FXI 372.5 375 FXII 352.5 320 FXIII 170 330 FXIV 162.5 175 described here. Furthermore, some regions had been poorly studied, as the Central Amazon, where studies revealed a high abundance of a single species, E. janetsclieki (Conde 1993; Adis et al. 1997). Biological surveys of some 350 caves in Brazil have been previously reported in the literature (Pinto-da- Rocha 1995; Trajano 1996, 2000; Gomes et al. 2000; Zeppelini Filho et al. 2003; Prous et al. 2004; Ferreira 2005; Silva 2006). In these papers, few palpigrades have been reported, except for iron ore caves (Ferreira 2005). The reports of Trajano (1996, 2007) for the caves of Olhos oAgua (MG) and the System of Areias (SP) include no taxonomic identification. Moreover, this author gave little emphasis to these arachnids, failing to include information about abundance and microhabitat, as well as the behavior of individuals prior to collection. This 422 THE JOURNAL OF ARACHNOLOGY Figures 18-20. — Eitkoenenia maquinensis new species (holotype): 18. First six flagellar segments; 19. Fifth flagellar segment; 20. Fourteenth flagellar segment. Scale bars; 600 pm (Fig. 18), 100 pm (Fig. 19), 150 pm (Fig. 20). lack of information, in conjunction with the limited number of records of palpigrades, has limited our understanding of their behavior and ecology. Moreover, inadequate collection methods in caves may have led to the misconception that palpigrades are rare in underground Brazilian systems. As shown in this paper, more than 20% of the caves surveyed have palpigrades, which shows that these organisms are not rare in Brazilian caves. Palpigrades have been discovered in epigean environments near cave entrances, as well as in their disphotic hypogean interiors, a fact that probably reflects the high humidity prevailing when collections were made. These populations can stay inside and outside of caves in the same territory. Conde (1996) affirms that the most likely places to find palpigrades inside caves are spaces under rocks, in the soil on the floor of the cave, or moving around on the ground or walls. He explains that the occupation of these different habitats is determined by factors such as the hygrometry and the movement of air in the cave. During the collecting process, we observed that the troglomorphic palpigrades were fre= quently found walking along the walls and on speleothems, whereas the edaphomorphic species were found only under rocks or in the soil. Similar information is reported for troglomorphic individuals of E. orghidani (in Cuba) and E. spelaea (in the Alps of Provence) found walking around exposed on stalagmites (Conde 1984b). The adoption of this exposed habitat may reflect behavioral modifications that have accompanied the morphological adaptations linked to the underground environment. Therefore, the occupation of habitats unlike those inhabited by palpigrades under other circumstances seems to be related both to environmental aspects of the cave in which they are found and to the degree of specialization of the species in the underground environ- ment. The greater frequency of palpigrades in iron ore caves may be explained by the structural and functional characteristics of this type of underground system. These rock systems, especially the weathered mantle (“canga”) covering the mother rock, include numerous minute conduits forming a network of interstitial spaces (meso- and microcaves) which are linked to the larger caverns, thus facilitating colonization of regions far from the entrance (Ferreira 2005). Furthermore, the epigean vegetation associated with iron ore is relatively sparse and provides rather limited leaf litter (Ferreira 2005), making the epigean environment unfavorable for the survival of soil fauna. The troglomorphic species described here as Eukoenenia maquinensis displays various troglomorphic characteristics. These include reduction of the length and number of setae of the propeltidium (7 + 7) as in E. naxos Conde 1989, a reduction in the number of setae of the basitarsus of leg IV, with six setae instead of seven (absence of an esp seta), and a flagellum composed of 14 narrow and much elongated segments. The presence of tiny setae in the propeltidium is shared with other troglomorphic species as E. gasparoi Conde 1988 (10 + 10 setae), E. thais Conde 1988 (10 + 10) and E. maws Conde 1992 (10 + 9). Moreover, this species has elongated appendages, such as the elongation of the basitarsus IV of E. maquinensis, which is longer than the tibia, with a bta/ti ratio of 1.07. This value is close to that found for E. naxos (bta/ti = 1.10), considered to be the species that has reached the highest level of underground evolution, with the exception of the lateral organs of the prosoma (Conde 1996). The mean value of Blbta SOUZA & FERREIRA— IN BRAZILIAN CAVES 423 Figure 21. — Habitus of Eukoenenui maqidnemis new species. IV in the two specimens is 1.64, close to that of other cave species such as E. naxos {\.1\), E. thais Conde 1988 (1.61), and E. grafittii Conde & Heurtault 1994 (1.59) (Conde 1998). Conde (1996) considers body length to be a very arbitrary criterion for determining troglomorphism, since it is related to how stretched out an individual is in the preserving solution when measurements are taken. The holotype of Eukoenenia maquinensis measured only 1.49 mm, but this would correspond to approximately 2 mm, since the specimen shrank upon contact with the 70% alcohol used for preservation. The paratype preserved in 60% alcohol mea- sured 2.14 mm, which is a more realistic representation of the adults’ real size. Furthermore, the flagellum is extremely fragile and is rarely preserved. If it survives the original capture, it seldom survives subsequent manipulation. Few specimens have been described with an intact flagellum, although Conde (1996) suggests that this length varies from 1.36 to 3.25 mm for a body length of 0.97 to 2.20 mm for adult cave species. If this is accurate, E. maquinensis has the longest flagellum reported for a species of Eukoenenia, with a length of 3.865 mm. The lateral organs of Eukoenenia maquinensis are composed of six elements, similar to that of E. spelaea (Peyerimhoff 1902) (5-6), E. depilata Remy 1960 (6), E. remyi Conde 1974 (4-6), E. maroccana Barranco & Mayoral 2007 (6) and E. guzikae Barranco & Harvey 2008 (6). It was not possible to establish the deutotritosternal chaetotaxy because one of the specimens had nine setae while the other had only eight setae. The number of specimens collected was inadequate to determine if this variation in the setae represented a population characteristic or if one of the specimens served as an exception. The holotype genitalia suggest asymmetry caused by dislocations due to the lack of regular and/or the presence of additional setae, not unusual in Eukoenenia. Unfortunately, the genital lobes of the paratype were damaged during slide mounting. It is impossible to determine if this asymmetry is an exclusive trait of the holotype female; in that case, the actual number of setae would be 11+11 or 12+12. In the opisthoso- mal sternites IV-VI the tip of the lateral setae is similar to the tip of the paramedian setae, hence, it is very difficult to distinguish the thickened setae (a) from normal setae (s). Eukoenenia maquinensis is certainly one of the most modified palpigrades adapted to the underground environ- mental conditions. Many models concerning the evolution of subterranean lineages state that isolation (due to climatic changes) was more pronounced in temperate regions than in tropical areas. This would be the reason why there are so many modified troglobiotic species in temperate areas. However, the strongly modified species described here suggests that the effects of climatic changes in Neotropics (leading to isolation in subterranean habitats) could have led to the same effects observed in the temperate areas, at least for some groups such as the palpigrades. ACKNOWLEDGMENTS We thank Marconi Souza Silva, who first found a specimen, Leopoldo Bernardi for methodological suggestions and Adalberto Santos for clarification related to the morphology of the group. We would also like to thank Paulo Rebelles Reis (EPAMIG-CTSM/EcoCentro Lavras) for use of the phase contrast microscope and the entire staff of the Laboratory of Underground Ecology of the Section of Zoologia of the Federal University of Lavras (UFLA) for their efforts in the 424 THE JOURNAL OF ARACHNOLOGY collection, and the Fundagao de Amparo a Pesquisa de Minas Gerais (FAPEMIG CRA-APQ-03526-09) for the financial support. Finally, we would like to acknowledge the assistance of the team of Maquinetur, the mayor of Cordisburgo, and especially Mario and Gilson, for their invaluable assistance during the field work in the Maquine cave. LITERATURE CITED Adis, J., U. Scheller, J.W. De Morais, B. Conde & J.M.G. Rodrigues. 1997. On the abundance and phenology of Palpigradi (Arachnida) from Central Amazonian upland forests. Journal of Arachnology 25:32^332. Barranco, P. & J.G. Mayoral. 2007. A new species of Eukoenenia (Palpigradi; Eukoeneniidae) from Marroco. Journal of Arachnol- ogy 35:318-324. Barranco, P. & M.S. Harvey. 2008. The first indigenous palpigrade from Australia: a new species of Eukoenenia (Palpigradi: Eu- koeneniidae). Invertebrate Systematics 22:227-233. Conde, B. 1956. Une Koenenia cavernicole de Sardaigne (Arachnides microteliphonides). Notes Biospeleologiques 11:13-16. Conde, B. 1974a. Eukoenenia remyi n. sp., palpigrade cavernicole d’Herzegovine. Annales de Speleologie 29:53-56. Conde, B. 1974b. Un palpigrade cavernicole du Libian {Eukoenenia juherthiei n. sp.). Annales de Speleologie 29:57-62. Conde, B. 1979. Palpigrades de Grece, de Guyane et du Kenya. Revue Suisse de Zoologie 86:167-179. Conde, B. 1981. Palpigrades des Canaries, de Papouasie et des Philippines. Revue Suisse de Zoologie 88:941-951. Conde, B. 1984a. Palpigrades (Arachnida) d’Europe, des Antilles, du Paraguay et de Thailande. Revue Suisse de Zoologie 91:369-391. Conde, B. 1984b. Les Palpigrades: quelques aspects morphobiologi- ques. Revue Arachnologique 5:133-143. Conde, B. 1986. Les palpigrades du nouveau monde: etat des connaissances. Memoires de la Societe Entomologique de Belgique 33:67-73. Conde, B. 1987. Les palpigrades des lies de la Mediterranee (Arachnida Palpigradida). Bulletin de la Societe Zoologique de France 112:215-219. Conde, B. 1988. Nouveaux palpigrades de Trieste, de Slovenie, de Make, du Paraguay, de Thailande et de Borneo. Revue Suisse Zoologie 95:723-750. Conde, B. 1989. Palpigrades (Arachnida) de grottes d’Europe. Revue Suisse de Zoologie 96:823-840. Conde, B. 1992. Palpigrades cavernicoles et endoges de Thailande et des Celebes (U'''"’ note). Revue Suisse de Zoologie 99:655-672. Conde, B. 1993. Description du male de deux speces de palpigrades. Revue Suisse de Zoologie 100:279-287. Conde, B. 1994. Palpigrades cavernicoles et endoges de Thailande et de Celebes (2'^ note). Revue Suisse de Zoologie 101:233-263. Conde, B. 1996. Les palpigrades, 1885-1995: acquisitions et lacunes. Revue Suisse de Zoologie hors-serie 1:87-106. Conde, B. 1998. Palpigradida. Pp. 913-920. In Encyclopaedia Biospeologica. Volume 11. (C. Juberthie & V. Decu, eds.). Societe Internationale de Biospeologie, Moulis (C.N.R.S.), France and Bucarest (Academic Rumanie), Roumanei. Ferreira, R.L. 2004. A medida da complexidade ecologica e suas aplicagoes na conservagao e manejo de ecossistemas subterraneos. Tese de Doutorado, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil. Ferreira, R.L. 2005. A vida subterrenea nos campos ferruginosos. O Carste 3:106-115. Gnaspini, P. & E. Trajano. 1994. Brazilian cave invertebrates, with a checklist of troglomorphic taxa. Revista Brasileira de Entomologia 38:549-584. Gomes, F.T.M.C., R.L. Ferreira & C.M. Jacobi. 2000. Comunidade de artropodes de uma caverna calcaria em area de mineragao; composigao e estrutura. Revista Brasileira de Zoociencias 2:77-96. Harvey, M.S. 2003. Catalogue of the Smaller Arachnid Orders of the World: Amblypygi, Uropygi, Schizomida, Palpigradi, Ricinulei and Solifugae. CSIRO Publishing, Melbourne, Australia. Krantz, G.W. & D.E. Walter. 2009. A Manual of Acarology. 3rd Edition. Texas Tech University Press, Lubbock, Texas. Mello-Leitao, C. & R. Arle. 1935. Consideragoes da ordem Palpigradi com a descrigao de uma nova especie. Anais da Academia Brasileira de Ciencias 7:339-343. Monniot, F. 1966. Un palpigrade interstitiel: Leptokoenenia scurra, n. sp. Revue d’Ecologie et de Biologic du Sol 3:41-64. Peyerimhoff, P. 1902. Decouverte en France du genre Kaenenia (sic) (Arachn. Palpigradi). Bulletin de la Societe Entomologique de France 1902:280-283. Pinto-da-Rocha, R. 1995. Sinopse da fauna cavernicola do Brasil (1907-1994). Papeis Avulsos de Zoologia 39:61-173. Prous, X., R.L. Ferreira & R.P. Martins. 2004. Ecotone delimitation: Epigean-hypogean transition in cave ecosystems. Austral Ecology 29:374-382. Rowland, J.M. & W.D. Sissom. 1980. Report on a fossil palpigrade from the Tertiary of Arizona, and a review of the morphology and systematics of the order (Arachnida: Palpigradida). Journal of Arachnology 8:69-86. Silva, F.J. 2006. Invertebrados de cavernas do Distrito Federal: diversidade, distribuigao temporal e espacial. Tese de Doutorado, Universidade de Brasilia, Brasilia, Brazil. Trajano, E. 1996. Biologia da Gruta Olhos d’agua. O Carste 8:85-89. Trajano, E. 2000. Cave faunas in the Atlantic tropical rain forest: composition, ecology, and conservation. Biotropica 32:882-893. Trajano, E. 2007. Biologia. Pp. 1^9. In Sistema Areias: 100 anos de estudos. Redespeleo, Sao Paulo, Brazil. Zeppelini, D.F., A.C. Ribeiro, G.C. Ribeiro, M.P.A. Fracasso, M.M. Pavanni, O.M.P. Oliveira, S.A. Oliveira & A.C. Marques. 2003. Faunistic survey of sandstone caves from Altinopolis region, Sao Paulo State, Brazil. Papeis Avulsos de Zoologia 43:93-99. Manuscript received 21 December 2009, revised 18 May 2010. 2010. Ihe Journal of Arachnology 38:425^32 On the Cryptocellus peckorum and Cryptocelius adisi groups, and description of a new species of Cryptocellus from Brazil (Arachnida: Ricinulei) Ana Lucia Tourinho and Regiane Saturnine: Institute Nacional de Pesquisas da Amazonia, INPA, Avenida Andre Araujo, 2936, Aleixo, CEP 6901 1-970, Caixa Postal 478, Manaus, AM, Brazil. Coordenagao de Pesquisas em Ecologia - CPEC; Coordenagao de Pesquisas em Entomologia - CPEN. E-mail: amtourinho@gmail.com Abstract. A new species of Ricinulei of the genus Cryptocellus Westwood 1874 is described from the Madeira-Purus Interfluve, Amazonas, Brazil. It shares a set of apomorphies with Cryptocellus peckorum Platnick & Shadab 1977 and Cryptocellus tarsilae Pinto-da-Rocha & Bonaldo 2007, with which it forms an assemblage of related species herein named the peckorum group. A second group of species, the Cryptocellus adisi group, is formed by the following species: Cryptocellus adisi Platnick 1988, Cryptocellus florezi Platnick & Garcia 2008, and Cryptocellus lishetliae Gonzalez-Sponga 1998. Keywords: New World, Amazon rainforest, Madeira-Purus interfluve, BR-319 highway, taxonomy, systematics The arachnid order Ricinulei has been increasingly captur- ing attention of New World specialists, and over the last seven years several new species have been described: Cryptocellus abaporu Bonaldo & Pinto-da-Rocha 2003; C. icamiabcis Tourinho & Azevedo 2007; C tarsilae Pinto-da-Rocha & Bonaldo 2007; C. florezi Platnick & Garcia 2008; C. platnicki Botero-Trujillo & Perez 2008; C. luisedieri Botero-Trujillo & Perez 2009 (Bonaldo & Pinto-da-Rocha 2003; Botero-Trujillo & Perez 2008, 2009; Pinto-da-Rocha & Bonaldo 2007; Platnick & Garcia 2008; Tourinho & Azevedo 2007); and advances undertaken in the fields of functional, structural biology, and genetics (Talarico et al. 2005, 2006). The taxonomy and systematic foundation of New World representatives of the order Ricinulei were established by Platnick and his co-workers (e.g. Platnick & Shadab 1976, 1977, 1981; Platnick & Paz 1979; Platnick, 1980; Platnick & Pass 1982). Several new species were described from the New World, and the current generic and species-group divisions were established. The groups proposed by Platnick are stable and mostly supported by the characters of the male tarsal process. However, species of Ricinulei seem to be conserva- tive, bearing very similar external morphology, and some species give the impression of a morphological mosaic of several different species (Tourinho & Azevedo 2007). In this paper we describe a new species from Brazil, Cryptocellus conori, and propose the Cryptocellus peckorum group. The group includes the new species herein described and three related species previously described: C. peckorum Platnick & Shadab 1977; C. tarsilae and C. lampeli Cooke 1967. The four species share characters present on the accessory piece, basitarsus and teleotarsus II of leg III, cucullus and ventral abdomen. A second group of species, the Cryptocellus adisi group, is also defined on the basis of characters of the male tarsal process. We also report some new records of C. becki Platnick & Shadab 1977 from the city of Manaus, Amazonas state. The new species was collected in a remote site in the Amazonas State in Brazil, between the rivers Madeira and Purus, one of the most important Amazonian regions for biodiversity (Py- Daniel et al. 2007) and under strong anthropogenic pressure (Fearnside 2005; Fearnside & Graga 2006). METHODS The specimens of C conori were collected using both Winkler extractors and visual nocturnal search and are preserved in 70% ethanol. The specimens of C. becki were collected using Winkler extractors, and all geographical coordinates provided were obtained from the site coordinates report in the website Large Scale Biosphere Experiment in Amazonia (online at http://www.lbaeco.org/cgi-bin/web/sites/ sites_report.pl). The morphological terminology follows Plat- nick & Shadab (1976, 1977, 1981) and Cokendolpher (2000); the general description is based on Cokendolpher (2000) and Cokendolpher & Enriquez (2004), who focused essentially on significant taxonomic structures and colors. All measurements are in mm. The specimens are lodged in the Instituto Nacional de Pesquisas da Amazonia (INPA), Manaus, Amazonas, Brazil. TAXONOMY Family Ricinoididae Ewing 1929 Genus Cryptocellus Westwood 1874 Cryptocellus foedus Westwood 1874:201. Type species. — Cryptocellus foedus Westwood 1874, by monotypy. Cryptocelius conori new species (Figs. 1-20) Type material. — BRAZIL: Amazonas: Male holotype, 30 km Igap6 -Agu, Careiro, 04°54'57"S, 61°06'45.4"W, 23 July 2007, visual nocturnal search, E.H. Wienskoski (INPA 23). Para- types: 1 female, Careiro (Area 1, Forest 2), 04°09'55.4"S, 60°08'00.37"W, 5 July 2007, Winkler extractor (INPA 25); 1 nymph, Careiro, 04°09'26.3"S, 60°07'53"W, 6 July 2007, visual nocturnal search, E.H. Wienskoski (INPA 24). Etymology. — A noun in apposition. In Amazonian mythol- ogy Conori was the powerful and brave queen of the female warriors “Icamiabas”, related to the Amazon warriors from Capadocia and described in the chronicles of Dominican friar Caspar de Carvajal. Diagnosis. — Cryptocellus conori shares with all other species of the peckorum group three basic synapomorphies: fixed male 425 426 THE JOURNAL OF ARACHNOLOGY Figure 1 . — Cryptocellus conori new species, male holotype, dorsal view. Scale bar = 1 .0 mm. accessory piece of leg III thin; both movable and fixed piece curved, and with a sinuous contour; and basitarsus of leg III moderately inflated (longer than wide). The species possess a ventral opisthosoma with three pairs of median pits containing tubercles, the anterior pair smaller than the posterior pair, those with darker and smoother tegument (Fig. 4); no abdominal pits have been reported for C. tarsilae, C. peckoriim Figures 2-5. — Cryptocellus conori new species, male holotype: 2. Prosoma, dorsal view; 3. Cucullus, dorsal view; 4. Opisthosoma, ventral view; 5. Tergite XL Scale bars = 1.0 mm. and C. lampeli; proximal telotarsus II distally without prolateral tubercles as in C. peckorum. Apex of accessory piece distally crenulated (Figs. 8, 9), inner half of spermathe- cae trilobate, outer half acuminate (Fig. 15), in C. tarsilae both inner and outer spermathecae bilobate, female of C. lampeli and C peckorum not known. Description. — Male holotype: Body total length, excluding pygidium, 3.9, cucullus 0.5 long, greatest width 0.8, prosoma 1.0 long, 0.9 wide between legs 11 and III; opisthosoma 2.9 long, 1.9 wide near middle of tergite; legs I 3.8, II 6.8, III 5.0, IV 4.8. General body color (in 70% ethanol) dark red, intersegmental membranes orange; median and posterior portion of ventral opisthosoma lighter. Cucullus dark red, but lighter than prosoma, orange in middle. Prosoma, opisthosoma, legs and cucullus covered with numerous tubercles (Fig. 1); iridescent tubercles, showing shades of purple and dark green, present on: prosoma (Fig. 2), opisthosoma, ventral opisthosoma (Fig. 4), legs and sternal region. Prosoma dark red, much darker at lateral and posterior margin. Tergal tubercles uniformly distributed (Fig. 5). Legs and lateral margin of prosoma covered by both straight and curved whitish setae; ventral opisthosoma covered with numerous setae, concolorous with the body. Prosoma as long as wide, with numerous tubercles uniformly distributed (Fig. 2). One pair of lateral eyes, cucullus densely covered with tubercles, wider than long; anterior margin straight, posterior depressed and rounded (Fig. 3); covered with many white setae, the longest on median portion. Chelicerae: fixed finger with four teeth (distal longer than others); movable finger with 9 teeth (basal almost vestigial). Sternal region with coxa I not meeting tritosternum; coxae 11, III and IV touching medially. Opisthosomal tubercles uniformly distributed. Pygidium with very slight distal dorsal notch on basal segment; no ventral notch. Pedipalps orange, with discrete tubercles. Distal leg coxae darker, leg tarsi light TOURINHO & SATURNINO— CRrRrOCfLLC/S PECKORUM AND ADIS! GROUPS 427 Figures 6-9. — Cryptocellus conori new species, male holotype: 6. Male leg III, anterior view; 7. Male leg III, posterior view; 8. Acessory tarsal process, anterior view; 9. Acessory tarsal process, posterior view. Scales bar = 0.5 mm. red. Leg formula II > III > IV > I; trochanter prolater- oventral apophysis of leg III and IV absent. Apex of accessory piece distally crenulated (Figs. 8, 9), basitarsus of leg III moderately inflated (longer than wide). Legs with numerous tubercles. Female paratype: Similar to male, except as follows. Body total length, excluding pygidium, 5.1, cucullus 0.8 long, greatest width 1.0, prosoma 1.9 long, 2.0 wide between legs II and III; opisthosoma 3.2 long; 2.7 wide near middle tergite; legs I 4.2, II 8.0, III 6.0, IV 6.1. Leg formula II > IV > III > 1. Opisthosoma: abdominal tubercles absent only in paramedian portion, lighter than other regions (Fig. 14). Cucullus: rounded tubercles concentrated on median portion and posterior margin (Fig. 12). Pygidium with “V” notch on dorsal margin of basal segment. Spermathecae wide and short, inner half trilobate, outer half acuminate (Fig. 15). Nymph paratype: Body total length, excluding pygidium 4.3, cucullus 0.5 long, greatest width 0.6, prosoma 1.3 long, 0.8 wide between legs II and III; opisthosoma 3.0 long; 2.8 wide near middle of tergite; legs I 3.1, II 5.8, III 3.6, IV 4.0. Leg formula II > IV > III > 1. General body color (in 70% ethanol) orange, intersegmental membranes dark yellow (dorsal view). Ventral opisthosoma lighter than prosoma, intersegmental membranes light yellow. Cucullus, legs (femur, patella and tibia of legs I and II: orange), sternal region and pedipalps (distal segment: dark yellow) light yellow. Carapace more densely covered with tubercles than male and female, absent on anterolateral border (Fig. 17). Cucullus slightly wider than long (Fig. 18), slightly depressed in posterior margin. Distribution. — This species is known only from the type locality (Fig. 20). Notes on the biotope. — Specimens of C. conori were collected at a location in the Solimoes formation (Araujo et al. 1978; Rosseti et al. 2005). Its topography is predominantly Bat, with tabular interfluvials and hills with Tertiary sedimentary deposits, known as “paleo-varzea” (Mauro et al. 1978). The soil is reddish-yellow (podzolic) and the vegetation is characterized as Dense Tropical Forest (Araujo et al. 1978; Doi 1978). This region has areas of savanna-like vegetation, spotted-like mosaic field islands distributed in large areas of continuous upland forest, and this landscape pattern contrib- 428 THE JOURNAL OF ARACHNOLOGY Figure 10. — Cryptocellus conori new species, female paratype, dorsal view. Scale bar = 1.0 mm. Figures 1 1-15. — Cryptocellus conori new species, female paratype: 11. Prosoma dorsal view; 12. Cucullus, dorsal view; 13. Tergite XI; 14. Opisthosoma, ventral view; 15. Spermathecae, ventral view. Scales bars 11-14 = 1.0 mm; scale bar 15 = 0.5 mm. utes directly to regional-scale diversity in the Amazon basin (Py-Daniel et al. 2007). This region is characterized by open vegetation, hosting a complex of several Amazonian natural fields, locally called “campinas”, which are very numerous in this interfluvial region of the Madeira and Purus Rivers. The fields and forests host several unique species and are regarded as one of the most diverse sites in the Amazon forest (Py- Danie! et al. 2007). Several species of animals and plants were collected during these campaigns. Along with C. conori several new species of harvestmen and spiders are still being processed in our laboratory at Institute Nacional de Pesquisas da Amazonia. New species of vertebrates and plants were collected on the same field trip: one new subspecies of saddleback tamarin, Saguinus fuscicolUs Rohe, Silva- Junior, Sampaio & Raylands, 2009 (R5he et al. 2009), a new species of palm has been diagnosed but not yet published (T. Emilio, pers. comm..), and a new species of crow, endemic to these local savanna-like spot fields in the interfluvials (M. Cohn- Haft, personal communication). All of these species are endangered by the construction of the BR-319 highway, which is planned to connect the city of Manaus (Amazonas State) to Porto Velho (Roraima State). The highway will link the Amazonas State with the deforesta- tion arc, the largest deforested area in the north of Brazil, including the states of Rondonia, Para, Acre and northern Mato Grosso and Tocantins, and it would be a focus for migration and unordered human occupation, illegal occupation of lands by land grabbers, and mechanized agricultural crops of soybean and rice. These factors together will accelerate the destruction of the forest along the highway, in several sites physically distant from the road, although under its influence (Fearnside 2005; Fearnside & Gra?a 2006; Fearnside et at. 2009). Along with the highway there are projects to build powerful hydroelectric power plants in this interfluvial region, which will contribute heavily to the environmental damage and degradation along the interfluvials, as has already occurred TOURINHO & SATURNINO—CRYPTOCELLUS PECKORUM AND AD!SI GROUPS 429 Figure 16. — Cryptocellus conori new species, paratype nymph, dorsal view. Scale bar = 1.0 mm. with the lake of Balbina, and the region where the Balbina power plant was built (Fearnside 1989). Cryptocellus becki Platnick & Shadab 1977 Cryptocellus becki Platnick & Shadab 1977:11, figs. 37-50. Type locality. — Brazil, Amazonas, near Manaus, Reserva Ducke. Figures 17-19. — Cryptocellus conori new species, paratype nymph: 17. Prosoma dorsal view; 18. Cucullus, dorsal view; 19. Opisthosoma, ventral view. Scales bars = 1.0 mm. Material examined. — Brazil; Amazoncis: Manaus, LBA Farm (Km 34, Br 174), 02°36'32"S, 60°12'32"W, 1 male, 12 August 2005, S.M. Ketelhut et ah; 1 female, 25 August 2006, F.B. Baccaro et ah; 1 nymph, 25 August 2006, F.B. Baccaro et ah; 1 nymph, 25 August 2006, F.B. Baccaro et ah; 2 nymphs, 16 May 2006, F.B. Baccaro et ah; 1 nymph, 15 November 2004, S.M. Ketelhut et ah; 3 nymphs, 8 August 2005, S.M. Ketelhut et ah Cabo Frio Farm, 02°25'S, 60°W, 1 nymph, 24 November 2005, S.M. Ketelhut et ah; 1 nymph, 24 July 2006, F.B. Baccaro et ah ZF 2 Farm (Km 14, Br 174), 02°35'20"S, 60°06'54"W, 3 nymphs, 20 October 2004, S.M. Ketelhut et ah; 1 nymph, 20 October 2004, S.M. Ketelhut et ah ZF 3 Farm (Km 37, Br 174), 02°35'55"S, 60°03'09"W, 1 nymph, 16 August 2005, S.M. Ketelhut et ah; 1 nymph, 16 August 2005, S.M. Ketelhut et ah; 1 nymph, 22 November 2005, S.M. Ketelhut et ah; 1 nymph, 20 July 2004, S.M. Ketelhut et ah; 1 nymph, 16 August 2005, S.M. Ketelhut et ah Distribution. — This species is now known from Reserva Ducke (type locality), Taruma-Mirim river (Adis et ah 1989), LBA Farm (Km 34, Br 174), ZF 2 Farm, ZF 3 Farm (Km 37, Br 174) and Cabo Frio Farm, located in Manaus municipality, Amazonas State, Brazil. Relationships. — The foeclus group was suggested by Platnick & Shadab (1977) as a monophyletic unit supported by the following synapomorphies: male cucullus with strong depres- sion below its dorsal margin; expanded male basitarsus 111; male with apophyses on trochanters of legs 111 and IV; and female spermathecae short and wide. For this study it is proposed that species included in this group also share a very inflated basitarsus (as long as wide) of leg III; a protuberance on the accessory piece of male leg III; and a large fixed male accessory piece forming an acute angle. Cryptocellus foeclus was said by Platnick &. Shadab (1977) to be closely related to the species C. peckorum, suggested as the plesiomorphic sister- taxon of C. foeclus. However, they stated that C. peckorum and C lampeli where not closely related to each other. Cryptocellus peckorum is herein included in a third assemblage of species, the peckorum group, which is composed of C. peckorum, C. tarsilcie, C. lampeli and the new species C. conori. The basic 430 THE JOURNAL OF ARACHNOLOGY 62”0'0'W erO'O’W 60°0'0"W 59°0'0'W Figure 20. — Type locality of Cryptocelliis conori new species, Careiro, Amazonas State, Brazil. synapomorphies supporting this group are fixed male acces- sory piece of leg III thin; both movable and fixed piece curved, and with a sinuous contour; basitarsus of leg III moderately inflated (longer than wide). The Cryptocelliis magmis group was first suggested by Platnick & Paz (1979) based on species that share a massive and straight accessory piece of the male tarsal process. This character was hypothesized to be a derived character also present in C. pseudocellatus, known only from female specimens. Cryptocelliis glenoides was originally included in the magmis group, as the most plesiomorphic lineage (Platnick & Paz 1979), but was later transferred to the centralis group which was based on a single apomorphic character, the presence of a peculiar anteroventral ledge on the tarsal process of male leg 111, which is the basic synapomorphy for the centralis group, composed of 10 species (Platnick & Shadab 1981). We suggest that C platnicki Botero-Trujillo & Perez 2008 also belongs in the centralis group, based on its close relationship with C. glenoides. However, evidence for further synapomorphies is needed in order to state precisely the position of C. glenoides + C. platnicki in the magmis or centralis group (Platnick & Shadab 1981). Among the five remaining species of Cryptocelliis, three of them are here hypothesized to form a fifth group, the adisi group: C. adisi, C. lisbetliae, and C. florezi. This group is based on a single shared character: the fixed accessory piece of male leg III is thin, curved and with a rounded contour. Platnick & Garcia (2008) suggested the last species is more closely related to C. lisbetliae, and that C. adisi resembles and could be related to C. albosqiiamatus {?Vdin\ck. 1988). Cryptocelliis alboscpiamatiis is known only from female specimens, and its relationships and placement inside the established groups can only be properly done when and if the male of this species is found. The current arrangement suggested in this paper for species included in the genus Cryptocelliis is: Adisi group Cryptocelliis adisi Platnick 1988 (Brazil, Amazonas, Tar- uma-Mirim river), known only from male. Cryptocelliis florezi Platnick & Garcia 2008 (Colombia, Department of Caqueta), known from male and female. Cryptocelliis lisbetliae Gonzalez-Sponga 1998 (Venezuela, Bolivar), known from male and female. Centralis group (Platnick & Shadab 1981) Cryptocelliis centralis Page 1921 (Costa Rica, Heredia, La Caja), known from male and female. Cryptocelliis cliirkjiii Platnick & Shadab 1981 (Panama, Chiriqui), known only from male. Cryptocelliis fagei Cooke & Shadab 1973 (Costa Rica, Golfito), known only from male. Cryptocelliis gamboa Platnick & Shadab 1981 (Panama, Canal Zone, Gamboa Pipeline), known from male and female. Cryptocelliis glenoides Cooke & Shadab 1973 (Colombia, Valle, 5 km of Delfina and Panama, Panama Province, Cerro Campana), known from male and female. Cryptocelliis goodnight i Platnick & Shadab 1981 (Costa Rica, Heredia, near Puerto Viejo de Sarapiqui), known only from male. Cryptocelliis hanseni Cooke & Shadab 1973 (Nicaragua and Honduras), known from male and female. Cryptocelliis isthmius Cooke & Shadab 1973 (Panama, Canal Zone, Gatun), known from male and female. Cryptocelliis hiisedieri Botero-Trujillo & Perez, 2009 (Co- lombia, Ipiales Municipality, Narino Department), known only from male. Cryptocelliis osa Platnick & Shadab 1981 (Costa Rica, Puntarenas, Peninsula of Osa), known from male and female. TOURINHO & SATURNINO— C/?KPTOC£LLt/5' PECKORUM AND ADISI GROUPS 431 Cryptocellus platnicki Botero-Trujillo & Perez 2008 (Co- lombia, Department of Choco) - dubious position (Botero- Trujillo & Perez 2008), known from male and female. Cryptocellus striatipes Cooke & Shadab 1973 (Costa Rica, Limon, Colombiana), known from male and female. Foedm group (Platnick & Shadab 1977) Cryptocellus abaporu Bonaldo & Pinto-da-Rocha 2003 (Brazil, Rondonia, Ji-Parana), known from male and female. Cryptocellus becki Platnick & Shadab 1977 (Brazil, Ama- zonas, near Manaus, Reserva Ducke), known from male and female. Cryptocellus foedus Westwood 1874 (Brazil, Amazon, somewhere between Belem in Para and Sao Paulo de Olivenga, in Amazonas), known only from female. Cryptocellus icamiabas Tourinho & Azevedo 2007 (Brazil, Amazonas, Presidente Figueiredo, Balbina), known only from male. Cryptocellus simonis Hansen & Sorensen 1904 (Brazil, Para, Belem), known from male and female. Cryptocellus whitticki Platnick & Shadab 1977 (Guyana, Rupununi, New River district), known only from male. Magnus group (Platnick & Paz 1979) Cryptocellus bordoni Dumitrescu & Juvara-Bals 1977 (Venezuela, Zulia), know from male and female. Cryptocellus brignolii Cokendolpher 2000 (Suriname, Para- maribo), know only from male. Cryptocellus magnus Ewing 1929 (Colombia, Magdalena), know only from female. Cryptocellus narino Platnick & Paz 1979 (Colombia, Antioquia), known from male and female. Cryptocellus pseudocellatus Roewer 1952 (Peru, Cajamarca), known only from female. Peckorum group Cryptocellus conori new species (Brazil, Amazonas, Car- eiro), known from male and female. Cryptocellus lampeli Cooke 1967 (British Guiana, Amatuk), known from male and female. Cryptocellus peckorum Platnick & Shadab 1977 (Colombia, Amazonas, Leticia), known from male and female. Cryptocellus tarsilae Pinto-da-Rocha & Bonldo 2007 (Brazil, Para, National Forest (FLONA) of Carajas), known from male and female. Nomina dubia Cryptocellus leleupi Cooreman 1977 (Platnick & Paz 1979) (Ecuador, Oriente, Rio Negro), know only from nymph. Cryptocellus emcirginatiis Ewing 1929 (Platnick & Shadab 1981) (Costa Rica, Cartago, Navarro farms), know only from nymph. Remaining species Cryptocellus albosquamatus Cooke 1967 (British Guiana, Amatuk), known only from female. Cryptocellus bocas Platnick & Shadab 1981 (Panama, Quebrada Alicia), known only from female. Cryptocellus verde Platnick & Shadab 1981 (Costa Rica, Puntarenas, Monteverde), known only from female. ACKNOWLEDGMENTS We thank Eduardo Wienskoski and Laura Miglio for collecting the specimens used in this study, and Osvaldo Villarreal for supplying a copy of G. Sponga’s paper. Alexandre Bonaldo helped with taxonomic questions. Au- gusto Henriques, Celio Magalhaes and Jose Rafael (INPA) provided laboratory facilities. Erika Portela helped with the database and sorting material. Renata Frederico prepared the map presented in this paper. Suzana M. Ketelhut and Jose Wellington gave us C. becki specimens from the TEAM network project. The authors are supported by grants: ALT Research Program for Biodiversity (PPBIO) research grant from MCT and RS # 382.072/2007-3 from CNPq (PCI/INPA/ PPBIO). This paper is an outcome of the project “Biodiversity of the Madeira-Purus Interfluve”, an activity of the GEOMA network of Brazil’s Ministry of Science and Technology. Field expeditions were supervised by Mario Cohn-Haft. LITERATURE CITED Adis, J.U., N.I. Platnick, J.W. Morais & J.M.G. Rodrigues. 1989. On the abundance and ecology of Ricinulei (Arachnida) from Central Amazonia, Brazil. Journal of the New York Entomological Society 97:133-140. Araujo, H.J.T., J.B.M. Rodarte & J.O. Del’arco. 1978. Geologia. Pp. 17-128. In Brasil. Departamento Nacional de Produpao Mineral. Projeto RADAMBRASIL. Folha SB-20 Purus. Geologia, geomorfologia, pedologia, vegetaijao e uso potencial da terra. Levantamento de recursos naturais, NME/DNPM, Rio de Janeiro. Bonaldo, A.B. & R. Pinto-da-Rocha. 2003. On a new species of Cryptocellus from the Brazilian Amazon (Arachnida, Ricinulei). Revista Iberica de Aracnologia 7:103-108. Botero-Trujillo, R. & G.A. Perez. 2008. A new species of Cryptocellus (Arachnida, Ricinulei) from northwestern Colombia. Journal of Arachnology 36:468-471. Botero-Trujillo, R. & G.A. Perez. 2009. A new species of Cryptocellus (Arachnida, Ricinulei) from the K6fan Territory in southwestern Colombia. Zootaxa 2050:56-64. Cokendolpher, J.C. 2000. First Cryptocellus from Suriname (Ricinu- lei). Memorie della Societa Entomologica Italiana 78:515-520. Cokendolpher, J.C. & T. Enriquez. 2004. A new species and records of Pseudocellus (Arachnida: Ricinulei: Rieinoididae) from caves in Yucatan, Mexico and Belize. Texas Memorial Museum, Speleo- logical Monographs 6:95-99. Cooke, J.A.L. 1967. Observations on the biology of Ricinulei (Arachnida) with description of two new species of Cryptocellus. Journal of Zoology 151:31M2. Cooke, J.A.L. & M.U. Shadab. 1973. New and little known ricinuleids of the genus Cryptocellus (Arachnida, Ricinulei). American Museum Novitates 2530:1-25. Doi, S. 1978. Vegetagao. Pp. 25-160. In Brasil. Departamento Nacional de Produgao Mineral: Projeto RADAMBRASIL. Folha SB 20 Purus. Rio de Janeiro. Fearnside, P.M. 1989. Brazil’s Balbina Dam: Environment versus the legacy of the pharaohs in Amazonia. Environmental Management 13:401M23. Fearnside, P.M. 2005. Carga pesada: O custo ambiental de asfaltar um corredor de soja na Amazonia. Pp. 397M23. In Amazonia revelada: Os descaminhos ao longo da BR-163. (M. Torres, ed.), Conselho Nacional de Desenvolvimento Cientifico e Tecnoldgico (CNPq), DF. Brasilia. Fearnside, P.M. & P.M.L.A. Graga. 2006. BR-319: Brazil’s Manaus- Porto Velho Highway and the potential impact of linking the arc of deforestation to central Amazonia. Environmental Management 38:705-716. Fearnside, P.M., P.M.L.A. Graga, E.W.H. Keizer, F.D. Maldonado, R.I. Barbosa & E.M. Nogueira. 2009. Modelagem de desmata- mento e emissoes de gases de efeito estufa na regiao sob influencia da Rodovia Manaus-Porto Velho (BR-319). Revista Brasileira de Meteorologia 24:208-233. 432 THE JOURNAL OF ARACHNOLOGY Gonzalez-Sponga, M.A. 1998. Aracnidos de Venezuela. Cryptocellus lishethae nueva especie de Ricinulei del estado Bolivar (Arachnida, Ricinulidae). Memoria Sociedad de Ciencias Naturales La Salle 57:49-54. Mauro, C.A., B.T.A. Nunes & M.S.M. Franco. 1978. Geomorfolo- gia. Pp. 161-215. In Brasil. Departamento Nacional de Produgao Mineral. Projeto RADAM BRASIL. Folha SB-20 Purus. Geologia, geomorfologia, pedologia, vegeta^ao e uso potencial da terra. Levantamento de recursos naturals. NME/DNPM, Rio de Janeiro. Pinto-da-Rocha, R. & A.B. Bonaldo. 2007. A new species of Cryptocellus (Arachnida, Ricinulei) from Oriental Amazonia. Zootaxa 1386:47-51. Platnick, N.I. 1980. On the phytogeny of Ricinulei. Pp. 349-353. In Verhandlungen des 8. Internationalen Arachnologen-Kongress, Wien. (J. Gruber, ed.). H. Egermann, Wien. Platnick, N.I. 1988. A new Cryptocellus (Arachnida: Ricinulei) from Brazil. Journal of the New York Entomological Society 96:363-368. Platnick, N.I. & L.F.H. Garcia. 2008. Taxonomic notes on Colombian Cryptocellus (Arachnida, Ricinulei). Journal of Arach- nology 36:145-149. Platnick, N.I. & G. Pass. 1982. On a new Guatemalan Pseudocellus (Arachnida, Ricinulei). American Museum Novitates 2733:1-6. Platnick, N.I. & N. Paz. 1979. On the Cryptocellus niagnus group (Arachnida, Ricinulei). American Museum Novitates 2677:1-9. Platnick, N.I. & M.U. Shadab. 1976. On Colombian Cryptocellus (Arachnida, Ricinulei). American Museum Novitates 2605:1-8. Platnick, N.I. & M.U. Shadab. 1977. On Amazonian Cryptocellus li (Arachnida, Ricinulei). American Museum Novitates 2633:1-17. ! Platnick, N.I. & M.U. Shadab. 1981. On Central American ! Cryptocellus (Arachnida, Ricinulei). American Museum Novitates 2711:1-21. Py-Daniel, L.R., C.P. de Deus, A.L. Henriques, D.M. Pimpao & O.M. | Ribeiro. 2007. Biodiversidade do Medio Madeira. Bases Cientificas para | Propostas de Conserva^ao. Manaus: INPA (Brasilia): MMA: MCT. Rohe, F., J.S.E. Silva-Junior, R. Sampaio & A.B. Rylands. 2009. A new subspecies of saddleback tamarin, Saguinus fuscicollis (Primates, Callitrichidae). International Journal of Primatology 30:533-551. Rossetti, D.F., P.M. Toledo & A.M. Goes. 2005. New geological framework for Western Amazonia (Brazil) and implications for biogeography and evolution. Quaternary Research 63:78-89. Talarico, G., J.G. Palacios- Vargas, M.F. Silva & G. Alberti. 2005. | First ultrastructural observations on the tarsal pore organ of Pseudocellus pearsei and P. boneti (Arachnida, Ricinulei). Journal of Arachnology 33:604-612. Talarico, G., J.G. Palacios-Vargas, M.F. Silva & G. Alberti. 2006. Ultrastructure of tarsal sensilla and other integument structures of two Pseudocellus species (Arachnida, Ricinulei). Journal of Morphology 267:441^63. Tourinho, A.L. & C.S. Azevedo. 2007. A new Amazonian Crypto- cellus Westwood (Arachnida, Ricinulei). Zootaxa 1540:55-60. Manuscript received 2 December 2009, revised 20 March 2010. 2010. The Journal of Arachnology 38:433-440 Patterns of abundance, habitat use and body size structure of Phoneutria reidyi and P. fern (Araneae: Ctenidae) in a Central Amazonian rainforest Martha Patricia Torres-Sanchez; Corporacion Propuesta Ambiental, Departamento de Ciencias Naturales, Universidad Central, Carrera 5 niimero 21-38, Bogota, D.C., Colombia. E-mail: mtorress@ucentral.edu.co Thierry Ray Gasnier: Departamento de Biologia/ ICB, Universidade Federal do Amazonas, Av. Gal. R.O.J. Ramos 3000, CEP 69067-000, Manaus, AM, Brazil Abstract. Phoneutria is one of the most medically important spider genera; however, its ecology is poorly known. In Amazonian upland rainforests, there are two sympatric species of the genus Phoneutria, P. reidyi (F.O. Pickard-Cambridge 1897) and P. fera Perty 1833. For 15 months we collected data on the spatial distribution, use of habitat (activity on the ground or vegetation) and temporal changes in body size structure in a forest reserve near Manaus city in three distinct habitats: dense forest, located on the plateaus on yellow latossol; swamp forest, located in the valleys; and heath forest or “campinarana,” on dry white sand soil in the Reserva Florestal Ducke. A total of 239 P. reidyi and 239 P. fera were captured in nocturnal searches during their period of activity. There were significant differences between the two species: 1) P. reidyi existed in higher abundance in the swamp forest than in the dense forest areas and was almost absent in the heath forest, while P. fera existed in similar abundance in the three habitats. 2) During their development, members of both species used the vegetation as an area of activity, but the subadults and adults of P. reidyi were less often found on the ground than the subadults and adults of P. fera. 3) P. reidyi more frequently used small or acaulescent palms as a substrate, and its abundance was directly related to the abundance of these palms, independent of the habitat, while P. fera did not show such relationship. 4) There was a strong temporal variation in the body size structure of the P. reidyi population indicating seasonal reproduction, but there was no evidence of seasonal reproduction by P. fera. We suggest that the differences in the use of habitat and in the seasonality of reproduction are related to the avoidance of intraguild predation between these species. Keywords: Amazon, “banana-spider,” coexistence, life cycle, microhabitat preferences, wandering spiders Spiders of the genus Phoneutria are large nocturnal hunting species common in South American forests and in synan- thropic areas close to these forests (Lucas 1988; Folly-Ramos et al. 1998). They are considered aggressive and among the medically most important spiders in the world, on the basis of the number of serious human accidents (Maretic 1987; Lucas 1988, 2002; Vetter & Isbister 2008). Despite their medical importance, little is known about their ecology, although they are frequently abundant in forests where they occur. Biicherl (1969, 1980), Biicherl et al. (1969), Lucas (1969, 1988), Folly-Ramos et al. (1998), and Almeida et al. (2000) presented data on development, activity, seasonality of reproduction, and habitat use of Phoneutria nigriventer (Key- serling 1891). In southeastern Brazil, Biicherl (1969) and Lucas (1988) found that Phoneutria keyserlingi (F.O. Pickard-Cam- bridge 1897) and P. nigriventer reach maturity in three years, with a marked seasonal reproduction. These studies were based mainly on synanthropic populations and on animals kept in captivity. Until recently, nothing has been published on the ecology of the Amazonian species Phoneutria fera Perty 1833 and Phoneutria reidyi (F.O. Pickard-Cambridge 1897), although they have a large distribution in most “terra-firme” forests (Simo & Brescovit 2001). The present paper is part of the first intensive ecological research on the genus in undisturbed areas. In previous publications, we presented results on temporal variation in adult size and sexual dimorphism and notes on natural history (Gasnier et al. 2002; Gasnier et al. 2009). Phoneutria reidyi and P. fera are among the most abundant large spider species in Amazonia. Healthy adults possibly have few enemies, but the juveniles are probably under constant risk. Most potential predators, such as army ants, spiders of the genus Ctenus and frogs, are more abundant on the ground (Vieira & Hofer 1994; Gasnier & Hofer 2001; Menin et al. 2008); therefore, the vegetation may be an important substrate, particularly for juveniles, to avoid predation. Both Phoneutria species are frequently found on plants with large leaves, especially acaulescent palms that may attract the spiders because their leaves can sustain their weight and transmit vibrations efficiently (Barth et al. 1988), or because of a higher availability of prey associated with the litter at its base (Vasconcelos 1990). Differences in the use of habitat, phenology, and general behavior (including prey capture and reproductive behavior) are factors that contribute to the coexistence of spiders (Enders 1976; Uetz 1977; Turner & Polish 1979; Uetz 1991; Cutler & Jennings 1992; Polls & Holt 1992; Wise 1993; Morse 1997; Wise & Chen 1999; Denno et al. 2004; Rypstra & Samu 2005). Areas with the sympatric occurrence of two or more species are not common in the genus (Martins & Bertani 2007); only P. reidyi and P. fera have a large area of coexistence, apparently most of the Amazonian Region (Simo & Brescovit 2001). Our purpose was to study certain aspects of the ecology of P. reidyi and P. fera in a forest area, including factors that affect their abundance, use of habitat and life history, and to furnish the basis for understanding their coexistence. METHODS Study site and species. — The study was conducted in the Reserva Florestal Ducke, a 10,000 ha primary rainforest 433 434 THE JOURNAL OF ARACHNOLOGY reserve of the Institute Nacional de Pesquisas da Amazonia (INPA). The fauna and flora of this reserve is one of the most heavily studied in Amazonia (Ribeiro et al. 1999; Adis 2002), including the spider fauna (H5fer & Brescovit 2001). We worked in the northern part of the reserve in the basins of the Barro Branco and Acara streams (2°55'00"-2°56'45"S, 59°57'08"-59°58'41"W). Well-drained clay soils in plateaus and slopes predominate in the basin of Barro Branco stream, with a dense forest (descriptions of vegetation in Guillaumet 1987), and well drained white sandy soils predominate in the basin of Acara stream, with a heath forest or “campinarana” offering a more open canopy than the dense forest. In both areas the soil is sandy and hydromorphic close to the streams, with vegetation called swamp forest or “baixio.” In all areas the understory vegetation is relatively open. A description of the study area is presented in Gasnier & Hofer (2001). The average annual temperature is 25.6° C, and the average annual rainfall is 2480 mm. The rainy season occurs between December and May, with the rainiest months in March and April and the driest months from July to September (Marques-Filho et al. 1981). We made our observations and collections between June 1998 and August 1999. We captured the spiders and observed them at night, with the help of headlamps strong enough to allow for observing spiders’ eye reflection (including small individuals) up to a distance of approximately 15-20 m. The body of the spider is visible up to about 3-8 m, depending of the spider’s size. The spiders could be located at a considerable height (up to 5 m), but were only captured at heights lower than 3 m. To avoid the effect of rain on the abundance estimates, we did not include data from nights with rain and nights after days with rain. We captured the spiders with glass or plastic vials proportional in size to the spiders (22-80 mm in diameter, 60-140 mm in height, and 20-60 mm opening) and preserved the specimens in 70% ethanol. The material is deposited in the Entomolog- ical Collection of INPA. We initially identified the species in the laboratory based on the reproductive structures of adults (palps and epigyne). However, we realized that the ventral and dorsal color patterns of the body and stripes of the palps allowed us to discriminate among species in the field, including very small juveniles. The patterns are described and illustrated in Martins & Bertani (2007). Censuses. — We used two types of census: one extensive census to evaluate variation in abundance of spiders in a large area including different habitats, and the other to compare abundance between dense forest on plateaus and swamp forest in valley areas. The extensive census consisted of 60 plots of about 50 X 10 m, separated by a distance of 100 m between plots in a trail of about 9 km inside an area of about 2X5 km. We collected spiders four times in each plot in June and October 1998, and in April and August 1999. One person searched for spiders from 0 to 3 m height for approximately 30 minutes in each plot. To investigate the influence of environmental factors on the abundance of the spiders, we counted the number of small or acaulescent palms (> 1 m diameter), collected samples of soil in each plot and calculated the volume of leaf litter. We measured the volume of leaf litter on the ground in the middle of the plot, placing the leaves from a 1-m^ area in a graduated container. The volume from each plot was the mean of the measures made on two occasions. Since the plots of the extensive census included few places with hydromorphic soils, we performed additional censuses comparing the dense forest areas on plateaus with the swamp forest areas on hydromorphic soils close to streams. We called these surveys plateau-swamp censuses. In these censuses, we searched for spiders in 15 plots on the plateau and in 15 plots in the swamp forest. The effort was standardized as a 2-h search in each plot by one person. The plateau-swamp censuses were made on four occasions, in August and November 1998 and in January and April 1999, but we collected spiders only once in each plot. Spiders collected apart from the censuses (on other occasions or by someone following behind the person searching) were not considered in the comparisons of abundance among habitats, but were included in the analysis of vertical distribution and population structure. For each spider collected, we recorded the species, size (prosoma length), sex (if adult), and the type and height of the substrate. On the first excursion, small juveniles were not collected because we could not identify small Phoneutria. Statistical Analyses. — The analysis of abundance in a sequence of plots helps to evaluate variation at different scales when the sample unit is not discrete and natural (Ludwig & Reynolds 1988), and to propose hypotheses on factors that affect the abundance. The number of spiders in each plot was low (up to 5); therefore, we used non-parametric tests (Mann- Whitney U-test and Spearman rank correlation, Tv). The linear regression between the number of spiders and the number of palms was modeled with Reduced Major Axis because the values of the independent variable were random (Sokal & Rohlf 1995). The interspecific correlation of P. reidyi and P. fera was tested with Spearman rank correlation, and the interspecific association was tested with Yates corrected test in a 2 X 2 contingency table with the presence/absence of both species (Ludwig & Reynolds 1988). We used the equality of proportions test to verify whether proportions of spiders on the ground or vegetation differed between species. The seasonality of reproduction was evaluated with a Kolmo- gorov-Smirnov test, comparing the distribution of body sizes graphically and testing the difference in sizes of juveniles in October 1998 and April 1999, because these were months with large data sets and were six months apart. The statistical package used was SYSTAT 12© (Wilkinson 1990). RESULTS Abundance Patterns. — A total of 239 P. reidyi, (181 juveniles, 36 adult males, and 22 adult females) and 239 P. fera (respectively 181, 27 and 31) was recorded. However, the proportions of species were different between the two types of censuses. In the 60 plots on the trail of the extensive census, made in areas where well-drained soils predominated, 56 spiders were P. reidyi and 108 were P. fera. In the plateau- swamp census, made in 15 areas of well-drained soil and 15 areas of swamp forest, 1 36 spiders were P. reidyi and 55 were P. fera. The spiders collected outside the censuses were 47 P. reidyi and 16 P. fera. In the extensive census, the patterns of abundance were similar within each species, and different between species in TORRES & GASNIER— ECOLOGY OF AMAZONIAN PHONEUTRIA 435 80 40 50 a. >, 'o so Number of plot (along a 9 km trail) Figure 1. — Superposition of graphs of abundance, showing both species and characteristics of the habitat (number of palms, volume of leaf litter, and type of soil) along the line of extensive census. Each bar corresponds to a transect of 50 m. The soils were categorized as sandy (white), hydromorphic (gray) and clayey (black), and they are closely related to the vegetation types (see Methods). the four counts throughout one year. P. reidyi's pattern was marked by a high abundance (1.36 spiders/100 m) in positions 1-39 and low abundance (0.14/100 m) in the basin of Acara stream (positions 40-60 in Fig. 1), in an area where heath forest over sandy soils predominated. This zone had lower abundances of small or acaulescent palms, but similar amounts of leaf litter compared to the other areas. Compared to P. reidyi, P. fera had a relatively homogeneous abundance along the sequence of plots (1.8/100 m). In addition, we compared the abundances of spiders between the habitats “dense forest” and “swamp forest”, using the plateau-swamp census samples. P. reidyi was more abundant in the swamp forest (Mann-Whitney U-test, U = in, P = 0.036), where palms predominate. The abundance of P. fera did not differ significantly between these two habitats (Mann-Whitney (7-test, U = 61, P = 0.051), with a tendency for smaller numbers in the swamp forest. This difference suggested that the abundance of small or acaulescent palms was an important factor. To test the correlation between abundance of palms and abundance of spiders, we used data from the extensive census samples. There was a positive correlation between the number of P. reidyi and the number of palms (Spearman rank correlation, r^ = 0.63, n = 60, P < 0.001; Fig. 2). As shown above, this relationship is in part a result of differences between habitats for the abundance of palms and the abundance of P. reidyi, which is weak evidence of a causal relationship. Therefore, we performed two additional tests for the relation- ship between the number of P. reidyi and the number of palms, separated by habitat, only in areas with dense forest (Spearman rank correlation, r, = 0.67, n = 2S, P < 0.001 ) and only in areas with heath forest vegetation (Spearman rank correlation, r., = 0.50, n = 25, P = 0.003) (sample size in swamp forest was low for this test). These tests show that this positive correlation is 436 THE JOURNAL OF ARACHNOLOGY 0 JD E 3 2 a: 0 XJ E 3 Figure 2. — Relationship between the abundance of palms and the number of spiders in each sample unit. The relationship was significant only for P. reidyi, for which a regression line is shown (Reduced Major Axis Regression, Y = -0.671 + 0.0545Y). independent of the type of habitat for P. reidyi. The number of P. [era and the number of palms did not differ significantly in the extensive census plots (Spearman rank correlation, = -0.14, n = 60, T = 0.28). We tested the correlation between leaf litter volume and the number of spiders using the extensive census data because the abundance of prey and availability of refuges for spiders may be affected by the volume of leaf litter. We found no TORRES & GASNIER— ECOLOGY OF AMAZONIAN PHONEUTRIA 437 Substrate ■ Palm la Other plants □ Ground 10 15 Prosoma length (mm) Figure 3. — Frequency of size (prosoma length) categories for P. reidyi and P. fera found on three types of substrate. significant relationship between the number of spiders and the amount of leaf litter for P. reidyi (Spearman rank correlation, = 0.15, « = 60, P = 0.23) or P. fera (Spearman rank correlation, = -0.05, n = 60, P = 0.72). To evaluate the interspecific association between these species, we used the sample units of the extensive census (n = 60) and of the plateau-swamp census {n = 30). There was no significant correlation between the abundance of P. reidyi and P. fera (Spearman rank correlation, = -0.04, n = 90, P = 0.72), nor in the presence/absence (P/A) of the species (PP = 51, PA = 24, AP = 11, AA = 4; r = 0.01, P = 0.91). Use of the habitat. — Use of habitat changed substantially during the development of both spider species. Juveniles with prosoma length < 9 mm were only found on the vegetation, and spiders gradually increased their use of the ground until they reached adult size (Fig. 3). However, there were differences between species. Use of the vegetation was still preponderant for P. reidyi with PL > 12mm, while most P. fera of this size were found on the ground. Furthermore, P. reidyi of all spider sizes used small or acaulescent palms much more frequently than other plants, while medium and large P. fera were found on both palms and other plants in similar frequency. A higher proportion of P. fera with PL >12 mm was found on the ground (Proportions test, 45% of 88 P. reidyi and 67% of 121 P. fera', Z = 3.10, P = 0.002). This difference was independent of the habitat because it was still significant when we restricted the data to swamp forest (Proportions test, 35% of 23 P. reidyi and 75% of 12 P. fera', Z = 2.25, P = 0.02) or dense forest (Proportions test, 49% of 61 P. reidyi and 68% of 88 P. fera', Z = 2.33, P = 0.02) (only 4 P. reidyi were found in the heath forest, which was not sufficient for a comparison). The height at which spiders were found on the vegetation (i.e., height > 0 cm) did not differ significantly between adult males (median = 120 cm, Q25 = 40, Q75 = 150) and females (median = 45 cm, Q25 = 37, Q75 = 57) of P. reidyi (Mann- Whitney f/-test, U - 109, P = 0.08) or between adult males (median = 30 cm, Q25 = 19.5, Q75 = 49.5) and females (median = 40 cm, Q25 = 12, Q75 = 1 13) of P. fera (Mann- Whitney U-test, U = 40, P = 0.74). There was no difference in the height of juveniles between P. reidyi (median = 85 cm, Q25 = 49, Q75 =150) and P. fera (median = 71 cm, Q25 = 40, Q75 = 134) (Mann-Whitney U-test, U = 10,024, P = 0.13); nor 438 THE JOURNAL OF ARACHNOLOGY 25 20 Phoneutria reidyi E 0 00 £ •00 ••• 0 0 00 00 •00 0 0 Female 1 5 - •••••• *Q 000000 0^0 •• •• O) •••00 •• 000000 00 000 •0 • Male <D •••• 000000000 00 000 0 •••• 00000 000000 00000 0 Juvenile CO 0^^ 00000 00000 0 0000 E 10 00000 0000 0 00 0000 o 000000 00 000 0000 (/) Q 00 00 000 000000 0 000 00 00 000000 0000000 □_ 00 000 0000000 000000 5 — 00000 0 000000000 000000 0 0 000000 0000 Jun Aug Oct Jan Apr Aug 25 Phoneutria fera E 20 E o) 15 - c CO E 10U o tn o 5 •0000 00 •ooo 00 00 •0000 ••00 •0 •00 ••0 ••oo ••0000 • •00 00^ 0 •••••••oo ooooooo*^ 000 00 0 000000000 0 000000000 0000 0 00000000 000000000 oooooo 000 0 000000000 0 oooooo 000 00 000000000 0 oooooo 000 0 00 0000000 00000 oooo 0 00 oooooo o oooo 000 00 000 000 00 0 0000 000 0 00 00 00 00 00 oo 0 00 oo 0 00 Ol I I J I I L_J Jun Aug Oct Jan Apr Aug 1998 1999 Figure 4. — Distribution of sizes of P. reidyi and P. fera on each of six collection dates during two years. between females of the two species (Mann-Whitney U-test, U = 42, P = 0.56). However, male P. reidyi were found in higher places than male P. fera (Mann-Whitney U test, U = 169, P < 0.001). Structure of body size. — There was strong temporal varia- tion in the structure of body size of P. reidyi (Fig. 4). Comparing data from October 1998 and April 1999, we found a significant difference (Kolmogorov-Smirnov test, D = 0.57, P < 0.001). This variation indicates a marked seasonal reproduction for P. reidyi. Another indication of seasonal reproduction was the high proportion of males captured in June 1998 (23d<J: 49?), most of them (93%) being on the ground, which, in this species, could indicate that they were searching for females. In P. fera, the structure of sizes remained relatively constant during the study (Fig. 4). Comparing data from October 1998 and April 1999, we found no significant difference (Kolmo- gorov-Smirnov test, D = Q.\5, P - 0.58). Apparently, these spiders reproduce throughout the year. Another indication of continuous reproduction in P. fera was that the sex ratio did not change much throughout the year, and both males and females were found on the ground in similar proportions (56% and 45% respectively). DISCUSSION The abundance of small and acaulescent palms correlated with the abundance of P. reidyi within habitats; consequently, this factor may explain why this species is abundant in swamp forests, where the palms are abundant and why they are rare in the heath forest, where the palms are almost absent. The association between Phoneutria and plants with large leaves, as we have found for P. reidyi, has been documented in previous TORRES & GASNIER— ECOLOGY OF AMAZONIAN PHONEUTRIA 439 papers (Schiapelli & Gerschman 1972; Biicherl 1980; Lucas 1988). Therefore, more remarkable was the absence of a relationship between the abundance of P. fera, collected in similar numbers in the same area, with the abundance of palms. We propose that the preponderance in use of the vegetation by juveniles of both species is an adaptation to avoid their common ground predators, and that the adults and subadults of P. reidyi use the vegetation more frequently, mainly to avoid predation by adults and subadults of P. fera on the ground. Barth et al. (1988), working with spiders of the genus Cupiennius, found that the palms are probably a safe place to stay because a spider is less visible and more able to sense vibrations of an approaching predator. Small juvenile P. fera were almost always on the vegetation; only individuals with a prosoma length greater than 8 mm were found on the ground. This is about the mean size of adults of two important potential predators, Ctenus amphora Mello-Leitao 1930 and Ctenus crulsi Mello-Leitao 1930, the most abundant medium- sized wandering spiders on the ground (Gasnier et al. 2002). Once P. fera spiders grow larger, their dispersal on the ground may be an advantage, because the Ctenus spiders will be smaller than them and thus become potential prey. We made no observations of Ctenus preying on Phoneutria, probably because small individuals were rare on the ground, but similar wandering spiders are among the main prey of Ctenus, and we observed that Ctenus was among the main prey of medium- sized to large Phoneutria. A similar pattern was noted with P. reidyi, but they also have P. fera as a larger predator on the ground. The mean prosoma length of adult P. reidyi (14.4 mm in males and 15.9 mm in females: Gasnier et al. 2002) is inferior to the mean prosoma length of P. fera (16.2 mm in males and 16.8 mm in females). Smaller spiders are probably more vulnerable to intraguild predation (Polis et al. 1989) and, consequently, less exposed (Johnson & Sih 2007), which would explain the more intensive use of the vegetation by P. reidyi. Direct evidence from a field experiment is necessary to demonstrate cases of coevolutionary divergence (Connell 1980) like the differentiation in use of the habitat suggested above. Meanwhile, there are additional arguments and evidence to sustain the hypothesis that the interaction between these two species, and possibly among other species of Phoneutria, is a relevant factor in their ecology. 1) The absence of a negative association between these sympatric species is not incompatible with this hypothesis because it is possible that coexistence in the present may have been facilitated by differentiation in the use of vegetation under intraguild predation pressure in the past. 2) The seasonality in reproduction of P. reidyi is consistent with this hypothesis because males searching for females probably have to disperse more frequently on the ground, and limiting this behavior to a part of the year could be an adaptation to prevent predation by P. fera. 3) P. boliviensis (F.O. Pickard-Cambridge 1897) is another Amazonian species with the smallest adults of the genus in central Amazonia. We have never seen it in “terra firme” forests, but it is relatively common in periodically inundated forests (T. Gasnier, pers. obs.). 4) Martins & Bertani (2007) showed that sympatry is practically restricted to the distribution limits of four species of Phoneutria in southeastern Brazil (only P. pertyi [F.O. Pickard-Cambridge 1897] was sympatric with other species). These contiguous allopatry patterns do not prove, but are consistent with, the importance of intraguild predation in the genus. Field experiments and study of the use of the habitat by sympatric and allopatric populations of Phoneutria species are necessary to verify these hypotheses and understand the ecology and evolution of this genus. Phoneutria may be considered key species in the forests where they occur. Most studies on Phoneutria were based on few individuals, probably because adults are not easy to find (Almeida et al. 2000). However, much data on abundance, distribution, and behavior may be acquired when juveniles are included in the study (e.g., Folly-Ramos et al. 1998; present study). Therefore, together with other ctenids, these spiders may be good indicators of disturbance in forest fragments (Jocque et al. 2005). The ecology of Phoneutria is not only important because of its medical importance. Comparative studies may also help to understand the importance of these predators in Neotropical forests. ACKNOWLEDGMENTS This paper is part of the thesis of the first author in the post- graduation program of the Agreement between the Instituto Nacional de Pesquisas do Amazonas (INPA) and the Universidade Federal do Amazonas (UFAM). Hubert Hdfer, Lucile Anthony, Antonio Brescovit, Pedro Gnaspini, Heraldo Vasconcelos, and Renato Cintra contributed much during the evaluation of the thesis. Stano Pekar, two anonymous reviewers, and Elizabeth Franklin provided many insightful comments on the preliminary version of the text. Steve Chilson reviewed the English. Clarissa de Azevedo and the personnel of the “Reserva Ducke” made our work easier with the warmth of their greeting, assistance, and conviviality. Financial support came from fellowship from the Brazilian institution CNPq and partial fellowship from the Colombian institution COLFUTURO (for Torres-Sanchez) and field support grants from CNPq (project. 400023/98). LITERATURE CITED Adis, J. (ed.). 2002. Amazonian Arachnida and Myriapoda.- Identification Keys to All Classes, Orders, Families, Some Genera, and Lists of Known Terrestrial Species. Pensoft Publishers, Sofia, Bulgaria. Almeida, C.E., E.F. Ramos, E. Gouvea, M. Carmo-Silva & J. Costa. 2000. Natural history of Ctenus mediiis Keyserling, 1891 (Araneae, Ctenidae) I: observations on habitats and the development of chromatic patterns. Revista Brasileira de Biologia 60:503-509. Barth, F.G., E.A. Seyfarth, H.J. Blekcmann & W. Schuch. 1988. Spiders of the genus Cupiennius Simon 1891 (Araneae, Ctenidae). I. Range distribution, dwelling plants, and climatic characteristics of the habitats. Oecologia 77:187-193. Biicherl, W. 1969. Biology and venoms of the most important South American spiders of the genera Phoneutria, Loxosceles, Lycosa and Latrodectus. American Zoologist 9:157-159. Biicherl, W. 1980. Aculeos que Matam. Quarta edicao. Livraria Kosmos Editora, Sao Paulo. Biicherl, W., S. Lucas & V. Eickstedt. 1969. Spiders of the family Ctenidae, Subfamily Phoneutrinae. IV. Bibliografia Phoneu- triarum. Memorias do Instituto Butanta 34:47-66. Connell, J.H. 1980. Diversity and coevolution of competitors, or the ghost of competition past. Oikos 35:131-138. Cutler, B. & D.T. Jennings. 1992. Habitat segregation by species of Metaphidippus (Araneae: Salticidae) in Minnesota. Journal of Arachnology 20:88-93. 440 THE JOURNAL OF ARACHNOLOGY Denno, R.F., M.S. Mitter, G.A. Langellotto, C. Gratton & D.L. Finke. 2004. Interactions between a hunting spider and a web- builder: consequences of intraguild predation and cannibalism for prey suppression. Ecological Entomology 29:566-577. Enders, F. 1974. Vertical stratification of orb- web spiders and a consideration of other methods of coexistence. Ecology 55:317-328. Folly-Ramos, E., C.E. Almeida, E. Gouvea & M. Carmo-Silva. 1998. Consideraqdes sobre a atividade de locomogao, preferencia por ecdtopos e aspectos territorials de Phoneutria nigriventer (Key- serling, 1891) (Araneae, Ctenidae). Revista Brasileira de Biologia 58:71-78. Gasnier, T.R. & H. Hofer. 2001. Patterns of abundance of four species of wandering spiders (Ctenidae: Ctemis) in a forest in Central Amazonia. Journal of Arachnology 29:95-103. Gasnier, T.R., C.S. de Azevedo, M.P. Torres-Sanchez & H. Hofer. 2002. Adult size of eight hunting spider species in central Amazonia: temporal variations and sexual dimorphisms. Journal of Arachnology 30:146-154. Gasnier, T.R., H. Hofer, M.P. Torres-Sanchez & C.S. de Azevedo. 2009. Historia natural de algumas especies de aranhas das familias Ctenidae e Lycosidae na Reserva Ducke: Bases para um modelo integrado de coexistencia. Pp. 223-230. In A Fauna de Artropodes da Reserva Florestal Ducke. Estado Atual do Conhecimento Taxonomico e Biologico. (C.R.V. Fonseca, C. Magalhaes, J.A. Rafael & E. Franklin, eds.). Editora INPA, Manaus, Brazil. Guillaumet, J.L. 1987. Some structural and floristic aspects of the forest. Experientia 43:241-251. Hofer, H. & A.D. Brescovit. 2001. Species and guild structure of a Neotropical spider assemblage (Araneae) from Reserva Ducke, Amazonas, Brazil. Andrias 15:99-119. Jocque, R., F. Samu & T. Bird. 2005. Density of spiders (Araneae: Ctenidae) in Ivory Coast rainforests. Journal of Zoology 266:105-110. Johnson, J.C. & A. Sih. 2007. Fear, food, sex and parental care: a syndrome of boldness in the fishing spider {Dolomedes triton). Animal Behaviour 74:1131-1138. Lucas, S.M. 1969. Aranhas da familia Ctenidae, subfamilia Phoneu- trinae. IV. Contribuiqao ao estudo da ooteca, dos ovos e da eclosao da aranha armadeira Phoneutria sp. Memorias do Instituto Butantan 34:37^1. Lucas, S.M. 1988. Spiders in Brazil. Toxicon 26:759-772. Lucas, S.M. 2002. Aranhas de interesse medico no Brasil. Pp. 141-149. In Animals Peqonhentos no Brasil: Biologia, Clinica e Terapeutica dos Acidentes. (J.L.C. Cardoso, F.O. de S Franqa, F.H. Wen, C.M.S. Malaque & V. HaddadJiinior, eds.). SARVIER, Sao Paulo. Ludwig, J.A. & J.F. Reynolds. 1988. Statistical Ecology: A Primer on Methods and Computing. John Wiley & Sons, New York. Maretic, Z. 1987. Spider venoms and their effect. Pp. 142-159. In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer Verlag, Berlin. Marques-Filho, A.O., M.N. Ribeiro, H.M. dos Santos & J.M. dos Santos. 1981. Estudos climatologicos da Reserva Florestal Ducke- Manaus AM. Precipitaqao. Acta Amazonica 11:759-768. Martins, R. & R. Bertani. 2007. The non-Amazonian species of the Brazilian wandering spiders of the genus Phoneutria Perty, 1833 (Araneae: Ctenidae), with the description of a new species. Zootaxa 1526:1-36. Menin, M., F. Waldez & A. Lima. 2008. Temporal variation in the abundance and number of species of frogs in a 10,000 ha of a forest in central Amazonia, Brazil. South American Journal of Herpe- tology 3:68-81. Morse, D.H. 1997. Distribution, movement, and activity patterns of an intertidal wolf spider Pardosa lapidicina population (Araneae: Lycosidae). Journal of Arachnology 25:1-10. Polls, G.A., C.A. Myers & R.D. Holt. 1989. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology and Systematics 20:297-330. Polls, G.A. & R.D. Holt. 1992. Intraguild predation: the dynamics of complex trophic interactions. Trends in Ecology & Evolution 7:151-154. Ribeiro, J.E.L.S., M.J.G. Hopkins, A. Vicentini, C.A. Sothers, M.A.S. Costa, J.M. Brito, M.A.D. Souza, L.H.P. Martins, L.G. Lohmann, P.A.C.L. Assunqao, E.C. Pereira, C.F. Silva, M.R. Mesquita & L. Procopio. 1999. Flora da Reserva Ducke: Guia de Identificaqao das Plantas Vasculares de uma Floresta de Terra Firme na Amazonia Central. INPA/DFID, Manaus, Brazil. Rypstra, A.L. & F. Samu. 2005. Size dependent intraguild predation and cannibalism in coexisting wolf spiders (Araneae, Lycosidae). Journal of Arachnology 33:390-397. Schiapelli, R. & B. Gerschman de Pikelin. 1972. Diagnosis de Phoneutria reidyi (F.O. Pickard-Cambridge, 1897) y de Phoneutria holiviensis (F.O. Pickard-Cambridge, 1897) (Araneae, Ctenidae). Revista de la Sociedad Entomologica Argentina 34:31-38. Sim6, M. & A.D. Brescovit. 2001. Revision and cladistic analysis of the Neotropical spider genus Phoneutria Perty, 1833 (Araneae, Ctenidae), with notes on the related Cteninae. Bulletin of the British Arachnological Society 12:67-82. Sokal, R.R. & F.J. Rohlf. 1995. Biometry: the Principles and Practice of Statistics in Biological Research. 3rd edition. W.H. Freeman and Company, New York. Turner, M. & G. Polis. 1979. Patterns of co-existence in a guild of raptorial spiders. Journal of Animal Ecology 48:509-520. Uetz, G.W. 1977. Coexistence in a guild of wandering spiders. Journal of Animal Ecology 46:531-541. Uetz, G.W. 1991. Habitat structure and spider foraging. Pp. 325-348. In Habitat Structure: the Physical Arrangement of Objects in Space. (S.A. Bell, E.D. McCoy & H.R. Mushinsky, eds.). Chapman & Hall, London. Vasconcelos, H.L. 1990. Effects of litter collection by understory palms on the associated macroinvertebrate fauna in Central Amazonia. Pedobiologia 34:157-160. Vetter, R.S. & G.K. Isbister. 2008. Medical aspects of spider bites. Annual Review of Entomology 53:409-429. Vieira, R.S. & H. Hofer. 1994. Prey spectrum of two army ant species in central Amazonia, with special attention on their effect on spider populations. Andrias 13:189-198. Wilkinson, L. 1990. SYSTAT: The System for Statistics. Systat Inc., Evanston, IL. SYSTAT Inc. Wise, D.H. 1993. Spiders in Ecological Webs. Cambridge University Press, Cambridge, UK. Wise, D.H. & B. Chen. 1999. Impact of intraguild predators on survival of a forest-floor wolf spider. Oecologia 121:129-137. Manuscript received 10 November 2008, revised 9 June 2010. 2010. The Journal of Arachnology 38:441^5 Ultraviolet light detection: a function of scorpion fluorescence Carl T. Kloock, Abraham Kubli and Ricco Reynolds; Department of Biology, California State University, Bakersfield, 9001 Stockdale Highway, Bakersfield, California 93311, USA. E-mail: ckloock@csiib.edu Abstract. The hypothesis that fluorescence in scorpions functions in the detection of ultraviolet light was tested. We reduced the fluorescence of scorpions by prolonged exposure to ultraviolet light on a 16:8 h light:dark cycle and compared their activity levels and light environment choices to unmodified scorpions in simple arenas that were half in shadow and half exposed to light. Three different lighting conditions were tested: infrared (IR) light only, IR + ultraviolet light and IR + white light. Treatments were illuminated by infrared light for videotaping. Activity level was measured by the number of transitions from the exposed to shadowed regions, and choice was measured by the percentage of time spent in the shadowed portion of the arena. Under IR + ultraviolet light, fluorescent scorpions reduced their activity levels and the variance in habitat choice increased, compared with reduced-fluorescence scorpions. There were no differences between fluorescent and non-fluorescent scorpions in the IR only condition or in the IR + white light condition. This is interpreted as evidence that lluorescence aids in the detection of and response to ultraviolet light, and possible implications of this result in natural habitats are discussed. This is the first experimental demonstration of a possible function for scorpion fluorescence. Keywords: Moonlight avoidance, habitat choice, light responses The fluorescence of scorpion cuticles is a well known, but little understood, phenomenon. Although two molecules associated with scorpion fluorescence have been isolated and identified - a B-carboline (Stachel et al. 1999) and 4-methyl, 7- hydroxycoumarin (Frost et al. 2001) - no function of scorpion fluorescence has previously been demonstrated. This article reports the first empirical support for a function for scorpion fluorescence. Several specific hypotheses regarding possible functions of scorpion fluorescence have been put forward, including the possibility that fluorescence functions in ultraviolet (UV) light detection (Blass & Gaffin 2008), mate identification and species discrimination (Kloock 2008), luring of prey (Kloock 2005), light amplification (Camp & Gaffin 1999), or as a sunscreen (Lourengo & Cloudsley-Thompson 1996). Some authors have hypothesized that fluorescence has no function, being either a relict trait (Frost et al. 2001) or correlated with some other functional aspect of the molecules responsible (i.e., sclerotization: Stachel et al. 1999). Those functions that have been tested to date have not received empirical support (Kloock 2005, 2008). A recent methodological development, the ability to significantly reduce fluorescence from live scorpions (Kloock 2009), makes new tests of functional hypotheses possible by allowing us to compare the behavior of fluorescent and reduced-fluorescence scorpions in different situations. We report here on a series of experiments supporting the hypothesis that scorpion fluorescence functions in the detection of ultraviolet light. METHODS Specimens. — Female Paruroctonus becki (Gertsch & Allred 1965) (Vaejovidae) were collected July-September 2008 in Kern County, California (voucher specimens deposited at the California Academy of Sciences, San Francisco). They were maintained in the laboratory in small, foam-plugged plastic vials, fed mealworm larvae and misted with water once per week until beginning the fluorescence reduction procedure. Only female scorpions were used in the experiments. Fluorescence reduction. — We used a modification of the method for reducing the fluorescence from living scorpions presented in Kloock (2009) to produce a group of scorpions with significantly reduced fluorescence. The original method involved exposing scorpions for 24 h/day to low level (1 IpW/ cm^) UV lights (two 40 W fluorescent GE black light tubes) until fluorescence faded (~ 4.5 wk). Because of the potential effect of constant light exposure on scorpion circadian rhythms (Fleissner 1977a, b, c; Schliwa & Fleissner 1980; Fleissner & Fleissner 2001), and therefore behavior, we modified this technique by exposing the scorpions on a 16:8 h light:dark cycle. This extended the time for complete loss of visible fluorescence to ~6 wk, but otherwise resulted in an effect similar to that reported by Kloock (2009). During UV exposure, scorpions were housed in small, open- topped plastic containers (13 cm length X 10 cm width X 7 cm height) with 12 ml of native soil: enough to provide a substrate, but insufficient for scorpions to bury themselves for protection from UV exposure. Scorpions were fed a single mealworm larva twice a week: if they failed to eat the larva, it was removed and a fresh larva provided at each feeding. Scorpions were also provided water twice a week by lightly misting the soil surface. Scorpions were maintained on this schedule until all experiments using them were completed. Control scorpions. — Control and reduced-fluorescence scor- pions were kept in identical containers with the same amount of soil and on identical feeding, watering, and lighting schedules with the exception that a layer of UV blocking film (Edmund Optics NT39-426) was interposed between the UV light source and the scorpions. Control and experimental scorpions were kept in a common environmental chamber with the same UV lights during UV exposure, so that environmental conditions were identical for all scorpions. Control scorpions showed no visible reduction in fluorescence over the course of the experiments. Basic setup. — The basic experimental design owes consider- able debt to Blass and Gaffin (2008), whose methods we have adapted and simplified for our purposes. All three experiments share the basic feature of placing scorpions in 14-cm diam. 441 442 THE JOURNAL OF ARACHNOLOGY 0©©€3 Figure 1. — Schematic of the 3x5 array of half-painted Petri dishes used in the experiment, showing a typical random orientation of dark and light halves. Petri dishes, ■which had one-half of their exterior surface (top and bottom) painted over with water-based, non-toxic black paint (Fig. 1). No paint was on the interior surface of the Petri dish to act as a possible cue for the scorpions. Petri dishes were washed with 70% ethanol and allowed to dry for more than 20 h before each use to remove any potential chemical cues previous occupants may have deposited (Steinmetz et al. 2004). We placed a single scorpion, either fluorescence-reduced or control, in each Petri dish, aligning the top and bottom halves so that half of the surface area of the Petri dish allowed light in, while the other half served as a light refuge. Petri dishes with scorpions were placed with the light and dark halves randomly oriented on a clear Plexiglas observation deck (Fig. 1) with an infrared sensitive video camera below and six infrared light emitting diodes (830 ± 35 nm) placed on the sides of the observation deck and reflecting off a light blue panel attached to the ceiling. This provided diffuse infrared illumination for videotaping. Scorpions have previously been shown not to respond to infrared illumination (Blass & Gaffin 2008). Control and reduced-fluorescence scorpions were alternated in a 15 Petri dish array (three rows, five columns) and the camera adjusted to allow good visualization of all scorpions. A monitor in another room was used to ensure that all Petri dishes were easily observed, and all experiments were recorded on 42-min videotapes, which set the length of each trial. Two variables were recorded from the videotape analysis: percent time spent in the dark-covered half and the number of times scorpions moved from the exposed side to the covered side of the Petri dish (transitions). The percent time in the dark half acts as a measure of habitat choice, while the number of transitions acts as a simple measure of activity level. In all experiments scorpions were allowed to acclimate to the array and lighting conditions for one hour before videotaping began, and the acclimation period began within 1 5 min of laboratory “sunset” to target the typical scorpion activity period (Hadley & Williams 1968). Just prior to the initiation of each experiment, each scorpion was moved into the clear portion of the Petri dish by gently tilting the dish toward the clear side; this ensured that all scorpions were visible at the beginning of the trial, and that they all experienced the lighting conditions, at least briefly, during the trial. Illumination other than the infrared needed for videotaping differed in the experiments that follow. ' Experiments 1 and 2. — The same set of 15 control and 15 ; reduced-fluorescence scorpions was used for experiments 1 and 2. The order of presentation was randomized, with eight ; of the control and seven of the reduced-fluorescence scorpions receiving IR only illumination first (Experiment 1). Seven ' control and eight reduced-fluorescence scorpions received ! IR+UV illumination first (Experiment 2). The scorpions were i then exposed to the alternate illumination, so that each experiment achieved a total sample size of 1 5 fluorescent and ■ 15 reduced-fluorescence scorpions. All trials were completed i over a four-night period. i Experiment I: No illumination. This experiment constitutes a control for possible effects of the fluorescent reduction | treatment. We placed scorpions in the standard experimental : setup described above and videotaped them for 42 min. Fluorescence reduction should have no effect in this environ- ; ment. If control and fluorescence-reduced scorpions exhibit : significant differences in this experiment, then behaviors being measured were changed by the fluorescence reduction I procedure itself. ! Experiment 2: IR -h UV illumination. As in Experiment 1, scorpions were placed in the standard experimental setup, but UV illumination was provided by a rectangular array (52 cm X 67 cm) of nine UV LEDs (Roithner-Lasertek RLT360- LO- IS, peak emission wavelength = 361 nm, spectral lA-width = 10 nm, viewing lA-angie = 15d) equally spaced. The array was placed 1.2 m above the observation deck, allowing room for the light from each LED to overlap, providing diffuse illumination across the observation deck. Although these LEDs caused scorpions to fluoresce at short distances, at a distance of 1.2 m, scorpions under the array did not visibly I fluoresce; measurement of UV power using a Mannix UV-340 light meter (range = 290-390 nm) yielded < 1 pW/cm^. Although fluorescence was not detectable by human vision, the question we are asking here is whether or not it is detectable by scorpions, whose vision is much more sensitive to low light levels than humans’ (Fleissner 1977c). Ultraviolet intensity was kept very low to mimic natural nocturnal ' conditions. If fluorescence functions in UV light detection, we expect to see significant differences between control and fluorescence-reduced scorpions in this experiment. Experiment 3, — IR + White light: The expected results of Experiment 2 under the hypothesis that fluorescence functions in UV light detection are also consistent with the possibility that the fluorescence-reduced scorpions suffered damage to the retina and/or the extra-ocular light sense (Zwicky 1970) during their exposure to UV light. To test this possibility we initiated a new experiment using white light. The basic experiment was identical to experiments 1 and 2, except that | additional illumination was supplied by white light from two ( 40W fluorescent tubes. Rather than being placed directly over the scorpions, the lights were placed to the side to provide 1 diffuse illumination because of the higher power of these lights. These lights did not produce measurable UV light on : the observation deck. This experiment used novel scorpions, j treated identically to those used in experiments 1 and 2, but | with a slightly smaller sample size due to the death of some of - the scorpions in this group during the preparation period. For ‘ KLOOCK ET AL.— A FUNCTION OF SCORPION FLUORESCENCE 443 Table 1. — Summary of transitions between light and dark regions of the Petri dish by fluorescence in experiments 1-3. All data presented use the transformation (transitions” Experiment 1 (IR only) Experiment 2 (IR+UV Light) Experiment 3 (IR+ White light) Light condition Control (fluorescent) (« = 15) Fluorescence reduced (n = 15) Control (fluorescent) (n = 15) Fluorescence reduced (// = 15) Control (lluorescent) (n = 14) Fluorescence reduced (« = 13) Transitions (mean ± SD) 4.1 ± 3.2 4.8 ± 2.2 3.0 ± 2.0 4.7 ± 2.2 1.2 ± 0.43 1.3 ± 0.44 95% Cl: Lower limit 2.3 3.6 1.9 3.5 1.0 1.1 Upper limit 5.9 6.1 4.1 5.9 1.5 1.6 Levene’s test for F = 2.447 F= 0.182 - Non-normally distributed variances P = 0.129 P = 0.673. Test for means Equal variances /-test Equal variances /-test Mann-Whitney U-test Test statistic / = 0.751, df = 28 / = 2.244, df ■ = 28 U = 90.0 P 0.459 0.033 0.961 Experiment 3, there were 14 control scorpions and 13 reduced- fluorescence scorpions. If scorpion vision was damaged by the fluorescence reduction treatment, we should see significant differences between control and reduced-fluorescence scorpi- ons, in a pattern similar that of Experiment 2. Statistical analysis. — Evaluation of the hypothesis that fluorescence affects the response to UV does not depend on differences between the light treatments, but rather on the overall pattern of differences between control and reduced- fluorescence scorpions within each of the three experiments. Thus we are not interested in differences between light treatments, but instead in the differences between control and reduced-fluorescence scorpions within each experiment. The number of transitions and the percentage of time spent in the dark were analyzed separately in each experiment to determine if significant differences existed. Both variables within each experiment were analyzed for normality using the Kolmogorov-Smirnov test for normality, and homogeneity of variances was tested using Levene’s test for equality of variances. Means within each experiment were compared with either the Mann-Whitney U test (for non-normally distributed data) or a t-test (for normally distributed data), with assumptions of equal or unequal variances as dictated by the data structure. Given that the number of transitions is expected to have a Poisson distribution, the square root transform was applied a priori to normalize this variable in all tests. RESULTS The number of transitions in the IR + white light condition was not normally distributed (Kolmogorov-Smir- nov Z = 1.993, P = 0.001). All other Kolmogorov-Smirnov tests of normality showed that the data were normally distributed (For % dark: IR only Z = 0.500, P = 0.964; IR+UV light, Z = 0.681, P = 0.743; IR + white light, Z = 1.088, P = 0.187. For transitions”''^': IR only, Z = 0.834, P = 0.490; IR +UV light, Z = 0.683, P = 0.740). Table 1 presents the data and results of statistical tests on the number of transitions between the light and dark sides of the Petri dish. Levene’s test for equality of variances revealed no differences between variances in experiments 1 and 2. Therefore, /-tests assuming equal variance were used to compare fluorescent and non-fluorescent scorpions while the non-parametric Mann-Whitney U test was used for Experi- ment 3 due to the non-normality of transitions'^' in white light. Fluorescent and reduced-fluorescence scorpions did not differ in the number of transitions when either IR only or white light was present, but did differ when UV light was present (Table 1). Fluorescent scorpions made fewer transitions when exposed to UV light than did fluorescence- reduced scorpions. Table 2 presents the data and results of statistical tests on the percentage of time spent in the darkened half of the Petri dish (% dark). Levene’s test for equality of variances showed no difference in the variance between fluorescent and reduced- fluorescence scorpions in either the IR only or white light experiments, but a significant difference in variance in the UV light experiment was found. Thus /-tests assuming equal variance were applied to experiments 1 and 3, while for Experiment 2, equal variances were not assumed. There were no significant differences between the means of fluorescent and reduced-fluorescence scorpions for % dark in any of the experiments. Confidence intervals all include 50%, as expected if scorpions exhibit no preference for either side of the Petri dish. Because of the difference in variances detected in Experi- ment 2, we decided to look more closely at the distribution of the percentage of time spent in the dark to see, post hoc, if any patterns emerged. Inspection of histograms suggested that the larger variation in fluorescent scorpions was caused by the tendency of these scorpions to stay in either the light or dark half of the Petri dish, with few of these scorpions having intermediate values of % dark. A simple categorization of % dark into “extreme” (< 25% or > 75%) vs. “moderate” (between 25% and 75%) values reveals this. With only IR light (Experiment 1), scorpions showed no tendency toward extreme values, regardless of fluorescence condition. Both fluorescent and reduced-fluorescence scorpions showed 7 “extreme” and 8 “moderate” values, = 0.067, df = 1, P = 0.80). With UV light present (Experiment 2) reduced- fluorescence scorpions showed no tendency toward extremes (6 “extreme” and 9 “moderate”, = 0.60, df = P = 0.44), while fluorescent scorpions displayed a strong tendency toward extreme values (13 extreme, 2 moderate, x~ = 8.1, df = 1, f* = 0.0045). Scorpions exposed to white light, despite reduced activity levels (Table 1) also showed no preference for extremes (Reduced-fluorescence; 8 extreme, 5 moderate, ')C = 1 444 THE JOURNAL OF ARACHNOLOGY Table 2. — Summary of the percentage of time spent in the dark half of the Petri dish by fluorescence in experiments 1-3. Light condition Experiment 1 (IR only) Experiment 2 (IR+UV light) Experiment 3 (IR+White light) Control (fluorescent) (n = 15) Fluorescence reduced {n = 15) Control Fluorescence (fluorescent) reduced (« = 15) (« = 15) Control (fluorescent) {n = 14) Fluorescence reduced in = 13) Mean % Dark (mean ± SD) 44.8 ± 32.8 40.4 ± 25.3 55.5 ± 37.9 56.8 ± 25.3 43.9 ± 39.6 44.2 ± 38.1 95% Cl; Lower limit 26.7 25.3 34.5 42.8 34.5 21.2 Upper limit 63.0 54.4 76.4 70.9 76.4 613 Levene’s test for variances 1.104, P = 0.302 F = 7.945, P = 0.009 0.192, P = 0.655 t-test assumption Equal variances Variances not equal Equal variances t(df) 0.418 (28) 0.118 (28) 0.022 (25) P 0.679 0.907 0.983 0.69, P = 0.41: Fluorescent; 9 extreme, 5 moderate, = 1.1, F = 0.29). We should caution, however, that both the small sample size for and post-hoc nature of this analysis call for conservative use of this information in interpretation. DISCUSSION These results support the hypothesis that scorpion fluores- cence serves as a means for the detection of ultraviolet (UV) light at very low levels. Significant differences between the activity levels of fluorescent and reduced-fluorescence scorpi- ons occurred only in the presence of UV light, with fluorescent scorpions changing their behavior by reducing their activity level (Table 1). Additionally, tests on variances in % dark revealed a difference between fluorescent and reduced- fluorescence scorpions only under UV light (Table 2). Although a difference in the variance in the % dark under UV light was detected, there was no difference in the mean time spent in the two sides of the dish by fluorescent and reduced-fluorescence scorpions; in fact, no differences in the mean % dark were observed in any experiment (Table 2). No preference for either side of the Petri dish was observed in any of the conditions, as all confidence intervals include the random expectation of 50% (Table 2). Although fluorescent scorpions reduced their activity levels in response to UV light, this did not change the mean time they spent in the different light environments. In other words, the change in activity level did not result in a change in the average use of the environment. The obvious question then is whether this response to UV light has any value in a natural environment. Scorpion surface activity is generally higher during moon- less nights than moonlit nights (Skutelsky 1996), and nocturnal UV light levels correlate to moon phase (Silberglied 1979). Thus, it is possible that UV light acts as a cue for moonlight avoidance. Blass & Gaffm (2008) demonstrated avoidance of UV light, but the data here indicate no avoidance of UV (Table 1). It is very likely that the low level of UV light used in this experiment did not reach a threshold for avoidance behavior. Blass & Gaffm (2008) used a greater intensity of UV light in their experiment (0.9 lux), so were more likely to observe avoidance. More work needs to be done to determine more precisely the intensity required to elicit avoidance behavior, and whether this differs between fluores- cent and reduced-fluorescence scorpions. In order to deter- mine if avoiding UV light results in moonlight avoidance, experiments comparing lluorescent and reduced-fluorescence scorpions in moonlight and UV-filtered moonlight are necessary. Conducting tests under natural illumination condi- tions is a natural follow-up to the current experiment that would help determine whether simple moonlight avoidance is the main function of fluorescence. Another interesting ; experiment could involve measuring individual scorpion ' responses to UV light before reducing their fluorescence, then ■ reducing their fluorescence and re-measuring, and finally, 1 measuring a third time after fluorescence recovers. This would i establish that the behavior of individuals changes in response I to the manipulation. Unfortunately, the time involved in this | set of manipulations (~ 6 weeks to remove fluorescence, plus several weeks to recover full fluorescence) would introduce the potential confounding of seasonal differences in responses. i The decision of whether to forage in moonlight or seek cover is influenced by factors other than UV light levels. ; Therefore, we must consider the possibility that the lack of UV light avoidance displayed in this experiment resulted from differences in decisions by individual scorpions about whether or not to seek cover. For example, Skutelsky (1996) found that scorpions with lower body massilength^ ratios were more likely to forage on moonlit nights than those with higher ratios, indicating that energy reserves were an important factor in the choice to forage while exposed to moonlight. We i controlled the food offered to scorpions, but we did not i control whether they actually ate, nor the scorpion size to prey ! size ratio. Therefore, motivation for foraging likely varied among the scorpions, though randomly with respect to fluorescence. If some fluorescent scorpions chose to seek • cover while others chose to “forage” in the open, we would f expect to see precisely what was observed in this experiment: ' reduced activity levels of fluorescent scorpions coupled with j' an increase in the variance of time spent in the dark caused by i fluorescent scorpions choosing to spend most of their time in ! either the darkened or exposed habitat, with the specific choice | influenced by hunger levels. 1 The post hoc inspection of fluorescent scorpions behavior [ when exposed to UV light is consistent this interpretation (with of course the caveat that it is post hoc). Fluorescent scorpions exposed to UV light exhibited a tendency to stay in . one light environment that was not evident in any other treatment combination, consistent with the idea that they are | making an active decision about where to spend their time in f response to UV light, even if the decisions of individual scorpions differed. The reduced-fluorescence scorpions did not I KLOOCK ET AL.— A FUNCTION OF SCORPION FLUORESCENCE 445 exhibit this tendency. If this interpretation is correct, an experiment comparing starved to fed scorpions with and without UV present should reveal that fed scorpions seek cover from UV light while starved scorpions choose exposure to UV light, and this difference should disappear in scorpions that have had their fluorescence reduced. Previous work on light-detection abilities of scorpions has shown UV light sensitivity in the lateral eyes (but not the median eyes; Machan 1968) and in an extra-ocular light sense localized in the metasoma (Zwicky 1968, 1970). Because the sensitivities of these senses are very similar (Zwicky 1970), we cannot at present attribute the observed changes in behavior to either one of these sensory mechanisms, and indeed it may involve both. Future experiments could attempt to determine the relative effect each of these senses has on the behavior with finer control of UV light wavelengths, and with more detailed information on the spectral sensitivities of these different mechanisms. Machan (1968) also showed that scorpion lateral eyes have a second peak in sensitivity near the wavelength of peak fluorescence emission, which is also the region of the peak in median eye sensitivity. This dual peak in scorpion vision has suggested the possibility that scorpions possess dichromatic vision (Machan 1968; Kloock 2008). However, our results suggest another possibility. The peak in sensitivity in the UV range measured by Machan (1968) may have been caused by the eye detecting fluorescence caused by UV light, rather than by directly detecting UV light. In other words, the observed peak in sensitivity to UV light may be partially or entirely a byproduct of sensitivity to the light produced by fluorescence. If the UV sensitivity peak is actually a byproduct of fluorescence, we should see the peak in sensitivity disappear (or at least drop in amplitude) in reduced-fluorescence scorpions relative to fluorescent scorpions when the cuticle is exposed to UV light. This study provides the first experimental evidence sup- porting a function for scorpion fluorescence. Further work will be necessary to link UV light detection to uses that scorpions may have for this ability in their natural habitat. Although the mechanism of detection is unclear, potentially being mediated through vision, the extra-ocular light sense or a third, as yet unknown, mechanism, the role of fluorescence in UV light detection may have implications for our understanding of fluorescence and the role of light in scorpions’ sensory world. ACKNOWLEDGMENTS The authors would like to thank M. Piris and E. Martinez for their patience and aid with our research efforts. R. Reynolds is a participant in the MARC U*STAR student research program at CSUB. We would also like to thank our anonymous reviewers for improving the quality of the manuscript. LITERATURE CITED Blass, G.R.C. & D.D. Gaffin. 2008. Light wavelength biases of scorpions. Animal Behaviour 76:365-373. Camp, E.A. & D.D. Gaffin. 1999. Escape behavior mediated by negative phototaxis in the scorpion Parwoctomis utahensis (Scorpiones, Vaejovidae). Journal of Arachnology 27:679-684. Fleissner, G. 1977a. Entrainment of the scorpion’s circadian rhythm via the median eyes. Journal of Comparative Physiology A 118: 93-99. Fleissner, G. 1977b. Scorpion lateral eyes: extremely sensitive receptors of zeitgeber stimuli. Journal of Comparative Physiology A 118:101-108. Fleissner, G. 1977c. The absolute sensitivity of the median and lateral eyes of the scorpion, Androctonus australis L. (Buthidae, Scor- piones) Journal of Comparative Physiology A 118:109-120. Fleissner, G. & G. Fleissner. 2001. Neuronal organization of circadian systems. Pp. 107-137. In Scorpion Biology and Research. (P. Brownell & G.A. Polls, eds.). Oxford University Press, Oxford, UK. Frost, L.M., D.R. Butler, B. O’Dell & V. Fet. 2001. A coumarin as a fluorescent compound in scorpion cuticle. Pp. 365-368. In Scorpions 2001: In Memoriam, Gary A. Polls. (V. Fet & P.A. Selden, eds.). British Arachnological Society, Burnham Beeches, Buckinghamshire, UK. Hadley, N.F. & S.C. Williams. 1968. Surface activities of some North American scorpions in relation to feeding. Ecology 49:726-734. Kloock, C.T. 2005. Aerial insects avoid fluorescing scorpions. Euscorpius 21:1-7. Kloock, C.T. 2008. A comparison of fluorescence in two sympatric scorpion species. Journal of Photochemistry and Photobiology B: Biology 91:132-136. Kloock, C.T. 2009. Reducing scorpion fluorescence via prolonged exposure to ultraviolet light. Journal of Arachnology 37:368-370. Lourengo, W.R. & J.L. Cloudsley-Thompson. 1996. The evolutionary significance of colour, colour patterns and fiuorescence in scorpions. Revue Suisse de Zoologie, volume hors serie 2:449-458. Machan, L. 1968. Spectral sensitivity of scorpion eyes and the possible role of shielding pigment effect. Journal of Experimental Biology 49:95-105. Schliwa, M. & G. Fleissner. 1980. The lateral eyes of the scorpion, Androctonus australis. Cell and Tissue Research 206:95-1 14. Silberglied, R.E. 1979. Communication in the ultraviolet. Annual Review of Ecology and Systematics 10:373-398. Skutelsky, O. 1996. Predation risk and state-dependent foraging in scorpions: effects of moonlight on foraging in the scorpion Bntlnis occitanus. Animal Behaviour 52:49-57. Stachel, S.J., S.A. Stockwell & D.L. Van Vranken. 1999. The fiuorescence of scorpions and cataractogenesis. Chemistry and Biology 6:531-539. Steinmetz, S.B., K.C. Bost & D.D. Gaffin. 2004. Response of male Centruroides vittatus (Scorpiones: Buthidae) to aerial and sub- strate-borne chemical signals. Euscorpius 12:1-6. Zwicky, K.T. 1968. A light response in the tail of Urodacns, a scorpion. Life Sciences 7:257-262. Zwicky, K.T. 1970. The spectral sensitivity of the tail of Urodacns, a scorpion. Specialia 15:317. Manuscript received 18 December 2009, revised 21 May 2010. 2010. The Journal of Arachnology 38:446-451 Social organization of the colonial spider Leucauge sp. in the Neotropics: vertical stratification within colonies Maxence Salomon, Carly Sponarski', Allen Larocque- and Leticia Aviles: Department of Zoology and Biodiversity Research Centre, University of British Columbia, Vancouver, B. C. V6T 1Z4, Canada. E-mail: salomon@zoology.ubc.ca Abstract. A first step toward understanding why sociality has evolved in a particular taxonomic group is to establish comparison points by studying the organization of different social systems. We examined the social organization and spatial distribution of individuals in colonies of the undescribed colonial spider Leucauge sp. (Araneae: Tetragnathidae). The social organization of this species was typical of a colonial species, with spiders maintaining individual territories (orb webs) within a scaffolding of shared support lines maintained by the group. Furthermore, we observed a size-dependent vertical stratification of spiders within colonies, with large spiders occupying the highest positions, followed by medium, extra-small and small individuals, a spacing pattern that was consistent across colonies of all sizes. Spiders captured and consumed prey individually and displayed territorial behaviors involving web defense. This study provides a new example of a colonial spider species that shows a distinctive within-group spatial distribution. We discuss possible reasons underlying this species’ spatial arrangement in the context of social evolution. Keywords: Aggregation, group living, orb web, sociality, spatial distribution Knowledge about the social organization of particular species provides key insight into the mechanisms and conditions involved in the evolution of sociality. Spiders have proven to be good model systems for the study of social evolution because they encompass a large range of social phenotypes (see Aviles 1997; Uetz & Hieber 1997; Lubin & Bilde 2007 for classic and recent reviews). The few species that express social tendencies fall into two broad categories defined by individuals’ level of cooperation and group structure: ‘colonial’ and ‘cooperative’ (Aviles 1997; Uetz & Hieber 1997). Colonial (or territorial permanent-social) species are generally orb-weaving spiders that spin individual webs and form permanent groups with low dispersal rates. A colonial lifestyle is characterized by the grouping of individual webs that serve as foraging or multi-purpose territories depending on the species, and which are usually maintained by single individ- uals. Group members engage in individual activities on their webs (e.g., foraging, brood rearing), and cooperation is usually limited to the maintenance of shared framework silk that joins the different webs (but see Fernandez Campon 2007 for an example of cooperative foraging). Conversely, cooper- ative (or non-territorial permanent social) spiders form permanent groups on communal webs without any spatial separation between group members, and individuals cooperate in various activities such as prey capture, web maintenance and parental care. Although both social structures have evolved independently multiple times, these two organization- al schemes represent distinct evolutionary pathways to sociality (Lubin & Bilde 2007). The social organization of a group of individuals may be characterized by various attributes, such as spatial arrange- ment, temporal pattern, behavioral interactions and genetic relationships. The characteristics of these social attributes Current addresses: 'Faculty of Environmental Design, University of Calgary, Calgary, A. B. T2N 1N4, Canada; "Department of Biology, McGill University, Montreal, Q. C. H3A IBl, Canada. depend largely on the tradeoffs between the benefits and costs of communal living. In colonial spiders, benefits of group living include decreased per capita silk investment due to a i shared silk framework (Uetz & Hieber 1997), increased i ' accessibility to areas of high prey availability (e.g., open space over bodies of water) otherwise out of the reach of solitary i ■ spiders (Buskirk 1975a; Smith 1983), enhanced predator warning (Uetz et al. 2002) and increased prey capture success ^ due to the proximity between individual webs that may cause : prey to ‘ricochet’ from one web to the next (Uetz 1989; 1. Whitehouse & Lubin 2005). However, coloniality also involves costs, such as increased vulnerability to predators or parasites and competition for local resources including food and web space (Buskirk 1975a, b; Uetz & Hieber 1997; Rayor & Uetz 2000). In this study, we document the colonial structure of a recently discovered and undescribed neotropical spider, Leucauge sp. (referred to as Plesiometa sp. in Aviles et al. 2001). This orb-weaving species has previously been catego- < rized as ‘colonial’ by Aviles et al. (2001), based on limited data i about its social organization. To address this shortcoming, we ) examined the social organization of Leucauge sp. colonies, I focusing on group composition, colony architecture and the 1 social dynamics involved in territory maintenance and * 1 foraging. Spatial structuring within colonies may reflect a b need for individuals to maximize resource acquisition and survival in a group-living situation where competition and predation pressure may impose constraints. Based on our j i preliminary observations suggesting a spatial arrangement of , ( I individuals within colonies, we tested the hypothesis that the j : distribution of Leucauge sp. individuals within a colony is p non-random with respect to spider size, a proxy for age class. METHODS .j. Study area. — We conducted this study in late August 2005 jl in the Jatun Sacha Biological Reserve (01°04'S, n°36’W, elev. 400-440 m), in the Napo Province of eastern Ecuador in the r 446 I SALOMON ET AL.— COLONIALITY IN LEUCAUGE SP. SPIDERS 447 Figure 1. — a) Leiicaiige sp. above a stream in the lowland rainforest of Ecuador: several visible orb webs (O) are joined together by framework silk (F) and anchored to the vegetation with long support lines (S). b) Adult male and c) adult female Leucauge sp. spider on an orb web (scale bars = 10 mm). Photo credits: a) and c) L. Aviles; b) A. Larocque. Amazon basin. The reserve comprises 2200 ha of 70% primary and 30% secondary tropical rainforest in a transitional region between the lower Andean slopes and the Amazonian lowlands (Jatun Sacha Foundation 2009). Study species. — The orb-weaving spider Leucauge sp. (Araneae: Tetragnathidae) is a territorial permanent-social species found in neotropical rainforests. This species was first characterized in Ecuador by Aviles et al. (2001) and bears significant morphological resemblance to Leucauge argyra (Walckenaer 1841) (= Plesiometa argyra: Platnick 2009). Leucauge sp. spiders are black with a prominent white patch on the abdomen outlined with silver bands and red coloring on their prosoma (Fig. lb, c). They spin individual orb webs that may reach close to 1 m in diameter and are usually found in riparian habitats of the rainforest above creeks or other bodies of water inside the forest (personal observation; Aviles et al. 2001). Colonies consist of clusters of individual orb webs joined together through a framework of shared silk scaffolding that is anchored to the vegetation on the banks of a body of water with thick silk threads (Fig. la). Their phenology and dispersal patterns are unknown, although we have observed spiders of all age classes in August. Data collection and analysis. — This study consisted of two parts: a population survey of Leucauge sp. colonies (n - 22) at the study site and behavioral observations of a single colony. Population survey: We collected the following population data over two days: (1) number of spiders per colony and, for each spider therein, (2) spider body size, (3) height of the spider from the ground and (4) location within the colony. Spider body size, measured with a ruler as total body length (distance from the front of the prosoma to the tip of the abdomen), reflects both developmental stage and feeding history in spiders (Jakob et al. 1996). We grouped spiders into four size classes based on the overall distribution of observed body lengths: large (‘L’, average length: 11-12 mm), medium (‘M’, 7-9 mm), small (‘S’, 5-6 mm) and extra-small (‘XS’, 3- 4 mm). L spiders were adult females and males, M spiders were sub-adults and both S and XS spiders were juveniles of various instars. We measured a spider’s height from the ground as the distance from the surface of a body of water directly below a colony to the center of the spider’s body. Spider location within the colony was categorized as on ( 1) an orb, (2) a support strand within the silk framework, (3) a dragline produced by the spider and attached to the web complex or (4) plant substrate at the periphery of a colony (usually the underside of a leaO- We collected these data on days without rain. For analysis, we first checked the data for normality and heteroscedasticity and applied transformations where appro- priate. We excluded three colonies with fewer than five spiders from the analyses (see Fig. 2). To examine differences in the distribution of spider sizes among colonies we used a log- likelihood ratio test with William’s correction for small sample sizes (Sokal & Rohlf 1995). We used a logistic regression model to test the effect of colony size on spider size distributions within colonies. We assessed the relationship between a spider’s size and its location within a colony using a log-likelihood ratio test with William’s correction. Finally, we used a general linear model (GLM) to examine the effects of colony size and spider size on a spider’s height above the ground (log-transformed), using averages for each spider size class within each colony to avoid pseudoreplication. We also assessed the relationship between colony size and the average height from the ground (log-transformed) of all spiders within a colony using a GLM and Spearman’s rank correlation. 448 THE JOURNAL OF ARACHNOLOGY i Number of spiders per colony Figure 2. — Distribution of Leucaiige sp. colony sizes (« = 22 colonies). The plain vertical line indicates the median colony size and the dashed vertical line the mean colony size. Behavioral observations: The behavioral component of the study consisted of the observation of a single colony for 10 h per day (0700-1700 hours) on two consecutive days. Our purpose was to record foraging activities and territorial interactions among colony members. The focal colony was chosen for ease of access, median colony size and the presence of all age classes (21 adults, subadults and juveniles). It was located on a sunny tree fall over a small creek. We observed the colony continuously over 10 h, and for each individual web in the colony we recorded ( 1 ) all successful prey capture events and the identity of the spider(s) involved in a capture and (2) the number of territorial interactions and the identity of both the intruding and resident spider. We continuously observed interactions between spiders until they led to a clear outcome (e.g., stay versus retreat). To increase our sample size, we also monitored prey capture at another colony of similar size. We estimated prey size (total body length) visually and recorded the prey’s taxonomic order. Prey biomass (dry weight) was calculated from total body length using insect-order specific regression models developed by Sage (1982). For analysis, we combined prey size and biomass data from both colonies. RESULTS Social organization and colony architecture. — Colonies ranged in size from 2-27 spiders, with a median group size of 19 (Fig. 2). Most colonies contained multiple adult females, adult males, sub-adults and juveniles. A colony typically consisted of a collection of individual orb webs arranged in a three-dimensional pattern and connected by a framework of silk forming a web complex (Fig. la). Orb webs were organized into multiple non-horizontal planes oriented either at the same angle or at different angles from each other (differences in orientation < 90°). Neighboring webs were closely arrayed and often faced similar directions (some were as close as 10 cm apart). The horizontal dimensions of a colonial web complex were usually proportional to the width and topology of the body of water above which they were placed and ranged from 0.5-3 m. Web complexes had a dense three-dimensional core with flattened edges connecting to the vegetation at the water’s edge (Fig. la), and colony height was proportional to the number of spiders (see ‘Spatial arrange- ment within colonies’). Web size (i.e. orb diameter) scaled positively with spider size: larger spiders occupied larger webs (personal observa- 1.6- 0^ , ^ ^ ^ XS S IVl L Spider size Figure 3. — Distribution of Leucauge sp. spiders’ heights from the ground (i.e., the water surface) as a function of body size (XS = extra- small; S = small; M = medium; L = large). Boxplots show medians (thick lines), means (filled squares), 25th and 75th percentiles (bottom and top of boxes) and 10th and 90th percentiles (cap of lower and upper whiskers). Letters inside the boxes denote differences between body-size classes based on Bonferroni-adjusted pairwise contrasts (see Results). tion). Individuals maintained individual territories (orb webs), although they were commonly seen moving between webs while travelling to other parts of the web complex, and some spiders did not own a web. When disturbed, spiders would typically flee along shared support silk threads to common retreats in the vegetation at the end of the anchor points to the substrate. Web owners regularly maintained their orb webs (i.e., as the need arose), and both spiders that owned webs and web-less spiders maintained the support silk framework. Web building was most common in the early morning and evening. Spatial arrangement within colonies. — There was significant heterogeneity across colonies in the proportion of spiders of different size classes (large, medium, small and extra-small; G^adj = 78.39, P = 0.02, df = 54). However, the distribution of spider sizes across colonies was independent of colony size (x^ = 6.43, df = 3, P = 0.09) except that extra-small spiders tended to be more common than small ones in large colonies (X^ = 5.67, P = 0.02). Large- and medium-sized spiders were more likely to occupy an orb, and small and extra-small spiders were more likely to be found on the framework, a dragline or the underside of peripheral leaves (Gadj = 120.58, df= 9, P< 0.0001). A spider’s vertical position within a colony was positively correlated with its size class = 22.54, P < 0.0001; Fig. 3). Large spiders assumed the highest positions in a colony (mean height ± SD: 1.10 ± 0.35 m, wl = 108), followed by medium-sized spiders (0.76 ± 0.26 m, «m = 70). Small spiders occurred closest to the bottom of the web complex (0.43 ± 0.22 m, «s = 73) whereas extra-small spiders clustered in between medium and small spiders (0.63 ± 0.28 m, Hxs — 1 14), apparently lacking individual orb webs. Therefore, from highest to lowest, the spiders from each size class were distributed as L > M > XS > S within a colony, with adults closer to the top and juveniles closer to bottom of the web complex. Bonferroni-adjusted contrasts revealed significant ( I 1 SALOMON ET AL.— COLONIALITY IN LEUCAUGE SP. SPIDERS 449 pairwise differences between each size class {P < 0.05) except between XS and S (P = 0.38; see Fig. 3). The correlation between height from the ground and spider body size held for colonies of all sizes as there was no significant interaction between spider size and colony size (^3,68 = 0.94, P = 0.43). Furthermore, the average height at which spiders of each size class occurred across colonies did not vary with colony size (F, 68 = 2.82, P = 0.10). However, the overall height from the ground of spiders in each colony increased with colony size (^1,17 = 5.25, P = 0.03), suggesting that colonies grow vertically. This relationship was stronger when the three smallest colonies with fewer than five spiders were included in the analysis (r^ = 0.71, n = 22, P = 0.0002). Territorial behavior. — We observed 20 attempts by eight different web-less spiders to displace web-holding individuals (“attacks”) on six different webs within the focal colony. Most of these attacks (16/20, 80%) took place on webs located in the center of the web complex and involved adult spiders. Responses to attacks followed an escalating pattern of agonistic behavior typically seen in colonial species (Buskirk 1975b; Hodge & Uetz 1995). Upon approach by an intruder, resident spiders positioned at the hub of the orb web would typically orient towards the intruder. The resident spider would then either contract all eight legs simultaneously, resulting in a rhythmic pulsing lasting 5-10 s that forced the intruder to halt and brace itself, or pluck the web by repeatedly jerking web radii with the front legs that also forced the intruder to brace itself (on two occasions jerks caused intruding spiders to fall out of the orb web). If intruders persisted and approached closer, the resident spider would rush out to face the intruder at the periphery of the web in a one-on-one encounter lasting from < 1 s to 3^ s, with legs and pedipalps entangled in a blur of activity. After such an encounter, the loser would retreat a short distance and the victor would scramble to the hub of the orb. Resident spiders most often won these encounters, repelling 85% (17/20) of attackers and maintaining their ownership of a web. Spiders not occupying orbs were generally tolerant of each other as close approach and even touching was observed on framework lines of the web complex without any agonistic interactions. Foraging behavior. — In the focal colony, we observed 23 prey capture events by 11 different spiders on 10 different orb webs. In the second colony we observed an additional 23 prey capture events. Most prey entered the web complex from the side; a few entered from the bottom. Spiders captured and consumed prey individually without cooperating. In both colonies, prey were mostly dipterans with a few hymenopter- ans and lepidopterans, and ranged from 1-15 mm in length (dry mass = 0.17-25.87 mg) with a majority of prey being small (median length = 2 mm; median dry mass = 0.32 mg; Fig. 4). Few prey were >10 mm in length (focal colony: 2/23, 8.7%; second colony: 3/23. 13.0%), but these accounted for 46.7% (focal colony) and 72.5% (second colony) of the total prey biomass. Because these data represent a small prey sampling effort, the exact shape of the distributions should be considered with caution, whereas the overall pattern is robust. Prey capture occurred throughout the day: 13/23 (56.5%) captures took place between 0800-1200 hours and the remainder (43.5%) took place between 1300-1700 hours; no prey were captured between 1200-1300. In the focal colony. 1- 0.8 0,6 0.4 0.2H 0 (a) Prey size 6 8 10 Body length (mm) 16 0 Q. 0 1- qI 0.8- 0.6- 0.4- 0.2- 0- (b) Prey biomass 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Dry mass (mg) Figure 4. — Distribution of prey captured by Leiicaiige sp. spiders in terms of a) size and b) biomass. five different spiders captured one prey each and six spiders captured two to four prey each. Spiders occupying webs located in the center of a colony secured 1 1 out of 23 (47.8%) prey captures, and those at the periphery had 12 out of 23 (52.2%) prey captures. We once observed a spider successfully stealing a prey item captured by another spider; no other conflict over prey between spiders was observed. DISCUSSION This study presents the first empirical evidence of a colonial social organization in the recently discovered neotropical spider Leucauge sp. (Aviles et al. 2001). In colonies from the lowland rainforest of Ecuador, spiders maintained individual orb web territories within a framework of shared silk and engaged in individual prey capture. Our population survey showed that the spatial distribution of individuals within colonies is vertically stratified, supporting our hypothesis of a non-random spatial distribution. Colonies were vertically stratified according to spider size so that large spiders positioned themselves closest to the top, medium-sized spiders were found below large ones, and small spiders occupied webs that were closest to the bottom of the colony. Extra-small spiders occurred within the vertical range of small and medium spiders, as they usually did not own a web but instead lived as floaters in the colony’s framework (i.e., an orb-less silk matrix within the colony’s web complex). This non-random spatial distribution of individuals within colonies suggests that the social organization of Leucauge sp. follows a hierarchical arrangement, which may be explained in at least three non- mutually exclusive ways. First, this size-dependent spatial distribution may result from environmental opportunity within a habitat. In Leucauge sp. colonies, larger spiders spin larger orb webs, as with other colonial (e.g. Metabus gravidus Pickard-Cambridge 1899 (Buskirk 1975a); Metepeira incrassata Pickard-Cambridge 1903 (Rayor & Uetz 2000) and usually solitary orb-weaving 450 THE JOURNAL OF ARACHNOLOGY species (e.g., Cytiosa spp. Menge 1866 (Miyashita 1997); Leucauge markma (Taczanowski 1881) (Eberhard 1988); Nephila clavipes (Linnaeus 1767) (Higgins & Buskirk 1992)). The highest positions in a colony may better accommodate large orb webs, causing size-related variation in space occupation. Likewise, habitat parameters such as topography and the availability and orientation of web attachment points may affect spacing patterns within spider colonies. For example, in Metabus gravidus colonies, the average height of individual orb webs depends on the water current of the stream beneath a colony and the distance between the stream banks (Buskirk 1975a). Further research is needed to determine how the physical characteristics of habitats occupied by Leucauge sp. colonies affect their spatial organization. Alternatively, the spatial stratification of individuals within Leucauge sp. colonies may be the result of interactions among colony members. The spatial arrangement of co-occurring spiders within a habitat is commonly thought to reflect competition between individuals for the occupancy of profitable locations (Wise 1993). Along these lines, Herber- stein (1998) showed through a manipulative experiment that competition for habitat space between two co-occurring species of web-building linyphiid spiders leads to vertical stratification of species within the habitat. Likewise, competition among conspecifics may produce spatial structure within a colony. Leborgne & Pasquet (1987) showed that the spatial organization of Zygiella x-notata (Clerck 1757) spiders living in aggregations is density-dependent. At high densities, cohabitation between spiders of different sizes involved modulations in web size because the presence of large spiders with large webs caused smaller individuals to spin smaller webs. Similar competitive interactions correlated with individual differences in age and size may affect spatial structuring in group-living species. In the colonial species Metepeira incrassata, spiders distribute themselves in a size-dependent pattern (Rayor & Uetz 1990). Large females compete for prime positions close to the core of the colony that afford the best protection from predators, whereas smaller immature spiders live closer to the edge of the web complex where both prey and predators are more common. This spatial arrange- ment reflects a tradeoff between the foraging and protective requirements of different age classes and results in a hierarchical distribution across different parts of the web complex based on competition for specific environmental conditions. In Leucauge sp. colonies, individuals may compete for the highest locations, and larger spiders may dominate due to their size advantage, just as in M. incrassata colonies (Rayor & Uetz 2000). To determine if the spatial arrangement of spiders within Leucauge sp. web complexes is based on competitive interactions, we would need to conduct field manipulations of spider size composition within colonies and environmental conditions (e.g., predation pressure) at differ- ent heights above the ground. Why would the highest positions in a web complex be the most coveted ones? One likely explanation is based on the fact that groups of colonial spiders are commonly viewed as ‘foraging societies’ that form to increase individuals’ foraging potentials (Whitehouse & Lubin 2005). In a foraging society, individuals may compete for locations within a colony where prey availability is higher or prey are more profitable. This may be especially true in tropical forests, where insect abundance, diversity and size vary with the height above ground, even at small spatial scales (Stork & Blackburn 1993; Basset et al. 2001). For example, Buskirk (1975a) found important spatial differences in insect species composition and abundance associated with the distance above and around streams occupied by Metabus gravidus spider colonies in tropical riparian habitats. Therefore, Leucauge sp. spiders positioned at different vertical locations within colonies may have access to different insect prey communities. In populations of Leucauge venusta (Walckenaer 1841) spiders from southern Mexico, the vertical distribution of webs on coffee plants is correlated with prey size and availability; large spiders build webs at an average height of 150 cm where prey are larger than at ground level, whereas small spiders build webs close to the ground where prey are smaller but more abundant (Henaut et al. 2006). If rainforest habitats occupied by Leucauge sp. colonies have a similar spatial distribution of prey, large prey would be more common at the top of colonies compared to the bottom, which large adult spiders living close to the top may capture - more easily than smaller subadult or juvenile spiders. Likewise, juveniles living closer to the bottom may have access to small but abundant prey. Another possible explanation for the greater intrinsic value of high positions within a colony invokes the architectural properties of a colonial web complex. Spiders living higher up in a web complex may enjoy more architectural stability because they are less susceptible to sources of physical disturbance such as seasonal changes in water levels that may destroy webs. Differences in the timing of web building between spiders of different sizes may also create heterogeneity in the spatial positioning of individuals. Rayor & Uetz (2000) found evidence for a sequential web-building pattern correlated with spider age and size in the colonial spider M. incrassata, with larger individuals securing prime web sites sooner at the expense of smaller ones. Similarly, the different positions occupied by Leucauge sp. spiders may depend on their age and web-building abilities. We found that individuals actively defend their webs against intruders, suggesting that they may also compete temporally to secure favorable positions. Individuals may then shift positions as they grow to progressively occupy more competitive locations. Further research is needed to examine this hypothesis in more detail. Spatial positioning ultimately depends on compromises between the foraging, protective and structural costs and benefits procured by different locations within of a colonial web complex. For example, locations with a higher incidence of large insects may also be more exposed to predatory insects, from which larger spiders may be better protected and thus afford to live in. Conversely, small spiders may settle where the prey and predator fauna may be more suitable. It is not known whether coloniality provides any individual fitness benefits to Leucauge sp. spiders, such as reduced web- building costs. In a congeneric species found in secondary forests of central Costa Rica, Leucauge mariana, spiders form local aggregations of adults during the dry season and tend to live solitarily the rest of the year (W. G. Eberhard, pers. SALOMON ET AL.— COLONIALITY IN LEUCAUGE SP. SPIDERS 451 comm.; Valerio & Herrero 1977). In these aggregations, individual orbs lack the surrounding tangle lines that support the ^veb and are solidly anchored to the substrate through shared support silk strands. Therefore, the main benefit invoked for group living in these colonies is an economy of silk (Valerio & Herrero 1977). In summary, this study documents the colonial social organization of Leucauge sp. spiders. We showed that colonies follow a size-dependent spatial arrangement with a positive vertical stratification correlated with spider body size. The characterization of a spatial pattern within colonial spider groups provides further evidence that social groups are organized in specific ways to meet environmental challenges, and provides insight into the forces that shape the evolution of social systems. Future research should determine the under- lying causes and mechanisms responsible for this observed spatial structure by conducting manipulative experiments and ecological studies. ACKNOWLEDGMENTS Data for this project were collected as part of a Field Ecology Methods course of the University of British Columbia. We thank the Department of Zoology at UBC for subsidizing this course, the Jatun Sacha Biological Station for logistic support, the Ecuadorian Ministry of the Environ- ment for research permits and the corporacion Sociedad para la Investigacion y Monitoreo de la Biodiversidad Ecuatoriana (SIMBIOE) for sponsorship. We also thank L. Higgins and two anonymous referees for their useful comments on the manuscript. LA was funded by a NSERC Discovery grant and MS by a postdoctoral fellowship from the Swiss National Science Foundation. LITERATURE CITED Aviles, L. 1997. Causes and consequences of cooperation and permanent-sociality in spiders. Pp. 476-498. In The Evolution of Social Behavior in Insects and Arachnids (J.C. Choe & B.J. Crespi, eds.). Cambridge University Press, Cambridge, UK. Aviles, L., W.P. Maddison, P.A. Salazar, G. Estevez, P. Tufino & G. Canas. 2001. Social spiders of the Ecuadorian Amazonia, with notes on previously undescribed social species. Revista Chilena de Historia Natural 74:619-638. Basset, Y., H.-P. Aberlenc, E. Barrios, G. Curletti, J.-M. Berenger, J.-P. Vesco, P. Causse, A. Haug, A.-S. Hennion, L. Lesobre, F. Marques & R. O’Meara. 2001. Stratification and diel activity of arthropods in a lowland rainforest in Gabon. Biological Journal of the Linnean Society 72:585-607. Buskirk, R.E. 1975a. Coloniality, activity patterns and feeding in a tropical orb-weaving spider. Ecology 56:1314—1328. Buskirk, R.E. 1975b. Aggressive display and orb defense in a colonial spider. Metabus gravidus. Animal Behaviour 23:560-567. Eberhard, W.G. 1988. Behavioral flexibility in orb web construction: effects of supplies in different silk glands and spider size and weight. Journal of Arachnology 16:295-302. Fernandez Campon, F. 2007. Group foraging in the colonial spider Parawixia bistriata (Araneidae): effect of resource levels and prey size. Animal Behaviour 74:1551-1562. Henaut, Y., J.A. Garcia-Ballinas & C. Alauzet. 2006. Variations in web construction in Leucauge venusta (Araneae, Tetragnathidae). Journal of Arachnology 34:234-240. Herberstein, M.E. 1998. Web placement in sympatric linyphiid spiders (Arachnida, Araneae): individual foraging decisions reveal inter-specific competition. Acta Oecologica 19:67-71. Higgins, L.E. & R.E. Buskirk. 1992. A trap-building predator exhibits different tactics for different aspects of foraging behaviour. Animal Behaviour 44:485-499. Hodge, M.A. & G.W. Uetz. 1995. A comparison of agonistic behaviour of colonial web-building spiders from desert and tropical habitats. Animal Behaviour 50:963-972. Jakob, E.M., S.D. Marshall & G.W. Uetz. 1996. Estimating fitness: a comparison of body condition indices. Oikos 77:61-67. Jatun Sacha Foundation. 2009. Online at http://www.jatunsacha.org. (Accessed 10 November 2009). Leborgne, R. & A. Pasquet. 1987. Influences of aggregative behavior on space occupation in the spider Zygiella x-notata (Clerck). Behavioral Ecology and Sociobiology 20:203-208. Lubin, Y. & T. Bilde. 2007. The evolution of sociality in spiders. Advances in the Study of Behavior 37:83-145. Miyashita, T. 1997. Factors affecting the difference in foraging success in three co-existing Cyclosa spiders. Journal of Zoology 242:137-149. Platnick, N.I. 2009. The World Spider Catalog, Version 10.0. American Museum of Natural History, New York. Online at http://research.amnh.org/entomology/spiders/catalog/ Rayor, L.S. & G.W. Uetz. 1990. Trade-offs in foraging success and predation risk with spatial position in colonial spiders. Behavioral Ecology and Sociobiology 27:77-85. Rayor, L.S. & G.W. Uetz. 2000. Age-related sequential web building in the colonial spider Metepeira incrassata (Araneidae): an adaptive spacing strategy. Animal Behaviour 59:1251-1259. Sage, R.D. 1982. Wet and dry-weight estimates of insects and spiders based on length. American Midland Naturalist 108:407^11. Stork, N.E. & T.M. Blackburn. 1993. Abundance, body size and biomass of arthropods in tropical forest. Oikos 67:483^89. Smith, D.R. 1983. Ecological costs and benefits of communal behavior in a presocial spider. Behavioral Ecology and Sociobiol- ogy 13:107-114. Sokal, R.R. & F.J. Rohlf. 1995. Biometry: the Principles and Practice of Statistics in Biological Research. 3rd edition. W.H. Freeman and Co., New York. Uetz, G.W. 1989. The “ricochet effect” and prey capture in colonial spiders. Oecologia 81:154-159. Uetz, G.W. & C.S. Hieber. 1997. Colonial web-building spiders: balancing the costs and benefits of group living. Pp. 458^75. In The Evolution of Social Behavior in Insects and Arachnids (J.C. Choe & B.J. Crespi, eds.). Cambridge University Press, Cam- bridge, UK. Uetz, G.W., J. Boyle, C.S. Hieber & R.S. Wilcox. 2002. Antipredator benefits of group living in colonial web-building spiders: the ‘early warning’ effect. Animal Behaviour 63:445M52. Valerio, C.E. & M.V. Herrero. 1977. Tendencia social en adultos de la arana Leucauge sp. (Araneae, Araneidae) en Costa Rica. Brenesia 10/11:69-76. Wise, D.H. 1993. Spiders in Ecological Webs. Cambridge University Press, Cambridge, UK. Whitehouse, M.E. A. & Y. Lubin. 2005. The functions of societies and the evolution of group living: spider societies as a test case. Biological Reviews 80:347-361. Manuscript received 16 November 2009, revised 24 June 2010. 2010. The Journal of Arachnology 38:452^59 1 Courtship and mating behavior of the wolf spider ScMzocosa Mlineata (Araneae: Lycosidae) j ! Rosanna Vacca^o'•^ George W, Uetz^ and J. Andrew Roberts'-'^: ‘Department of Evolution, Ecology, and Organismal j, ' Biology, The Ohio State University at Newark, Newark, Ohio 43055, USA; ^Department of Biological Sciences, University of Cincinnati, Cincinnati, Ohio 45220, USA | ' Abstract. Of all the wolf spiders {Araneae: Lycosidae), the genus ScMzocosa Chamberlin 1904 is probably the most widely studied, and has become an established model in studies of behavior, sexual selection, and speciation. Much of the work to date involves the complex, often multimodal courtship behaviors and secondary sexual traits used by males to elicit receptivity from potential mates. ScMzocosa Mlineata (Emerton 1885) is one of those species where males possess decorative tufts of bristles on the forelegs that likely play a role in sexual selection, but little is known of male courtship behavior or its role in mate choice. In the present study, we provide the first description of visual and seismic behaviors performed by males in response to female silk and chemical cues, and examine male-female behavioral interactions in live mating trials. Males clearly recognized and responded to female chemical cues by displaying several species-specific visual signaling behaviors accompanied by seismic signals from stridulation. As these behaviors rarely occurred in the absence of female cues, we suggest they function primarily in a courtship context. In live mating trials, females typically responded to male [ courtship with visual receptivity behaviors, which were seen prior to mounting and copulation. While both visual and [ seismic signals of males are clearly implicated in courtship and mate choice, future work will be necessary to fully ! understand the interaction between modalities in this species. The description of behavior provided here should help resolve r the relationship between male ornamentation and courtship behavior in the genus ScMzocosa. | Keywords: Chemical signaling, multimodal communication, sexual selection i Probably the most widely studied of all the wolf spiders (Lycosidae), are members of the genus ScMzocosa Chamberlin 1904. The genus has become an established model for exploring many aspects of evolution and speciation, multi- modal communication, and sexual selection (reviewed in Uetz 2000; Uetz & Roberts 2002; Hebets «fe Papaj 2005; Framenau & Hebets 2007), for which studies of members of the S. ocreata clade (Stratton 2005) have been especially informative. The importance and utility of this genus for scientific study is due not only to the fact that most ScMzocosa species are relatively easy to collect and maintain in the laboratory, but more importantly because males of many species possess complex, sexually selected courtship elements and secondary sexual characteristics (decorative tufts and/or pigmentation) that can be manipulated in a number of ways for study (reviews in Uetz & Roberts 2002; Hebets & Papaj 2005; Stratton 2005; Framenau & Hebets 2O07). In the most recent comprehensive morphological phylogeny of the North American ScMzocosa, Stratton (2005) divided the genus into three major clades: Clade A, containing most of the species from eastern North America as well as the well-studied 5. ocreata clade, and the much smaller Clades B and C, containing many western and southern species. Despite extensive work on several members of this genus, there remain a number of described species for which little or no behavioral data have been collected (Stratton 2005). This is unfortunate as it prevents definitive conclusions about any correlation between male ornamenta- tion and courtship behavior in ScMzocosa (Stratton 2005). Currently, courtship and mating behavior has been de- scribed for all but three of the 17 species contained within Clade A [5. Mlineata (Emerton 1885), S. hiimilis (Banks 1892), ^Present address: U.S. Food and Drug Administration, 580 South High Street Suite 140, Columbus, Ohio 43215, USA. ‘’Corresponding author: E-mail: roberts.762(gosu.edu. and S. segregata Gertsch & Wallace 1937]. ScMzocosa Mlineata is the focus of the present study. Stratton (2005) placed S. Mlineata as a sister taxon to S. crassipalpata Roewer 1951 within Clade A, but outside the S. ocreata clade. This placement is consistent with the more recent molecular phylogeny by Hebets and Vink (2007). While specimens of S. Mlineata have turned up periodically in collections, species descriptions, and taxonomic studies (Montgomery 1902, 1904; Chamberlin 1908; Comstock 1912, 1940; Kaston 1948; Don- dale & Redner 1978, 1990; Sierwald et al. 2005; Stratton 2005; Finkes et al. 2006; Framenau & Hebets 2007; Hebets & Vink 2007), they have otherwise received little attention. A unique opportunity to address this lack of information arose when we discovered a sizable population of S. Mlineata on and around the campus of The Ohio State University at Newark in Newark, Ohio, USA in May 2006. ScMzocosa Mlineata is a Neartic species, thought to be widely distributed throughout the eastern part of North America, from Canada south to Georgia and Texas, and from the East coast to as far west as Kansas and Nebraska (Comstock 1940; Kaston 1948; Dondale & Redner 1990; Sierwald et al. 2005; Stratton 2005). Females of this species are light brown to yellow and cryptically colored. Males are also light brown to pale yellow (often lighter than females) but, as in several other ScMzocosa species, have dark tufts of bristles on the tibia of their forelegs at maturity which may play a role in courtship and mate attraction. Montgomery (1903) provided the only known description of courtship and mating behavior for this species based on direct observation, in which he specified that he could find no evidence of visual courtship from males prior to mounting and copulation. Based on this finding, Kaston (1936) included S. bilineata in his comparative analysis of courtship behavior as an example of a species possessing secondary sexual traits but lacking visual displays (without confirming by direct observation), and this has likely 1 I |i li 452 VACCARO ET AL.— BEHAVIOR OF SCHIZOCOSA BILINEATA 453 hindered further work on the species. In the present study, we first describe visual display behaviors performed by the males during mate searching and courtship, challenging the original assessment of male courtship by Montgomery (1903). We then provide evidence of seismic communication by males, confirm several male visual behaviors that occur primarily in a courtship context in response to female chemical and multimodal cues, and describe visual receptivity behaviors shown by females in live mating trials. METHODS Animal collection and maintenance. — Schizocosa bilineata were collected in open grassy habitats along riparian zones on the campus of The Ohio State University at Newark (OSUN) (40°04.155'N, 82°26.743'W). Spiders were returned to the laboratory and housed individually in opaque, square plastic containers (150 mm X 150 mm X 50 mm, 740 ml), each with a clear lid and containing a short piece of garden hose (250 mm) for shelter. Individuals were fed two or three 10-day old crickets (Ache t a domes ticus) twice a week, and provided ad libitum access to 10 ml of fully hydrated, high molecular weight polyacrylamide gel (Watersorb®, medium crystal polymer) for water and humidity. Spiders were maintained at 24 ± 1° C with a 13:11 h light: dark photoperiod. At the conclusion of each set of experiments, spiders were euthanized by freezing and preserved in 70% ethanol. Voucher specimens are available in the collections of the corresponding author (JAR) and the Denver Museum of Nature and Science. Male Schizocosa are known to respond to silk and chemical cues with courtship behavior (reviewed in Uetz & Roberts 2002; Roberts & Uetz 2004a, b; Roberts «&: Uetz 2005), and we followed the methods outlined in Roberts & Uetz (2005) to elicit male courtship. Specifically, cues were collected from females by placing the female on a piece of filter paper (Fisherbrand® 90 mm diam) inside a clean, glass Petri dish. Females were allowed to deposit silk, chemical cues, and excreta on the filter paper for a period of 24 h, at which time they were returned to their individual containers. Filter paper disks were then used to elicit male behavioral responses. Experiment 1: Observation and description of male behav- iors.— We collected 38 S. bilineata (16 males & 22 females) at OSUN on 25 May-1 June 2006 for use in the description of male behaviors. All individuals were collected as adults, and therefore we had no definitive way to determine mating status or previous experience with conspecifics before starting the experiment. We assumed, therefore, that all individuals had experience with adult conspecifics and that all had likely already mated. Mating status of males was of little concern because male Schizocosa are likely to mate multiply (Norton & Uetz 2005), and will display in response to silk of mated females, though at considerably reduced frequency, rate, and total duration (Norton & Uetz 2005; Roberts & Uetz 2005). We therefore felt confident that we could elicit behaviorally appropriate responses from males, even using the silk of mated females. Females were maintained until they produced (and hatched) egg sacs and/or died naturally, at which time they were preserved in 70% ethanol. Spiderlings from each egg sac were counted to obtain an average number of offspring per egg sac. Females {n = 16) were selected randomly (here and throughout using a random digits table: Rohlf & Sokal 1969) from the 19 collected without egg sacs and were used to collect silk and chemical cues to elicit male behavior. Filter paper disks containing female cues were transferred to clear plastic containers (100 X 100 X 250 mm), and males (n = 16) were gently deposited on these disks from above. Each male was filmed for 20 min using a digital video camera (Sony, Model # DCR-HC42) for later analysis. Observations of male behavior were used only to describe basic behavioral elements for the construction of an ethogram (Table 1), and we made no attempt to determine frequency, rate, and/or duration of male display elements in this experiment. Plastic containers were cleaned between trials using lens paper and 70% ethanol to remove all chemical and silk cues from previous trials, and then allowed to air dry. To explore seismic signals of male S. bilineata, we followed the recording methods of Gibson & Uetz (2008), using a randomly selected set of five males and five females from the laboratory population. Females were confined to a small area (130 X 70 mm) on a poster board substrate for 24 h to deposit cues, after which they were returned to their containers. In each recording trial, we placed the poster board substrate on a non-conductive block of carpet foam on top of a heavy table within a sound isolation chamber. An acetate ring (100 mm diam) was placed over the area containing female cues, and a male was gently placed into the apparatus from above. We utilized acetate because it is transparent, allowing direct observation of male behavior, and light enough to reduce detrimental surface loading that might interfere with seismic signal transmission. Males that began courting in response to female cues were recorded for 30 s blocks using a laser Doppler vibrometer (Polytech PI, Model # PDV-100) set to a sampling rate of 12.5 kHz, a four-channel analyzer and software set to 48 dB gain (Oros Inc., Dulles, VA, USA, Model # OR24), and a laptop computer (Dell Inspiron 4100). Experiment 2: Male response to conspecific cues. — Scliizo- cosa bilineata were hand collected as juveniles in late March and early April 2007 at OSUN to ensure that all experimental individuals were virgin at the time of the study and to control for any experience with adult conspecifics that might influence behavior (Hebets 2003). Spiders were returned to the laboratory and maintained as in Experiment 1, except that individuals were checked daily for molts to obtain an exact date of maturity. All spiders used in this experiment were between one and three weeks of maturity to maximize courtship response (Roberts & Uetz 2005). We randomly selected male spiders from the laboratory population and placed them into one of four cue treatments (resulting in slightly unequal sample sizes) as follows: “no-cue control” where males were exposed to blank filter paper (n = 10), “male silk-cues” where males were exposed to male silk and chemical cues on filter paper (n = 8), “female silk-cues” where males were exposed to female silk and chemical cues [n = 10), or “female multi-cues” where males were exposed to female silk and chemical cues as well as any potential visual and/or seismic cues from live females corralled in the same apparatus {n = 10). Additional males and females were selected randomly from the laboratory population to serve as stimulus individuals. We collected silk, chemical cues, and excreta from stimulus individuals as in Experiment 1 . In each 454 THE JOURNAL OF ARACHNOLOGY Table 1. — Ethogram of behaviors performed by Schizocosa hilineata. Behavior Description Male behaviors Chemoexplore Active exploratory behavior where anterior, lateral surfaces of pedipalps are brushed on the substrate in rapid succession (adapted from Tietjen 1977; Stratton & Uetz 1986) Quick Tap One or (rarely) both extended forelegs, is/are very quickly dropped toward the substrate from an above parallel position (in live observation often perceived as a flicker of motion), often striking the substrate; simultaneous downward motion of the distal abdomen; generally performed while Stationary, but may be produced in combination with Incremental Leg Descend or Slow Jerky Walk Incremental Leg Descend One or (rarely) both forelegs is/are partially or fully flexed, extended vertical to the substrate, then slowly lowered (while extended) in a series of slow, short, incremental movements; generally from a stationary position, sometimes interrupted by, or culminating in, one or more Quick Taps Slow forward locomotion characterized by short, jerky, forward leg movements; may be produced independently or in combination with chemoexploratory behavior Female lowers body to the substrate, often with forelegs extended anteriorly Generally from a stationary position, female slowly turns body either clockwise or anticlockwise one-third to almost one full turn (also called Pivot, see Miller et al. 1998) Directed locomotion toward the stimulus Legs (or pedipalps) are drawn through the chelicerae (both sexes) and/or legs are brushed together rapidly _ (males) One or more legs are raised above parallel with the cephalothorax and then lowered without striking the ground Walking with no other behaviors expressed Turning the body to direct the posterior median eyes toward the stimulus Directed movement away from the stimulus Motionless with no other behaviors expressed Both forelegs are raised above parallel with the cephalothorax in the direction of a stimulus; often culminates in approach or lunge One foreleg (or the first pair of legs), fully extended, raised above the cephalothorax and then lowered back to the substrate; in females this sometimes precedes Settle or Slow Turn Slow Jerky Walk Female behaviors Settle Slow Turn Shared behaviors Approach Groom Leg Raise Locomotion Orient Retreat Stationary Threat Display Wave “female multi-cues” trial, the live stimulus female was the same individual from which the silk and chemical cues were collected, and the female was corralled within a transparent acetate ring (25 mm dia.) on the filter paper containing her own silk and chemical cues. This prevented direct contact between the male and female, but specifically allowed transmission of all other multimodal signals/cues that might play a role in courtship. Using the same clear plastic containers ( 10 X 100 X 250 mm) from Experiment 1, we gently deposited males onto filter paper disks from above and filmed for 5 min using a digital video camera (Sony DCR-HC42) for later analysis. Male behavior was scored according to the ethogram developed in Experiment 1 (see Table 1). We determined the frequency (total number of bouts per 300 s trial) and total duration of male behaviors described in Table 1 using JWatcher (Version 1.0), a behavioral analysis software package freely available for download from the University of California, Los Angeles. One behavior (Quick Tap - see Table 1) was scored only for frequency, as it occurs too quickly to establish a precise duration for each bout and often occurs during bouts of other behaviors. Frequency results were square root transformed, and total duration results were log transformed for analysis using the statistical software JMP version 7 (SAS Institute, Cary, North Carolina). We analyzed results by ANOVA using the Bonferroni-adjusted critical value (a = 0.005) in all significance tests to account for multiple comparisons (Shaffer 1995). Post hoc comparisons among treatments were conduct- ed using Tukey-Kramer HSD tests (Zar 1999). Experiment 3; Female receptivity and mating behavior. — Following the completion of Experiment 2, we selected males {n = 12) and females (n = 12) randomly from the laboratory population and paired them arbitrarily for mating trials. Males all had prior experience with female silk and chemical cues, but not with live females. Females had no previous (adult) experience with conspecifics. Individuals used in this experiment were between 2 and 6 wk of maturity. The translucent plastic arenas (140 X 130 X 100 mm) used in this experiment were filled to a depth of 20 mm with white sand (Quikrete, #1113) to provide a semi-natural substrate with high contrast for filming. Males and females were fed two 10- day old crickets {Acheta domesticus) 48 h prior to the start of experiments to standardize hunger levels, and females were then placed into the apparatus 24 h before introduction of a male to deposit silk and chemical cues. At the start of a trial, males were gently deposited into the arena in the corner most nearly opposite the location of the female. We filmed pairs from above for one hour using a video camera (Watec, model # 902H2) wired to a remote recording device (Sony Digital Videocassette Recorder, model # DSR-11) for later analysis. At the end of the recording period all individuals not in copula (or cannibalized) were returned to their individual containers. Mating pairs were allowed to separate naturally and (survivors) were then returned to individual containers. VACCARO ET AL.— BEHAVIOR OF SCHIZOCOSA Bill NEAT A 455 Figure 1. — Waveform and spectrogram of male Schizocosa hili- neata seismic signaling. RESULTS Experiment 1: Observation and description of male behav- iors.— Basic behaviors performed by male S. bilineata in response to female cues are similar to behaviors seen in other Schizocosa species (e.g., chemoexplore), but three behaviors are described which we propose as species-specific visual courtship behaviors (Incremental Leg Descend, Slow Jerky Walk, and Quick Tap). Table 1 contains a full description of all male and female behavioral elements. Further, based on recordings of seismic signals, males of this species pro- duce distinct vibratory signals (Fig. 1). Seismic signaling bouts by male S. bilineata occur as short bursts of vibration, approximately one second in duration, at regular 3-5 s intervals. Each individual burst is a pulse train consisting of four pulses of stridulation, and a pulse lasts approximately 0.1 s. Though stridulatory pulses represent a sound spectrum of 1-3 kHz (with harmonics to 6 kHz), most sound energy appears to be below 1200 Hz. Unfortunately, we were only able to record brief signaling bouts of two males, so we make no attempt to further analyze aspects of the signal. Of the 22 females collected for this experiment, 20 produced (or were collected with) egg sacs. Of these twenty, 12 egg sacs (60%) successfully hatched with the number of spiderlings per egg sac ranging from two to 63 (mean = 23 ± 5.51 SE). Experiment 2: Male response to conspecific cues. — ANOVA results for frequency and total duration of male behaviors performed in response to conspecific cues are summarized in Table 2. We did not observe some behaviors described in Experiment 1 in this experiment (Leg Raise and Retreat: Table 1), and we excluded these from analysis. Grooming and threat display did not vary significantly across treatment categories for either frequency or total duration (Table 2). Frequency and total duration of male behavior varied significantly by treatment across most other behaviors (Table 2). Orient and approach behaviors occurred only in the presence of a live female (female multi-cues), and wave Table 2. — ANOVA results for behaviors of male Schizocosa bilineata in response to conspecific cues in experiment 2. Key behaviors in bold. Behavior Frequency Total duration ^3,34 P T3,34 P Approach 7.4749 0.0006* 7.0919 0.0008* Chemoexplore 15.7700 < 0.0001* 21.2598 < 0.0001 Groom 0.9787 0.4142 1.9706 0.1369 Locomotion 6.4839 0.0014* 7.7682 0.0004* Orient 8.3509 0.0003* 6.7151 0.0011* Quick Tap 10.5033 < 0.0001* - - Incremental Leg Descend 6.0274 0.0021* 10.6720 < 0.0001 Slow Jerky Walk 11.2184 < 0.0001* 20.9891 < 0.0001* Stationary 2.3067 0.0941 5.5349 0.0033* Threat Display 0.7645 0.5219 0.7393 0.5360 Wave 6.2188 0.0017* 8.5837 0.0002* * Indicates significance using Bonferroni correction (a = 0.005) occurred only in the absence of all conspecific cues (Table 2). The frequency of bouts of stationary behavior was not signif- icantly different by treatment; however, there were significant differences by treatment for total duration (Table 2). Loco- motion varied significantly across treatment categories for both frequency and total duration of behavior (Table 2), and chemoexploratory behavior was also significantly different across treatment categories with bouts occurring at higher frequency and longer duration in response to female silk cues (Table 2, Fig. 2). The behaviors proposed as species-specific courtship be- haviors all varied significantly by treatment for both frequency and total duration (as appropriate) of bouts (Table 2). Quick Tap was only observed in the presence of conspecific cues, but very rarely in response to male silk cues (Fig. 3). Quick Tap was considerably more frequent in response to female silk and multimodal cues, and frequency was not different between these treatments (Fig. 3). Incremental Leg Descend and Slow Jerky Walk both varied significantly by treatment for frequency and total duration of behavioral bouts (Table 2). We observed Incremental Leg Descend across all treatments, but it occurred most frequently in the presence of female cues (Fig. 4a). Bouts of Incremental Leg Descend behavior were of significantly longer duration in the female multi-cues treat- ment, and not significantly different across the other treatment categories (Fig. 4b). We never observed Slow Jerky Walk in response to male cues and only very rarely in the no-cue control treatment (Fig. 5a). Neither frequency nor total duration of Slow Jerky Walk was different between the female silk-cues or multi-cues treatments (Fig. 5), but both frequency and total duration were significantly higher than in the no-cue control or male silk-cue treatments (Fig. 5). Experiment 3: Female receptivity and mating behavior. — Males exhibited chemoexploratory and courtship behaviors in all but one trial, which was excluded from further analysis, since both the male and the female remained stationary for the entire 1 h trial period. Mean latency to begin chemoexplora- tory behavior in the remaining trials was 7.2 s (± 2.84 SE, n = 11), and mean latency to begin courtship (defined as first instance of Quick Tap, Incremental Leg Descend, or Slow Jerky Walk) was 34.5 s (± 5.00 SE, n = 11). Female receptivity 456 THE JOURNAL OF ARACHNOLOGY ^ 20 D O' 1 c o) 15 ^ 10 <D s 5 (a) A 140' (b) B No-Cue Male Female Female Control Silk-Cues Silk-Cues Multi-Cues Figure 2. — Mean + SE a) frequency and b) total duration of bouts of Chemoexplore behavior for male Schizocosa bilineata exposed to blank control, conspecific silk cues, or multimodal female cues. Shared letters above the bars indicate no significant difference between treatment categories by Tukey-Kramer post-hoc analysis (a = 0.05). B ^ -L i li No-Cue Male Female Female Control Silk-Cues Silk-Cues Multi-Cues Figure 4. — Mean + SE a) frequency and b) total duration of bouts of Incremental Leg Descend behavior for male Schizocosa bilineata exposed to blank control, conspecific silk cues, or multimodal female cues. Significant differences for post-hoc analysis indicated as in Fig. 2. No-Cue Male Female Female Control Silk-Cues Silk-Cues Multi-Cues Figure 3. — Mean + SE frequency of bouts of Quick Tap behavior for male Schizocosa bilineata exposed to blank control, conspecific silk cues, or multimodal female cues. Significant differences for post- hoc analysis indicated as in Fig. 2. behaviors were typical of other Schizocosa (summarized in Table 1) and were observed in 8 of 11 trials (72.7%). We observed mounting and copulation in 5 of 11 trials (45.5%) and mean latency to copulate in these trials was 1402.8 s (± 559.2 SE, n = 5). In all cases, males initiated courtship prior to attempting to mount, and females adopted a Settle position prior to mounting by the male. Four of the five females performed at least one Slow Tern prior to Settle. Males mounted females from the anterior or anterior lateral position, and copulation position was of the normal type for Lycosidae with male above on the female dorsum and facing the female posterior (Foelix 1996). We observed two instances of sexual cannibalism by females out of 11 mating trials (18.2%), one pre-copulatory and one post-copulatory. DISCUSSION In the Lycosidae, ornamentation of male forelegs (e.g., tufts of bristles, pigmentation) is generally associated with active leg-waving displays that play a role in visual communication. This seems especially true within the genus Schizocosa (Hebets & Uetz 2000; Stratton 2005; Framenau & Hebets 2007), but a few species remain within the genus for which little or no behavioral data exist (Stratton 2005). Unfortunately, VACCARO ET AL.— BEHAVIOR OF SCHIZOCOSA BILINEATA 457 Male Silk-Cues Female Female Silk-Cues Multi-Cues Figure 5. — Mean + SE a) frequency and b) total duration of bouts of Slow Jerky Walk behavior for male Schizocosa hilineata exposed to blank control, conspecific silk cues, or multimodal female cues. Significant differences for post-hoc analysis indicated as in Fig. 2. this ‘missing data’ has hindered any thorough phylogeny- based exploration of the correlation between leg decoration/ ornamentation and visual display behavior (Stratton 2005). The work presented here represents the first comprehensive analysis of courtship and mating behavior for Schizocosa hilineata, a species whose males possess leg decoration but for which little behavioral work has been done. Results clearly indicate the presence of visual display elements during male mate searching and courtship (Table 1), and refute earlier assertions by Montgomery (1903) that males of this species do not exhibit discernable visual courtship. This is not entirely surprising, as a careful review of methods reveals that the assertion is based on a single, successful male/female pairing (Montgomery 1903). In fact, three visual display behaviors (Quick Tap, Incremental Leg Descend, and Slow Jerky Walk) are described here for S. hilineata males and, as all three were expressed predominantly in the presence of conspecific female cues and prior to copulation in mating trials, any or all could play a role in mate choice. Of the three male visual display behaviors. Quick Tap seems least likely to play a role in visual communication. During the real-time video analysis of male behaviors, we observed Quick Taps only as a ‘flicker’ of motion. This probably corresponds to the “quiver” of leg motion described by Montgomery (1903). Frame-by-frame analysis of bouts of this behavior demonstrate that each bout is conducted within three video frames at the NTSC video standard of 29.97 frames per second (corresponding to an approximate bout length of 0.1 s). Considering the visual system characteristics of wolf spiders, which have flicker fusion rates similar to humans (Land 1985; Uetz 2000), at this speed it seems likely that the behavior would serve as an attention signal or driver of seismic signals, but not function effectively as a courtship signal in and of itself The other two display behaviors. Incremental Leg Descend and Slow Jerky Walk, seem more promising for a role in mate choice. These behaviors are similar to behaviors in other Schizocosa species that have been demonstrated to be important in female mate choice. In particular, Incremental Leg Descend resembles the “Extension” behavior described for S. crassipes (Miller et al. 1998) and Slow Jerky Walk, the “Jerky Walk” behavior described for S. ocreata (Stratton & Uetz 1983, 1986). We stress, however, that the behaviors seen in S. hilineata are distinctly different from these other behaviors. In Incremental Leg Descend, the extended leg is slowly and incrementally lowered (in a series of pauses) to the substrate (never quickly tapped). Slow Jerky Walk is not only slower than Jerky Walk, but also lacks the distinctive cheliceral strikes and leg taps characteristic of bouts of S. ocreata courtship. The assertion of a role for these behaviors in sexual selection, however, is speculative at present, despite the coincidental production of these behaviors immediately prior to copulation and in response to female cues. Additional work will be necessary to explore the actual function of the display traits in conspecific interactions. Kaston (1936) raised an important point concerning sexual traits and display behaviors (even if it was based on the faulty assumption of no visual display in S. hilineata). He emphasized that possession of such traits and behaviors does not necessarily mean that said characteristics are actually important for or directly involved in mate choice. This point is especially important concerning the recent work by Hebets (2008) where she demonstrates that although male S. strklulans Stratton 1991 produce a visual signal, only simultaneously produced seismic signals are important for female choice. We made no attempt to isolate signaling modalities (visual/seismic) in mating trials and as such, we are not able to determine whether visual or seismic cues are of greater or equal importance in this species. We can, however, draw some important conclusions for males of this species. Clearly, as in other Schizocosa species, females signal to males using chemical signals associated with silk (reviewed in Roberts & Uetz 2004a, b; Roberts & Uetz 2005), and signals from chemical cues are lacking in males (Figs. 2-5). Although only rarely performed in the absence of silk and chemical cues, Slow Jerky Walk was never performed in the presence of male silk cues (Fig. 5), and thus may offer evidence for a male inhibitory chemical as suggested by Ayyagari &Tietjen (1987). Male S. hilineata recognize and respond to chemical cues alone with species-specific courtship behavior (Figs. 3-5), and perform bouts of one behavior (Incremental Leg Descend) for a significantly longer duration in response to live females (Fig. 4b). This suggests that males alter their behavior in response to some behavioral feedback from females, though aside from visual receptivity behavior 458 and chemical signals, we did not find evidence of other signaling by females. The significant increase in frequency of Quick Taps in response to female multimodal cues further supports this suggestion. Females, whether responding to visual, seismic, of multimodal male cues, do respond to male courtship with typical receptivity behaviors. These behaviors precede mounting and copulation, and males must recognize female receptivity behaviors and respond accordingly. It is unfortunate that we were unable to fully describe male seismic signals in our study. There is a good chance that these signals play a critical role in courtship and mate choice in this species (Stratton 2005), maybe more so than visual signals alone (Hebets 2008). Future work will be necessary to fully understand the interaction between visual and seismic modalities, but both are likely important and our description of behavior in this species should help resolve the relationship between male ornamentation and courtship behavior in the genus Schtocosa. ACKNOWLEDGMENTS This work was funded by an Ohio State Newark Scholarly Activities Grant (to JAR) and an Ohio State Newark Student Research Grant (to RV). LDV recording was supported by National Science Foundation grant IBN 0239164 (to GWU and JAR). We would like to thank B. Moskalik for help with recording seismic signals and G. Stratton for identifying voucher specimens. We thank B. Angle, S. Crawford, S. Cordasco, and M. Rector for animal care, thoughtful discussion, and comments on draft manuscripts. LITERATURE CITED Ayyagari, L.R. & W.J. Tietjen. 1987. Preliminary isolation of male inhibitory pheromone of the spider Schizocosa ocreata (Araneae, Lycosidae). Journal of Chemical Ecology 13:237-245. Banks, N. 1892. The spider fauna of the Upper Cayuga Lake Basin. Proceedings of the Academy of Natural Sciences of Philadelphia 44:11-81. Chamberlin, R.V. 1904. Notes on generic characters in the Lycosidae. Canadian Entomologist 36:145-148. Chamberlin, R.V. 1908. Revision of the North American spiders of the family Lycosidae. Proceedings of the Academy of Natural Sciences of Philadelphia 60:158-318. Comstock, J.H. 1912. The Spider Book. A Manual for the Study of the Spiders and Their Near Relatives, the Scorpions, Pseudoscor- pions, Whipscorpions, Harvestmen and Other Members of the Class Arachnida, Found in America North of Mexico, with Analytical Keys for Their Classification and Popular Accounts of Their Habits. Garden City, New York. Comstock, J.H. 1940. The Spider Book. A Manual for the Study of the Spiders and Their Near Relatives, the Scorpions, Pseudoscor- pions, Whipscorpions, Harvestmen and Other Members of the Class Arachnida, Found in America North of Mexico, with Analytical Keys for Their Classification and Popular Accounts of Their Habits. (W.J. Gertsch, ed.). Doubleday, Doran & Company, Inc, New York. Dondale, C.D. & J.H. Redner. 1978. Revision of the Nearctic wolf spider genus Schizocosa (Araneida: Lycosidae). Canadian Ento- mologist 1 10:143-181. Dondale, C.D. & J.H. Redner. 1990. The Insects and Arachnids of Canada, Part 17. The Wolf Spiders, Nurseryweb Spiders, and Lynx Spiders of Canada and Alaska. Biosystematics Research Centre, Ottawa. Emerton, J.H. 1885. New England Lycosidae. Transactions of the Connecticut Academy of Arts and Sciences 6:481-505. THE JOURNAL OF ARACHNOLOGY Finkes, L.K., A.B. Cady, J.C. Mulroy, K. Clay & J.A. Rudgers. 2006. Plant-fungus mutualism affects spider composition in successional fields. Ecology Letters 9:344—353. Foelix, R.F. 1996. The Biology of Spiders. Second edition. Oxford University Press, New York. Framenau, V.W. & E.A. Hebets. 2007. A review of leg ornamentation in male wolf spiders, with the description of a new species from Australia, Artoria schisocoides (Araneae, Lycosidae). Journal of Arachnology 35:89-101. Gertsch, W.J. & H.K. Wallace. 1937. New American Lycosidae with notes on other species. American Museum Novitates 919: 1-22. Gibson, J.S. & G.W. Uetz. 2008. Seismic communication and mate choice in wolf spiders: Components of male seismic signals and mating success. Animal Behaviour 75:1253-1262. Hebets, E.A. 2003. Subadult experience influences adult mate choice in an arthropod: exposed female wolf spiders prefer males of a familiar phenotype. Proceedings of the National Academy of Sciences USA 100:13390-13395. Hebets, E.A. 2008. Seismic signal dominance in the multimodal courtship display of the wolf spider Schizocosa stridulans Stratton 1991. Behavioral Ecology 19:1250-1257. Hebets, E.A. & D.R. Papaj. 2005. Complex signal function: developing a framework of testable hypotheses. Behavioral ' Ecology and Sociobiology 57:197-214. Hebets, E.A. & G.W. Uetz. 2000. Leg ornamentation and the efficacy of courtship display in four species of wolf spider (Araneae: Lycosidae). Behavioral Ecology and Sociobiology !■ 47:280-286. Hebets, E.A. & C.J. Vink. 2007. Experience leads to preference: experienced females prefer brush-legged males in a population of i syntopic wolf spiders. Behavioral Ecology 18:1010-1020. 1 Kaston, B.J. 1936. The senses involved in the courtship of some p vagabond spiders. Entomologica Americana 16:97-167. ! Kaston, B.J. 1948. Spiders of Connecticut. State Geological and 1 Natural History Survey of Connecticut Bulletin 70:1-874. Land, M.F. 1985. The morphology and optics of spider eyes. I Pp. 53-78. In Neurobiology of Arachnids. (F.G. Barth, ed.). | Springer-Verlag, Berlin. ' Miller, G.L., G.E. Stratton, P.R. Miller & E. Hebets. 1998. i Geographical variation in male courtship behaviour and sexual | isolation in wolf spiders of the genus Schizocosa. Animal I Behaviour 56:937-951. Montgomery, T.H. 1902. Descriptions of Lycosidae and Oxyopidae p of Philadelphia and its vicinity. Proceedings of the Academy of j Natural Sciences of Philadelphia 54:534-592. i Montgomery, T.H. 1903. Studies on the habits of spiders, particularly j| those of the mating period. Proceedings of the Academy of Natural ‘ Sciences of Philadelphia 55:59-149. Montgomery, T.H. 1904. Descriptions of North American Araneae of i the families Lycosidae and Pisauridae. Proceedings of the Academy > ' of Natural Sciences of Philadelphia 56:261-323. Norton, S. & G.W. Uetz. 2005. Mating frequency in Schizocosa ocreata (Hentz) wolf spiders: evidence for a mating system with , female monandry and male polygyny. Journal of Arachnology ' ) 33:16-24. Roberts, J.A. & G.W. Uetz. 2004a. Chemical signaling in a wolf ' I spider: a test of ethospecies discrimination. Journal of Chemical j ^ Ecology 30:1271-1284. Roberts, J.A. & G.W. Uetz. 2004b. Species-specificity of chemical : ) signals: silk source affects discrimination in a wolf spider (Araneae: 1 1 Lycosidae). Journal of Insect Behavior 17:477^91. '< Roberts, J.A. & G.W. Uetz. 2005. Information content of female chemical signals in the wolf spider, Schizocosa ocreata: male discrimination of reproductive state and receptivity. Animal ' Behaviour 70:217-223. i VACCARO ET AL.— BEHAVIOR OF SCHIZOCOSA BILINE ATA Roewer, C.F. 1951. Neue namen einiger Araneen-Arten. Abhandlun- gen des naturwissenschaftlichen Vereins Bremen 32:437-456. Rohlf, F.J. & R.R. Sokal. 1969. Statistical Tables. W.H. Freeman and Company, San Francisco. Shaffer, J.P. 1995. Multiple hypothesis testing. Annual Review of Psychology 46:561-584. Sierwald, P., M.L. Draney, T. Trentice, F. Pascoe, N. Sandlin, E.M. Lehman, V. Medland & J. Louderman. 2005. The spider species of the Great Lakes states. Proceedings of the Indiana Academy of Science 1 14:1 1 1-206. Stratton, G.E. 1991. A new species of wolf spider, Schizocosa stridulam (Araneae, Lycosidae). Journal of Arachnology 19:29-39. Stratton, G.E. 2005. Evolution of ornamentation and courtship behavior in Schizocosa: insights from a phylogeny based on morphology (Araneae, Lycosidae). Journal of Arachnology 33:347-376. Stratton, G.E. & G.W. Uetz. 1983. Communication via substratum- coupled stridulation and reproductive isolation in wolf spiders (Araneae; Lycosidae). Animal Behaviour 31:164-172. 459 Stratton, G.E. & G.W. Uetz. 1986. The inheritance of courtship behavior and its role as a reproductive isolating mechanism in two species of Schizocosa wolf spiders (Araneae; Lycosidae). Evolution 40:129-141. Tietjen, W.J. 1977. Dragline-following by male lycosid spiders. Psyche 84:165-178. Uetz, G.W. 2000. Signals and multi-modal signalling in spider communication. Pp. 387^05. In Animal Signals. Signalling and Signal Design in Animal Communication.. (Y. Espmark, T. Amundsen & G. Rosenqvist, eds.). Tapir Academic Press, Trondheim, Norway. Uetz, G.W. & J.A. Roberts. 2002. Multi-sensory cues and multi- modal communication in spiders: insights from video/audio playback studies. Brain Behavior and Evolution 59:222-230. Zar, J.H. 1999. Biostatistical Analysis. Prentice-Hall, Upper Saddle River, New Jersey. Manuscript received 22 December 2009, revised 27 June 2010. 2010. The Journal of Arachnology 38:460^65 A new species of Santinezia (Opiliones: Cranaidae) from Panama Victor R. Townsend, Jr.: Department of Biology, Virginia Wesleyan College, 1584 Wesleyan Drive, Norfolk, Virginia 23502, USA. E-mail: vtownsend@vwc.edu Marc A. Milne: Department of Biological Sciences, Old Dominion University, Norfolk, Virginia 23529, USA Abstract. A new species, Santinezia noctiscansor, most similar to members of the S. festae group, is described from central Panama; this species is the only member of the genus that occurs in Central America. Our examinations of adults and nymphs collected from Code and Panama Provinces revealed that males are sexually mature as penultimate nymphs. The subadult strongly resembles the adult with respect to coloration, scutal outline, armature of femora III-IV, and penis morphology, but differs in having fewer tarsomeres on each leg, a smaller body size (scutal length and width), and a much smaller tubercle on the ventral surface of coxa IV. Natural history observations are provided for specimens collected from the field site in Code Province. Keywords: Harvestman, Laniatores, natural history, Neotropics In the catalogue of the Laniatores of the New World, Kury (2003) listed only three species of cranaids (all members of the subfamily Cranainae) occurring in Central America: Coin- boyus alhlineatus Roewer 1943 and Phareicranaus inagnus (Roewer 1932) from Panama, and P. ornatus Roewer 1932 from Costa Rica and Panama. In Pinto-da-Rocha & Kury (2003), P. magmis was transferred to Neocranaus. In a brief review of the geographic distribution of the Cranaidae, Pinto- da-Rocha & Kury (2007) stated, but did not specify, that only two species in the family occur in Central America (Panama and Costa Rica). Given that C. alhilineatus was not examined by Pinto-da-Rocha & Kury (2003), it seems likely that this species was overlooked in the later review. In this study, we describe Santinezia noctiscansor based upon examinations of adult and subadult specimens collected from two sites in Panama. The genus Santinezia consists of 21 species of medium to large harvestmen (scutal length: 6-16 mm) that occur primarily in northern South America (Pinto-da-Rocha & Kury 2003, 2007). Males possess a large tubercle on the ventral surface of coxa IV, a feature that represents the most reliable character used for distinguishing between members of the genera Santinezia and Phareicranaus (Pinto-da-Rocha & Kury 2003). In comparison to adult females, cranaid males have more pronounced armature on leg IV (especially the femur, but also the tibia in several species) and somewhat enlarged chelicerae (Pinto-da-Rocha & Kury 2007). During postem- bryonic development, the sexually dimorphic femur IV is observable in the antepenultimate nymph (Townsend et al. 2009). Presently, three species groups {ciirvipes, festae, and gigantea) are distinguished on the bases of the armature on femur IV and tibia IV, the position of the ventral tubercle on coxa IV, and penis morphology (Pinto-da-Rocha & Kury 2003). Little is known about the natural history of these harvestmen. On the Caribbean island of Trinidad, adult S. serratotibialis (a member of the curvipes group) climb vegetation after dusk (Townsend et al. 2008a), occupy diurnal shelters within rotting logs and palm frond sheaths (Townsend et al. 2008a), exhibit parental care through the guarding of eggs (Machado & Warfel 2006) and nymphs (Townsend et al. 2009), and are infrequently parasitized by larval erythraeid mites (Townsend et al. 2008b). METHODS The adult male holotype and subadult male paratype of Santinezia noctiscansor were collected from montane rain- forests in Code Province near the village of El Cope and are deposited in the American Museum of Natural History (AMNH). We examined an adult male paratype and three subadult paratypes of S. noctiscansor from the collections of the Museo de Invertebrados G.B. Fairchild (MIUP) at the Universidad de Panama. All specimens were observed and photographed in 70% ethanol with a digital Leica EZ 4D stereomicroscope. We took measurements with the aid of an image capturing system. Because the taxonomic literature for the Laniatores is rich with potentially ambiguous terms for macroscopic cuticular structures (e.g., spines, tubercles, and apophyses), collectively referred to as “macrosculpture” (Acosta et al. 2007), we employed the terminology suggested by Acosta et al. (2007) with regard to spines (structures that insert into sockets) and tubercles (small, blunt projections that may or may not bear setae). However, we refer to larger structures as “spiniform tubercles” rather than “apophyses” (Acosta et al. 2007) or “spines” (e.g., Goodnight & Goodnight 1947). For comparative purposes, we examined the female holotype of Phareicranuas ornatus Roewer 1932 (RII/ 2597/ 68-42) from Senckenberg Museum, Frankfurt, Germany (SMF) as well as multiple adult and nymphal specimens of Santinezia serratotibialis Roewer 1932 from AMNH. KEY TO THE CRANAIDAE OF CENTRAL AMERICA 1 . Only second free tergite with a pair of median tubercles; sulci I-II with conspicuous white stripes; femur IV of male armed with a large basal pair of spiniform tubercles; tarsus II with 28 segments Comboyus albilineatus Roewer 1943 (Panama) Second and third free tergites with a median pair of tubercles; sulci I-II without conspicuous stripes; femur IV of male not armed with large, basal, spiniform tubercles on the prolateral and retrolateral surfaces; tarsus II with less than 20 segments 2 460 TOWNSEND & MILNE— NEW SPECIES OF SANTINEZIA 461 2. Coxa IV of male armed ventrally with large, median tubercle tarsus I with 9 segments Santinezia noctiscansor new species (Panama) Coxa IV of male ventrally unarmed; tarsus I with less than 9 tarsomeres 3 3. Leg II greater than 50 mm; tarsal formula 8:17:7:7 (Roewer 1932) or 7:11-12:7:7 (Pinto-daRocha and Kury 2003) Neocranaus magniis (Roewer 1932) (Panama) Leg II less than 50 mm; tarsal formula 8:13-14:9:10 Phareicranaus ornatus Roewer 1932 (Costa Rica) SYSTEMATICS Family Cranaidae Roewer 1913 Santinezia Roewer 1923 Inezia Roewer 1913:392 (preoccupied by Inezia Cherrie 1909); Mello-Leitao 1926:39; Mello-Leitao 1932:113 (type species Inezia gigantea Roewer 1913, by monotypy). Santinezia Roewer 1923:553 (replacement name); Mello- Leitao 1932:122; Roewer 1932:289; Mello-Leitao 1935:96; Kastner 1937:389; Soares & Soares 1948:616; Roewer 1963:69; Gonzalez-Sponga 1989:59 (type species Inezia gigantea Roewer 1913); Pinto-da-Rocha & Kury 2003:181; Kury 2003:97. Niebla Roewer 1925:27; Roewer 1932:348; Soares & Soares 1948:610 (type species Niebla festae Roewer 1925, by monotypy); Synonymy established by Pinto-da-Rocha & Kury 2003:181. Chondwcranaus Roewer 1932:341; Soares & Soares 1948:592 (type species Chondrocranaus scrip tus Roewer 1932, by monotypy). Synonymy established by Pinto-da-Rocha & Kury 2003:181. Macuchichola Mello-Leitao 1943:4; Soares & Soares 1948:606 (type species Macuchicola arthrocentrica Mello-Leitao 1943, by original designation. Synonymy established by Pinto-da- Rocha & Kury 2003:181. Carvalholeptes H. Soares 1970:330 (type species Carvalboleptes singularis H. Soares 1970, by original designation). Synon- ymy established by Pinto-da-Rocha & Kury 2003:181. Santinezia noctiscansor new species (Figs. 1-7) Type material. — PANAMA: Code Province'. Holotype male, Parque Nacional General Division Omar Torrijos H., El Cope, 08°49.2'80"N, 80° 5'45.7"’W, 23-28 February 2007, collected by hand along hiking trails at night in montane rainforest by V. Townsend, A. Savitzky, and J. Ray (AMNH). Paratypes: 1 subadult male, same location (AMNH); 1 adult male, Carretera El Llano Carti, Est. Burbayar, Panama Province (MIUP: no coordinates available); 2 subadult males, 1 subadult female, same location, 9-21 July 2007, R.J. Miranda (MIUP). Etymology. — The name of this species is based on the Latin words noctis for “night” and scansor for “climber.” It is a noun in apposition. This name refers to field observations of behavior in which several individuals (adults and nymphs) were observed climbing tree trunks up to heights of 1-2 m above the ground after dusk. Diagnosis. — Large, ventral tubercle on coxa IV located on middle region; rows of subequal tubercles on prolateral and retrolateral surfaces of male femur IV, with retrolateral tubercles slightly larger than prolateral tubercles; femur IV with three distal spiniform tubercles, largest is prolateral, other tubercles ventral; male tibia IV straight and without 8- 12 spiniform tubercles on proximal half; stylus without subdistal pointed apophyses. Superficially, this species resem- bles Phareicranaus ornatus with respect to scutal outline, dorsal pattern, general penis morphology, and the armature of male femora III-IV. However, Santinezia noctiscansor differs from P. ornatus in having a prominent ventral tubercle on male coxae IV (the major character used to distinguish between Phareicranaus and Santinezia). In addition, Santine- zia noctiscansor differs from P. ornatus with respect to body size, the distribution and relative sizes of spines on the mesal surface of the tibia of the pedipalp, and tarsal formula. With respect to genital morphology, the penis of S. noctiscansor has a ventral plate with 3 distal, curved setae and the glans lacks a dorsal process. In contrast, the penis of P. ornatus has 4 distal curved setae and the glans features a dorsal process (Pinto-da- Rocha & Kury 2003). Santinezia noctiscansor should be considered a member of the festae group (Pinto-da-Rocha & Kury 2003) and represents the only known member of this genus that occurs in Panama and Central America. In addition to the position of the large, ventral tubercle on coxa IV in the male, S. noctiscansor also shares similar patterns of armature on trochanter III and the femur of the pedipalp with members of the festae group (Pinto-da-Rocha & Kury 2003). However, in contrast to other species in this group, S. noctiscansor exhibits distinctive morphology with respect to the armature of femur IV of the male. In addition, this species differs from S. arthrocentrica (Mello-Leitao 1943) in the morphology of the stylus and the armature of coxa and tibia IV and differs from S. festae (Roewer 1925) with respect to color pattern and the distribution of tubercles on the carapace. Description. — Male (holotype): Measurements (mm): Dor- sal scutum length 9.32; width 10.07; cephalothorax length 5.12; width 6.93; leg segments (Table 1). Dorsal scutum (Fig. 1): anterior border with a median projection between the chelicerae and a lateral row of 6-7 tubercles on each side. Eye mound with two large, divergent, spiniform tubercles and two smaller granular tubercles on each side, posterior to the larger tubercles. Carapace smooth. Lateral margin of scutum with 1-3 tubercles adjacent to area I and 8-9 tubercles adjacent to areas II-III. Area I with a median pair of larger tubercles and one lateral smaller seta-bearing tubercle on each side; area II with transverse row of seta-bearing tubercles; area III with a median pair of large, spiniform tubercles and three smaller seta-bearing tubercles on each side; area IV smooth; posterior border with three lateral seta-bearing tubercles on each side. Free tergite I with two lateral tubercles on each side; II and III with median pair of larger tubercles, with median tubercles on the second free tergite larger than those on the third. Free tergite II with two lateral tubercles on each side; free tergite III with a single lateral tubercle on each side. Anal operculum with several scattered seta-bearing tubercles. 462 THE JOURNAL OF ARACHNOLOGY Figures 1-5. — Scmtinezki uoctiscansor, adult male holotype: 1. Habitus, dorsal view; 2. Habitus, ventral view; 3. Habitus, lateral view, arrow indicates position of large, ventral spiniform tubercle on coxa IV; 4. Pedipalpus, left, ventral view; 5. Trochanter-tibia of Leg IV, left, dorsal view. Scale bars = 5 mm. Venter (Fig. 2): Coxa I with scattered arrangement of four larger tubercles and 8-12 smaller tubercles; coxa II with median row of six tubercles, two apical tubercles, four anterior tubercles, and 4-5 posterior tubercles; coxa III with median row of 9-10 tubercles, three anterior tubercles, five posterior tubercles; coxa IV with one stout, larger tubercle covered with many scattered setae near mid-ventral portion (Fig. 3) and an additional 24-30 scattered smaller tubercles. All coxal TOWNSEND & MILNE— NEW SPECIES OF SANTINEZIA 463 Figures 6, 7. — Santinezia noctiscansor, adult male holotype, penis. 6. Dorsal view, is = inflatable sac of glans, st = stylus, tr = truncus, vp = ventral plate; 7. Lateral view. Scale bars = 0.5 mm. tubercles bear setae. Stigmatic area and free sternites with few, scattered seta-bearing tubercles. Chelicerae: Basichelicerite with five tubercles; hand with two frontal tubercles; fixed finger with four teeth; moveable finger with four teeth. Table 1. — Length of the leg segments for adult (holotype) and subadult (paratype from El Cope) males of Santinezia noctiscansor. All measurements are in mm. Leg Segment Leg I Leg II Leg 111 Leg IV Adult male Trochanter 1.02 1.33 1.42 2.28 Femur 6.85 14.78 11.45 16.06 Patella 1.55 2.26 2.64 3.19 Tibia 4.05 10.57 6.51 8.40 Metatarsus 7.95 13.35 10.81 14.89 Tarsus 3.34 8.45 5.34 6.17 Total length 24.76 50.74 38.17 50.99 Subadult male Trochanter 0.82 1.28 1.52 2.14 Femur 6.22 13.14 10.79 13.96 Patella 1.31 1.86 2.40 2.73 Tibia 3.71 9.44 6.04 7.80 Metatarsus 6.92 12.16 9.81 13.61 Tarsus 2.95 7.83 4.84 5.44 Total length 21.93 45.71 35.40 45.68 Pedipalpus (Fig. 4): Trochanter length: 1.61 mm; femur length: 4.53 mm; patella length: 2.03 mm; tibia length: 3.69 mm; tarsus length: 4.39 mm; total length: 16.25 mm. Coxa with one ventral tubercle. Trochanter with two ventral and two dorsal tubercles. Femur with ventral row of seven tubercles, retrolateral row of five tubercles, and dorsal row of seven tubercles (most apical tubercle is larger and spiniform). Patella granular with scattered setae. Tibia dorsally granular, ventrally with four ectal (ilii) and four mesal (lili) spines; tarsus dorsally smooth and ventrally with four ectal (lili) and four mesal (lili) spines. Legs: Coxae I-II each with one large dorsal tubercle; coxa III smooth; coxa IV with 4-5 latero-dorsal tubercles and one large spiniform apical tubercle. Trochanter I dorsally smooth and ventrally with one prolateral, one median, and 1-2 retrolateral seta-bearing tubercles; trochanter II with one dorsal, three prolateral, one ventral and three retrolateral seta- bearing tubercles; trochanter III dorsally smooth and ventrally with three prolateral, 3^ median and three retrolateral seta- bearing tubercles, apical retrolateral tubercle is largest; trochanter IV dorsally smooth and ventrally with three prolateral, 5-6 median, and three retrolateral seta-bearing tubercles; the apical prolateral and retrolateral tubercles are larger. Femora I-IV granular with organized rows of setae; femora III-IV with prominent tubercles organized into rows with tubercles on femur IV larger than those on III; femur III basally and with one larger dorsal prolateral and one dorsal 464 THE JOURNAL OF ARACHNOLOGY I retrolateral tubercle and apically with two larger distal retrolateral tubercles; femur IV (Fig. 5) with three larger distal spiniform tubercles on ventral surface, one prolateral and two retrolateral; all three tubercles curved; dorsal row of tubercles with alternating larger and smaller tubercles; tubercles on the retrolateral and prolateral rows largest basally decreasing in size distally. Patellae I-IV granular. Tibiae I-IV granular. Tarsal formula: 9:16:10:11-12, distitarsi I-II with 3 segments. Male genitalia (Figs. 6, 7): Ventral plate subrectangular, deeply cleft, distal corners without flange. Two groups of setae: two straight latero-basal and three curved latero-apical. Gians without dorsal process, but with inflatable sac with many similar folds arranged in a stack. Stylus smooth, slightly curved. Apex bent, without stylar apophysis. Color: Body dark reddish brown and legs dark brown to black. Eye mound and median region of carapace lighter brown than rest of carapace. Large yellow spots occur on margin of abdominal scutum adjacent to areas I-III. Median borders of abdominal scutal areas partially outlined by yellow lines (Fig. 1). Median tubercles on areas I-III brown. Posterior borders of free tergites II-III outlined in yellow. Tubercles on second free tergite yellow. Venter dark brown. Chelicerae mottled, darker than eye mound, but lighter than carapace and legs. Suhadult male paratype: Measurements (mm): Dorsal scute length 7.78; width 8.82; cephalothorax length 4.16; width 6.09; leg segments (Table 1). Dorsal scutum: same as adult holotype. Venter (Fig. 2): same as adult holotype with the exception that the stout, larger tubercle on the ventral surface of coxa IV is less than half the overall size as that of the adult holotype. Chelicerae: same as adult holotype. Pedipalpus: Trochanter length: 1.28 mm; femur length: 3.87 mm; patella length: 1.95 mm; tibia length: 2.90 mm; tarsus length: 4.24 mm; total length: 14.24 mm. Legs: same as adult holotype except tarsal formula: 8:15:9:10. Male genitalia (Figs. 5, 6): same as adult holotype. Color: same as adult holotype. Distribution. — This species is known from specimens col- lected from forests in the central Panamanian provinces of Code and Panama, areas west of the Canal Zone. Habitat. — The adult male holotype and the subadult male paratype from El Code were collected from the external surfaces of trunks and moss-covered buttresses of medium to large trees (0.5-1. 5 m in diameter) along hiking trails in a moderately sloping montane rainforest. Specimens were collected after dark between 2100-2300 h during moderate periods of rainfall. No information regarding habitat was available for the MIUP specimens collected from Panama Province. Remarks. — During our sampling at the El Code field site, we collected adults, subadults, and three other distinct groups of nymphs (based on size and coloration) at the end of February, which corresponds to the early part of the dry season in central Panama. Due to the elevation, the field site was relatively moist and it also rained briefly each evening. The presence of multiple instars in the population indicate that S. noctiscansor is probably sexually active over an extended period of time at this location and that reproduction begins during the later part of the wet season or perhaps earlier. The collection of sexually mature subadults indicates that cranaids share the novel life history pattern exhibited by other members [ of the Grassatores, in which there are two instars, the [ penultimate nymph and adult, in the adult phase (Gnaspini f et al. 2004). 1 The similarities and differences that we observed in the | morphology of the subadult and adult male also raise [ interesting questions concerning the reliability of the identify | cations of cranaids in regions in which Santinezia species are believed to be syntopic with those of Pharekranaus. The most useful character for distinguishing between these genera is the presence of a prominent ventral tubercle on coxa IV of the male (Pinto-da-Rocha & Kury 2003). It is our opinion that subadult male Santinezia probably have, at least occasionally, been mistaken for adult male Pharekranaus due to the inconspicuous nature of the ventral tubercle on coxa IV | during the penultimate nymph stage. In comparing the adult male holotype (El Code Province) and adult male paratype (Panama Province), we noticed several asymmetries in the holotype that were not present in the paratype. For example, in the holotype, there is a paracheliceral projection only on the right side of the anterior scutal margin (Fig. 1). In the paratype, this projection is absent and the two sides are symmetrical. Similarly, in the holotype, the left side had a greater number of tubercles than the right with respect to the abdominal scutal areas and the ventral surfaces of the coxae. In the paratype, the tubercles on each side were equivalent in number or differed by only one. The subadult paratypes were also symmetrical with respect to the anterior scutal margin and the distribution of tubercles. In the field we did not observe any instances of feeding or predation; however, we found that several individuals were ; parasitized externally by larval mites. In the living condition these mites are orange to red, but when preserved, they quickly ^ turn white (within a few weeks when stored in ethanol). We observed two distinct sizes of these mites on the specimens i from El Code, which probably represent different species of parasites. Attachment sites for mites included the dorsal | scutum, ventral surfaces of coxae and free tergites. The [ holotype had a small mite attached to the ventral surface of [ the first coxa, whereas the subadult paratype had two large 'i mites attached to the dorsal scutum near the posterior margin ji of area III. The intensity of infestation was generally 1-2 mites 1 per host. t ■I ACKNOWLEDGMENTS j This research was made possible through the support of a | Batten endowed professorship (VRT) and the VWC science I undergraduate research fund. Assistance in the field was j provided by A. Savitzky and J. Ray. We are grateful to P. ji Jager (SMF), D. Quintero (MIUP), L. Prendini (AMNH), and 1 J. Huff (AMNH) for assistance with the loans of specimens. ; Special thanks to B. Halier (VWC) for assistance with and i suggestions for Latin names. This research was made possible through logistical support provided by ANAM and STRI. , Specimens collected in the field were exported under scientific permit number SE/A-11-07 (VRT). i LITERATURE CITED j Acosta, L.E., A.P. Gonzalez & A.L. Tourinho. 2007. Methods for j taxonomic study. Pp. 494-510. In Harvestmen: The Biology of j TOWNSEND & MILNE— NEW SPECIES OF SANTINEZIA Opiliones. (R. Pinto-da-Rocha, G. Machado & G. Giribet, eds.). Harvard University Press, Cambridge, Massachusetts. Gnaspini, P., M.B. da Silva & F.C. Pioker. 2004. The occurrence of two adult instars among Grassatores (Aracnida: Opiliones) - a new type of life cycle in arachnids. Invertebrate Reproduction and Development 45:29-39. Gonzalez-Sponga, M.A. 1989. Dos especies nuevas de opiliones del genero Santinezia de Venezuela y redescripcion de Santinezia heliae Avram 1983. Boletin Academia Ciencias Fisicas y Naturales de Caracas 49(1 55-1 56):57-77. Goodnight, C.J. & M.L. Goodnight. 1947. Studies on the phalangid fauna of Central America. American Museum Novitates 1340:1-21. Kastner, A. 1937. Chelicerata 7. Ordnung der Arachnida: Opiliones Sundeval Weberknechte. Pp. 300-393. In Handbuch der Zoologie. Volume 3, Number 2. (W. Kukenthal & T. Krumbach, eds.). Walter de Gruyter & Co., Berlin & Leipzig. Kury, A.B. 2003. Annotated catalogue of the Laniatores of the New World (Arachnida, Opiliones). Revista Iberica de Aracnologia 31:1-337. Machado, G. & J. Warfel. 2006. First case of maternal care in the family Cranaidae (Opiliones, Laniatores). Journal of Arachnology 34:269-272. Mello-Leitao, C.F de. 1926. Notas sobre Opiliones Laniatores sul- americanos. Revista do Museu Paulista 14:327-383. Mello-Leitao, C.F. de. 1932. Opilioes do Brasil. Revista do Museu Paulista 17:1-505. Mello-Leitao, C.F. de. 1935. Algumas notas sobre os Laniatores. Archos Museu do Nacional Rio de Janeiro 36:87-1 16. Mello-Leitao, C.F. de. 1943. Aracnidos recogidos en el Ecuador y en el Peru por la senora H. E. Frizell Don. Comunicaciones Zoologicas Museo Historia Natural Montevideo 5:1-8. Pinto-da-Rocha, R. & A.B. Kury. 2003. Phylogenetic analysis of Santinezia with description of five new species (Opiliones, Laniatores, Cranaidae). Journal of Arachnology 31:173-208. Pinto-da-Rocha, R. & A.B. Kury. 2007. Cranaidae. Pp. 185-188. In Harvestmen: The Biology of Opiliones. (R. Pinto-da-Rocha, G. 465 Machado & G. Giribet, eds.). Harvard University Press, Cam- bridge, Massachusetts. Roewer, C.F. 1913. Die Familie der Gonyleptiden der Opiliones- Laniatores. Archiv fiir Naturgeschichte, Berlin, Abteilung A:257^73. Roewer, C.F. 1923. Die Weberknechte der Erde. Systematische Bearbeitung der bisher bekannten Opiliones. Gustav Fischer, Jena. Roewer, C.F. 1925. Opilioniden aus Sud-Amerika. Bolletino dei Musei di Zoologia ed Anatomia Comparata della Universita di Torino, NS 34:1-34. Roewer, C.F. 1932. Weitere Weberknechte VII. (7. Ergiinzung der Weberknechte der Erde, 1923). Archiv fiir Naturgeschichte, Berlin 1:275-350. Roewer, C.F. 1943. Weitere Weberknechte XI: Uber Gonyleptiden. Senckenbergiana 26:12-68. Roewer, C.F. 1963. Opiliones aus Peru und Colombia (Arachnida Arthrogastra aus Peru V). Senckenbergiana Biologica 44:45-72. Soares, B.A.M. & H.E.M. Soares. 1948. Monografia dos generos de opilioes neotropicos I. Arquivos de Zoologia do Estado de Sao Paulo 5:553-636. Soares, H.E.M. 1970. Novas especies de opilioes da Regiao Amazonica (Opiliones, Cosmetidae, Gonyleptidae, Phalangiidae, Stygnidae). Revista Brasileira Biologia 30:323-338. Townsend, V.R. Jr., D.N. Proud & M.K. Moore. 2008a. Harvestmen (Opiliones) of Trinidad, West Indies. Living World 2008:53-65. Townsend, V.R. Jr., D.N. Proud, M.K. Moore, J.A. Tibbetts, J.A. Burns, R.K. Hunter, S.R. Lazarowitz & B.E. Felgenhauer. 2008b. Parasitic and phoretic mites associated with neotropical harvest- men from Trinidad, W. 1. Annals of the Entomological Society of America 101:1026-1032. Townsend, V.R. Jr., N.J. Rana, D.N. Proud, M.K. Moore, P. Rock & B.E. Felgenhauer. 2009. Morphological changes during post- embryonic development in two species of Neotropical harvestmen (Opiliones, Laniatores, Cranaidae). Journal of Morphology 270:1055-1068. Manuscript received I June 2009, revised 16 March 2010. 2010. The Journal of Arachnology 38:466-474 Natural history of the Iberian solifuge Gluvia dorsalis (Solifuges: Daesildae) Martina Hruskova-Martisova and Stano Pekar': Department of Botany and Zoology, Faculty of Sciences, Masaryk University, Kotlafska 2, 611 37 Brno, Czech Republic Pedro Cardoso: Smithsonian Institution, National Museum of Natural History, 10th & Constitution NW, MRC 105, Room E-509, Washington, DC 20560=0105, USA and Azorean Biodiversity Group, University of Azores, Angra do Heroismo, Portugal Abstract. In this paper we present a detailed study of the natural history of Gluvia dorsalis (Latreille 1817), a representative of the family Daesiidae, the only solifuge species known to occur in southwestern Europe. We studied its distribution, habitat preference, circadian activity, seasonal occurrence, burrowing, predatory and post-mating behavior, prey, fecundity, ontogenesis, and sexual dimorphism. Gluvia dorsalis occurs in lowlands across the entire Iberian Peninsula, preferring grassland or similar open-ground habitats with little summer rain. According to pitfall trap data, the species was active on the surface from May until the beginning of November. It is a nocturnal epigean predator, feeding principally on ants and spiders. However, under laboratory conditions, specimens captured and consumed a variety of arthropods. G. dorsalis seems to hide in underground burrows only when molting, overwintering, or laying eggs. Reproduction occurred in early summer, and females usually produced a single egg clutch containing, on average, 84 eggs, and died soon after. Our results indicate that the G. dorsalis is a biennial species. There was sexual dimorphism in several morphological structures that might be used for sex matching in juvenile instars. Keywords: Camel-spider, activity, prey, reproduction, sexual dimorphism Solifuges are one of the most important predators of arid environments (e.g., Polis & McCormick 1986; Cloudsley- Thompson 1977; Punzo 1997). Despite being locally very abundant, their natural history remains largely enigmatic, as researchers have investigated only a very few species so far. Out of 12 families and slightly more than 1,000 solifuge species currently recognized (Harvey 2003), scientists have performed a thorough natural history studies on only eight species in three families: four eremobatid (e.g., Muma 1966a, 1967; Punzo 1997, 1998a); three galeodid (Heymons 1902; Clouds- ley-Thompson 1961a, b; Junqua 1966; Hruskova-Martisova et al. 2007); and one solpugid species (Wharton 1987). Biologists have studied only particular aspects of behavior, such as prey and predatory behavior (e.g.. Dean &. Milton 1991), repro- ductive behavior (Perreti & Willemart 2007), or burrowing behavior (Gore & Cushing 1980). Yet, even these data are largely fragmented. With the exception of three species (Junqua 1966; Muma 1966a; Punzo 1998a), the postembryonic development or life cycle of solifuges is virtually unknown. Only Punzo (1998a) documents the number of instars, duration of intermolt intervals, and lifespan in a single species. Solifuges are restricted to arid regions of the world with the exception of Australia, where they do not occur. In Europe, 18 species from 4 families have been found so far (Harvey 2003). The majority of the species occur in the eastern Mediterra- nean, but a single species, Gluvia dorsalis (Latreille 1817) (Daesiidae), is endemic to the Iberian Peninsula (Fig. 1). This species was originally only described from Spain, but later collectors found it in different places across the Iberian Peninsula (Pablos 1967; de Matos 1978; Schenker 1980; Grosso-Silva 1998). Our aim in this study is to provide a detailed natural history account of G. dorsalis, found to be locally quite abundant, ' Corresponding author. E-mail: pekar@sci.muni.cz particularly in southern Portugal. We know of no other natural history studies of any species of the Daesiidae. We performed field and laboratory investigations in order to predict its distribution, to reveal seasonal and circadian activity, habitat preference, burrowing, foraging and post- mating behavior, fecundity, and ontogenetic development. Further, our goal has been to identify morphological characters correlated with sexual dimorphism in order to find measures that could be used to identify the sex of juvenile specimens. METHODS Field collection. — From February to November we moni- tored pitfall traps (see Cardoso et at. 2004, 2007 for details) in 1 1 different sites across Portugal over three different years: in 2001 in the northern part (Douro Intemacional Nature Park), in 2002 in the central part (Serras de Aire e Candeeiros Nature Park), and in 2003 in the southern part of Portugal (Vale do Guadiana Nature Park). Solifuges occurred in three different habitat types: pseudosteppe or pasture with solitary trees, scrubland, and forest (Table 1). We emptied traps in two-week intervals. Further, we collected solifuges by hand in a garden in Valverde da Mitra (near Evora), southern Portugal. We preserved all captured specimens in 70% ethanol and deposited them in the collection of arachnids at the Department of Botany and Zoology, Masaryk University, Brno. Altogether, we collected 702 individuals of G. dorsalis in the field over the research period, with most of them (78%) coming from southern Portugal, followed by 8% from central, and 14% from northern Portugal. We classified each collected individual to a developmental stage (juvenile or adult) and sex (only adults). Additionally, for each individual, we measured the following four morphological characters: length and the width of propeltidium, number of malleoli (on the ventral side 466 hruskovA-martisovA et al.— natural history of gluvia 467 Figure 1. — Adult male (A) and adult female (B) of Gluvia dorsalis consuming their prey. of coxa, trochanter, and femur of the fourth leg), and width of the fourth malleolus at the distal end. These characters were used to distinguish probable instars, to estimate developmen- tal trajectory, and to find sexually dimorphic characters. We chose these characters because they were found to change during development, and their size is constant within intermolt intervals (Junqua 1966). Potential distribution, — There are a large number of techniques proposed for modeling the predicted potential distribution of species based on presence data for a restricted number of points in space and environmental data for an entire region. In general, they are based on the principle that species are restricted by the environmental conditions they can survive. If we know a number of sites where a species is present, it is possible to create a “suitable climatic envelope” for the species. This climatic envelope is then translated to space through geographic information systems, making it possible to obtain spatial suitable maps for the species. It should be noted, however, that such maps usually only reflect the potential distribution of species, not their true distribution, which may be smaller. This is due to historical (e.g., local extinction, inability to disperse to many regions) or biological (e.g., competition from ecologically similar species) factors. Such factors may limit the effective distribution of species in areas that should be environmentally adequate for their populations. We used the Maxent method (Phillips et al. 2006; Phillips & Dudik 2008) because it uses presence-only data, and because it is considered the most accurate method available (Hernandez et al. 2006). The map of distribution is based on Gonzalez-Moline et al. (2008) and unpublished data from the authors (total of 254 records). One-fifth of the records were randomly chosen for a test dataset, not used in the training of the algorithm. The spatial variables used in the modeling were eight bioclimatic variables available from the Worldclim database (Hijmans et al. 2005): annual precipita- tion, precipitation of the driest and wettest months, precipi- tation seasonality (coefficient of variation of monthly values), average annual temperature, maximum temperature of the warmest month, minimum temperature of the coldest month, and temperature seasonality (standard deviation of monthly values). Additionally, we used a digital elevation model. All variables had a resolution of one square kilometer. Field observations. — We investigated circadian activity in June 2005 in Valverde da Mitra, Portugal, for 9 days by censusing active solifuges for 20 min/h from 05:00 to 03:00 h in four 100 m^ areas. We observed burrowing and prey hunting behaviors in July and September 2006 at the same site. We obtained data on potential prey by censusing represen- tatives of invertebrate orders occurring on the ground during the major activity period (20.00-24.00 h) on 3 days in June in two 100 m^ areas. Additionally, we observed natural prey by witnessing prey capture, inspecting the prey, and identifying its remnants to order. Only 15 individuals of 196 inspected were found to have prey in their chelicerae. Laboratory observations. — To determine diet breadth, we brought 62 individuals (20 males, 20 females, 22 juveniles) into the laboratory and put them in individual dishes (8 cm diameter, 4 cm tall) with a thin layer of sand substrate and a lightly moistened piece of gauze. We offered solifuges various arthropods (Table 2) found in the study site. The prey was Table 1. — Overview of 11 study sites in three nature parks in Portugal according to their habitat classification. Nature park Habitat type Douro Internacional Serras de Aire e Candeeiros Vale do Guadiana Pseudosteppe or pasture Vila Cha da Braciosa - Algodor Braciais Scrubland Picote Vale Garcia Corredoura Ribeira de Limas Forest Fonte d’Aldeia Cerro das Antenas, Sao Domingos, Perimetro Florestal de Mertola 468 THE JOURNAL OF ARACHNOLOGY Table 2. — Attack and feeding frequencies on a variety of prey offered to large juvenile and adult specimens of G. dorsalis in the laboratory, n = sample size, Size = length of prey. Order Family, Genus, Species n Size [mm] Attack [%] Feeding [%] Acari Trombidium sp. 10 2.5-3.5 20 0 Solifuges G. dorsalis 38 15-22 100 100 Araneae Agelenidae, Araneidae, Corinnidae, Filistatidae, Gnaphosidae, Lycosidae, Theridiidae 62 15-22 83.9 83.9 Opiliones Leiobimum sp. 5 2 100 100 Isopoda Armadilliditim sp. 10 7-11 20 0 Porcelio sp. 10 10 70 70 Diplopoda Julidae 14 16-21 16.7 0 Thysanura Lepisma sp. 9 7.5-9 77.8 77.8 Ephemeroptera unidentified 11 7.5-11 27.3 27.3 Dermaptera Forficula sp. 2 9 100 100 Ensifera Acheta domestica 52 6-25 100 100 Caelifera unidentified 45 5 93.3 93.3 Isoptera Reticulitermes sp. 10 3.5^ 80 80 Heteroptera Coreus sp., Lygus sp. 35 ^12 88.6 80 Coleoptera Tenebrio molitor larvae 41 5 100 100 Brachinus crepitans adults 12 9 0 0 other Carabidae 42 4.5-16 47.6 45.2 Diptera Muscidae 10 5.5-9 90 90 Hymenoptera Tapinonui sp. 8 2-3 100 100 Camponotus sp. 10 9-13 20 0 Messor sp. 17 6-9 64.7 64.7 Lepidoptera Noctuidae adults 10 8.5-15 60 50 offered to each solifuge in a random order in one-day intervals. For each feeding trial we recorded the size of prey, whether it was attacked and consumed, and noted the predatory behavior. We replaced a prey item that was not attacked within 5 min with another one. Once the prey was consumed, we cleaned the prey remnants from the dishes. Adult males {n = 39) and adult females (« = 21) were captured in the field in June and then mated in the laboratory (see Hruskova-Martisova et al. 2010). We paired each female successively with one to four males until it mated. After successful copulation, we placed female solifuges singly in vials (5 cm diameter, 8 cm tall) with a sand layer (2 cm deep) with darkened sides, to lay eggs. The vials were kept in a chamber at 23° C (humidity was not measured) until hatching. The sand substrate was moistened once a week by few drops of water. We checked egg clutches every day during the three following months. We counted and measured (diameter) 10 eggs in each clutch. Larvae were kept in conditions similar to the eggs until molting to the first instar, for which we measured morpho- logical characters similarly to the adults (see above). Data analysis. — We studied the effect of average monthly precipitation and average monthly temperature on seasonal activity using multiple regression within Generalized Linear Models with Poisson errors (GLM). Data on long-term average monthly temperatures and average monthly precipi- tation were those for Evora, because most of the material came from Alentejo and this was the nearest place for which meteorological data were available at www.worldclimate.com. We compared the relative composition of potential and actual prey using chi-square tests. We used the Spearman correlation to study the relationship between the size of eggs and the clutch size. For the analyses of sexually dimorphic characters, we used specimens from all study sites, because there was no significant difference in the propeltidium width of adult males or females between locations (ANOVA, F < 0.8, P > 0.46). Models of the relationship between selected morphological ! characters for juveniles, males, and females were compared by means of ANCOVA. We performed all analyses in R (R Development Core Team 2009). We present means and their standard errors in the text. We estimated confidence intervals (CI95) using normal approximation. i RESULTS Distribution and habitat preference. — Gluvia dorsalis sped- ‘ mens have been collected throughout most of the Iberian Peninsula (Fig. 2). This is reflected in its predicted presence in j all but the northern and northwestern regions and the highest : mountain chains. As previously mentioned, the output of Maxent, or any other technique, are maps of predicted distribution based on current climate suitability and not realized distributions, which can be smaller due to historical or biological reasons. The AUC (area under curve) of the ROC (Receiver Operating Characteristic) plot of the test dataset was 0.815 in a scale of 0 to 1, where 0.5 represents a random . prediction, indicating that the maps can be considered as \ relatively accurate. The variable found to be most strongly I driving the species distribution was the precipitation of the | driest month. According to our abundance data collected in Portugal, the I species shows a strong preference for open-grass plains with i low vegetation, such as pseudosteppes or pastures (55%, n = 643) or slopes with low-height shrub cover (39%). Only 6% of the individuals were found in forested (Quercus or Eucalytus) areas. Seasonal and circadian activity. — Solifuges were found to be active from late May until the beginning of November (Fig. 3). Their activity was negatively correlated with monthly average precipitation (GLM, = 227, df = I, P < 0.0001, « HRUSKOVA-MARTISOVA ET AL.— natural history of GLUVIA 469 Figure 2. — Potential distribution of G. dorsalis on the Iberian Peninsula. Darker shades represent higher suitability for the species. Points represent actual records. Figure 3. — Seasonal activity (i.e., total abundance per month [bars]) of G. dorsalis and the mean monthly precipitation (points). Fig. 3) and positively correlated with monthly average temperature (GLM, X" = 8.5, df = \, P = 0.004). The adult males and females appeared in late May and were most abundant in June and July. Males disappeared in mid-August, and females disappeared a month later. The average sex ratio (males: females) was highly skewed in favor of females (0.38, n = 157) in the field (Binomial test, P = 0.004). G. dorsalis are strictly nocturnal (Fig. 4). Individuals emerge after sunset and are active until midnight, with an activity peak between 21:00 and 22:00 (n = 68). Principal activities observed included prey-capturing or burrow-digging. During the rest of the day solifuges hid in stonewall crevices, in shallow depressions under stones, or in the debris. Two individuals that we observed running on the ground during the day were presumably disturbed when we stepped on the rocks where they had hidden. Burrowing behavior. — We observed females (n — 20) and juveniles (n = 9), but no males, digging burrows between 22:00 and 24:00 h in sandy soil. Solifuges dug and raked the loose sand, mainly with the second pair of legs. Chelicerae, pedipalps, and the first pair of legs were shaped into a basket to push sand heaps aside to a distance of a few centimeters. 470 THE JOURNAL OF ARACHNOLOGY I II Figure 4. — Mean number (±SE) of individuals per hour found during 20 min in Mitra. Time of inactivity (01:00-11:00 h) are not shown. They repeated these movements until completing the burrow. All burrows had open entrances. Prey and prey capture. — Hunting solifuges, juvenile or adult, ran quickly over the ground with thte pedipalps and first pair of legs stretched forward. The movement appeared haphazard, with frequent changes in direction. Once prey was encountered, solifuges seized it quickly by means of the sucking organs on the pedipalpal tips, and subsequently grasped it with the chelicerae. Solifuges grasped bodies of small prey, but legs of large prey (e.g., grasshoppers). The prey sometimes escaped by pulling away from its own leg. In all cases we observed, the solifuges remained at the site of capture while eating the prey. In the field, we found the main available (potential) prey on the ground to be ants (42%, n = 325), woodlice (32%), and beetles (10%), but we observed solifuges catching mainly ants and spiders in such conditions. The captured prey, therefore, differed from the available options {X~ = 47.4, df = 9, P < 0.0001, Fig. 5). In the laboratory, solifuges accepted and consumed a wide variety of arthropods ranging in size from 2 to 25 mm (Table 2), so that the relative size (prey to solifuge) was between 0.2 and 1 . Solifuges rejected or attacked but did not consume mites, some woodlice, millipedes, large ants, and carabid beetles. Male solifuges further rejected hard sclero- tized beetles. In the laboratory, solifuges accepted a wider prey variety than was available in the field. Ovipositfon and ontogenetic development. — In June, on average 11 days (SE = 0.7, n = 11) after mating, females produced one clutch of spherical, whitish eggs. They deposited a mass of eggs on the surface of the sand in the vials. The clutch size varied from 47 to 163 eggs (mean = 83.4, n = 11), and the eggs had an average diameter of 1.78 mm (SE = 0.09, n = 96). The size of the eggs was not related to the clutch size (Spearman correlation, p = —0.3, S — 26, P = 0.68). The females died on average 9 days (SE = 2.5, n = 11) following the oviposition. In one case, the female was observed laying Figure 5. — Comparison of available and captured prey. Bars represent proportion of individuals. There were 3,254 prey individuals f available and 15 prey captured. ; f another clutch (10 eggs) one week after the first one (137 eggs) had been deposited. On average, 56 days (SE = 4, « = 85) after laying eggs, pale ! I yellow immobile larvae hatched. These inactive individuals i It molted into the first non-feeding instar, on average in 17 days I i (SE = 0.9, n = 25). Measurements of the morphological j : characters of the first instar are shown in Table 3. Juveniles 1 3 captured in the field were separated into two ontogenetic categories based on the number of malleoli: as early juveniles i j with 3 malleoli, and later juveniles with 5 malleoli. In the field, > the smallest early instars (propeltidium width < 1 mm) ' occurred from mid-July to mid-October (Fig. 6). The occur- j; rence of early and late juveniles overlapped, as they occurred . every month from May to October. | Because we sampled early juveniles, late juveniles, and adult |j ' stages each month from June to August, we assume that the li . life cycle is biennial, with a longevity of about 700 days. We < suggest two extreme developmental trajectories. Specimens Ij i that hatched into the first instar in mid-July (dashed line in |, i Fig. 6) would probably pass through 2-3 instars before ■ overwintering to reach about 1.8 mm (propeltidium width). > Next season, these specimens would pass through several t " instars to reach about 3.7 mm (propeltidium width) before the |i .’ onset of winter, so that the following season they would reach " maturity in the beginning of June. Specimens that hatched into j the first instar later, in late September, did not molt and |j overwintered at 1.2 mm (propeltidium width). Next season j, they would pass through a few instars to reach 3 mm |! (propeltidium width). In the following season they would f pass through another few instars to reach adulthood in f ■ August. I Sexual dimorphism. — Adult males (n = 55) were smaller and I had a narrower propeltidium and wider malleoli than adult i females (n = 66) (Table 3). Male sclerotized body parts (prosomatic parts and appendages) were densely covered with 4 hruskovA-martisovA et al.— natural history of gluvia 47 Table 3. — Measurements (mean ± SE) of some body traits in the first-instar specimens, adult males and adult females of G. dorsalis. Character First instars {n = 10) Males (n = 55) Females (n = 66) Length of propeltidium [mm] 0.797 ± 0.047 2.024 ± 0.363 2.728 ± 0.441 Width of propeltidium [mm] 1.061 ± 0.087 2.778 ± 0.405 4.495 ± 0.701 Number of malleoli 3 5 5 Width of the fourth malleolus [mm] 0.230 ± 0.040 0.945 ± 0.129 0.835 ± 0.128 Malleolus to propeltidium width ratio 0.216 ± 0.030 0.355 ± 0.038 0.188 ± 0.026 spines, while those of females were not (Fig. 1). The relationship between the width of propeltidium and width of malleoli was isometric for females, but allometric for males (Fig. 7A). The malleoli size of females increased with propeltidium size in similar fashion to juveniles (ANCOVA, ^2.650 = 0-91, P = 0.4). But in males the relationship was significantly steeper than in juveniles (ANCOVA, F2,652 = 1021, i* < 0.0001, Fig. 7A), suggesting that the size of malleoli increased markedly at the last molt of males. The relationship between the width and length of propeltidium was, in turn, isometric for males, but allometric for females (Fig. 7B). For juveniles, this relationship was significantly different from that of females (ANCOVA, = 13.3, P < 0.0001) as well as males (ANCOVA, F2J27 = 72.5, P < 0.0001, Fig. 7B). Clearly, there was a change in the shape of propeltidium at the last molt in both sexes. The ratio of width to length of propeltidium was 1.62 (CI95 = 1.59, 1.65) in females and 1.37 (CI95 = 1.32, 1.41) in males. DISCUSSION We found that Gluvia dorsalis is widespread on the Iberian peninsula, but has a quite restricted seasonal activity. It feeds on a variety of epigean arthropods that are captured only during the first three hours of the scotophase. Its reproduction occurs at the beginning of summer, and the life cycle is probably biennial. Several morphological characters revealed apparent sexual dimorphism at maturity, such as shape of the propeltidium or size of malleoli. We will now compare traits studied with those from other solifuge families, as there are few comparative data. Our results strongly suggest that the distribution of the solifuge G. dorsalis is mostly limited to areas of the Iberian Peninsula with scarce summer rain. This supports the fact that solifuges are restricted to arid environments of the world at variable altitudes and temperatures. Similar to other solifuge species (Griffin 1990; Punzo 1998b), G. dorsalis has a marked preference for habitats of a semi-desert character (Gonzalez- Figure 6. — Relationship between the size of propeltidium (width) and the occurrence of adult males, adult females, late juveniles (with 5 malleoli), and early juveniles (with 3 malleoli) during season from May until October. Bold lines mark the hypothesized borders between three consecutive seasons (1., 2., 3.). The slopes of the bold lines are parallel with two extreme hypothesized developmental trajectories displayed by thin solid (for late hatchlings) and thin dashed (for early hatchlings) lines with arrows. 472 THE JOURNAL OF ARACHNOLOGY Width [mm] Figure 7. — A. Relationship between the width of malleoli and the width of propeltidium for juveniles, adult males and females. B. Relationship between the length and width of propeltidium for juveniles, adult males and females with linear models. Males = solid line, females = gray line, juveniles = dashed line. Moline et al. 2008; this study): sand substrate, rocks, and low and sparse vegetation cover. Females and juveniles can easily dig burrows in the sand, whereas stones, grass and shrubs provide cover and protection for the males. Although the use of burrows as retreats during the periods of inactivity and oviposition is widespread among solifuges (Muma 1966b; Cloudsley-Thompson 1977; Punzo 1998b; Hruskova-Marti- sova et al. 2007), our field observations show that solifuges sheltered mainly under stones, in debris and rock crevices while resting during the day and as a protection during ecdysis. Burrows were excavated exclusively by juveniles and females; thus, we expect that burrows function for overwin- tering and for egg deposition. Similar observations were made by Punzo (1998c) of Eremobates marathoni Muma 1951 (Eremobatidae), where females were found in the non-plugged burrows, while males utilized simple depressions under a rock or decaying vegetation. The burrowing behavior of G. dorsalis was generally similar to that reported for other solifuge species (Kingston 1925; Junqua 1966; Muma 1966b; Cloudsley- Thompson 1977; Gore & Cushing 1980). Their preference for dry habitats is also reflected in their restricted seasonal activity. As revealed by previous studies (Rambla & Barrientos 1986; Gonzalez-Moline et al. 2008), G. dorsalis was active during months with the highest tempera- ture and very few summer rains, like many other solifuge species (Heymons 1902; Cloudsley-Thompson 1961a; Whar- ton 1987; Chandra 1989; Punzo 1997). Although both these variables are strongly correlated, rainfall was found to be a better predictor of their seasonal activity pattern than temperature. It is not clear why rain can limit the seasonal occurrence of solifuges. We have observed that on rainy nights in late summer, the density of their prey, such as ants, decreased dramatically. So solifuges may avoid wet days due to low profitability resulting from a low capture rate. High abundance of prey must be very important for G. dorsalis, because its foraging activity period is very short, lasting only about three hours. Whether such a short period is an adaptation to avoid desiccation or predators is not clear. Berland (1932) and Lawrence (1956), however, suggested that G. dorsalis is a day-active species, because it is small. Nocturnal activity is very common in many other solifuge species (e.g.. Turner 1916; Lawrence 1955; Cloudsley-Thomp- son 1961a; Hruskova-Martisova et al. 2007; Punzo 1998b), as well as many other animals inhabiting arid environments (Cloudsley-Thompson 1991; Rose & Mueller 2006). The majority of solifuge species has been reported to be extremely active and extraordinarily voracious predators (Muma 1966c); however, researchers have made only a few observations on solifuge foraging behavior and feeding habits under natural conditions (Turner 1916; Bolwig 1952; Wharton 1987; Hruskova-Martisova et al. 2007). In most of those studies, juveniles and/or adults were observed to search and hunt their prey on the ground. Our observations indicate that hunting, capturing, and handling prey in G. dorsalis are similar to descriptions for other solifuge species (Hutton 1843; Turner 1916; Bolwig 1952; Cloudsley-Thompson 1961b; Punzo 1995). G. dorsalis is clearly a polyphagous predator, as are other solifuge species (Turner 1916; Muma 1967; Cloudsley- Thompson 1977; Wharton 1987; Punzo 1994, 1997). In the field, the diversity of captured prey was lower than in the laboratory, reflecting the restricted availability. Gluvia dorsalis seems to prefer soft-bodied prey smaller than itself, refusing arthropods with chemical or other types of defenses like in other solifuge species (Punzo 1993, 1994). It is probable that there is frequent cannibalism of small specimens by larger ones. The longevity of solifuges is largely unknown. Some authors have suggested that they are univoltine and do not live for more than one year (Muma 1963, 1966a; Punzo 1998a). But Wharton (1987), using data from pitfall traps and field observations, concluded that Metasolpuga picta (Kraepelin 1899) is biennial. Junqua (1966) observed in the laboratory that Othoes molted once a month. Having 8-10 instars, this I i I hruskovA-martisovA et al.— natural history of gluvia 473 species lived for ~3 yr. Lawrence (1963), studying change in the cheliceral teeth, concluded that Zeria monteiroi (Pocock 1895) live for several years. Our data suggest that the longevity of G. dorsalis is about 2 yr. Although it is as large as E. marathoni (Punzo 1998a), G. dorsalis has longer life cycle, likely due to longer hibernation period. In the laboratory we observed that the embryological development lasted almost 2 mo. We assume that in natural conditions development is much shorter because the ambient temperature is much higher than it was in laboratory. Based upon our previous experience, we avoided using higher temperature to avoid egg desiccation (if kept in dry conditions) or attack by fungi (if kept in humid conditions), both conditions accelerated by higher temperatures. As mating takes place in mid-June and the first instars were found in pitfall traps in mid- July, it appears that the incubation period together with the larval development takes about one month in nature. In a similarly large E. durangonus Roewer 1934, the average duration of intermolt interval was 25 days at 20° C. Assuming a similar duration for G. dorsalis and considering the temperature records over the season in Portugal, we expect that individuals that hatched in mid July would molt about 2- 3 times before overwintering, then continue to molt next season about 4-5 times, finally molting into the adult stage the next season. There would thus be 7-9 juvenile instars altogether. This corresponds closely to the rate observed for E. marathoni (i.e., increase in propeltidium length by 37% on average [Punzo 1998a]). We were not able to distinguish particular juvenile instars due to the large variation in all measured morphological traits. The only discrete character was the number of malleoli: three in early stages and five in later juvenile stages and in adults. Junqua (1966) and Muma (1966a) observed that the juveniles of the first to the fourth instars exhibit three pairs of malleoli and the fifth to the ninth instars possess five pairs. Our data appear to be in agreement with their conclusions. In most of the solifuge species studied to date, males and females differ in body size, in the presence of sex-specific structures, and in shape and size of some body structures (Punzo 1998b). In G. dorsalis, we found conspicuous sexual dimorphism in total body setosity, propeltidium, and malleoli size. We suspect that some features might already be dimorphic in the juvenile stages. However, the size of malleoli and the shape of propeltidium (expressed as width to length ratio) become clearly dimorphic only in the adult stage. However, certain differences in the shape of propeltidium were already apparent in the juvenile stage. This character deserves further investigation because it might be used for the recognition of sex in juvenile instars. By revealing basic natural history data of G. dorsalis, we raise many questions, such as why their foraging activity is so short, what portion of their diet involves cannibalism, or how long is their life cycle. These remain to be addressed in future studies. Unfortunately, the scarcity of individuals in nature and the difficulty of observing and rearing them makes investigation of these remarkable arachnids infeasible. ACKNOWLEDGMENTS We would like to thank C. Meierrose and P. Surovy for kind help with the logistics in Portugal and P. Vahhara for help in the field. We are very grateful to L. Higgins for useful comments on the manuscript and for the improvement of English. The study was supported by grants no. 1P06ME851 and no. 0021622416 of the Ministry of Education, Youth and Sports of the Czech Republic. Fieldwork was also supported by the Portuguese Foundation for Science and Technology (SFRH/BD/1 195/2000). LITERATURE CITED Beiiand, L. 1932. Les Arachnides. In Encyclopedic Entomologique (A). Volume XVI. Pp. 1M85. Paul Lechavalier & Fils, Paris. Bolwig, N. 1952. Observations on the behavior and mode of orientation of hunting Solifugae. Journal of the Entomological Society of Southern Africa 15:239-240. Cardoso, P., 1. Silva, N.G. Oliveira & A.R.M. Serrano. 2004. Higher taxa surrogates of spider (Araneae) diversity and their efficieney in conservation. Biological Conservation 117:453-459. Cardoso, P., I. Silva, N.G. Oliveira & A.R.M. Serrano. 2007. Seasonality of spiders (Araneae) in Mediterranean eeosystems and its implications in the optimum sampling period. Eeological Entomology 32:516-526. Chandra, H. 1989. Some observations on the predatory behaviour of Galeodes sp. (Solifugae: Galeodidae) on acridids. Plant Protection Bulletin (Faridabad) 40:23-24. Cloudsley-Thompson, J.L. 1961a. Observations on the natural history of the ‘camel-spider’, Galeodes arahs C.L. Koch (Solifugae: Galeodidae) in the Sudan. Entomologist’s Monthly Magazine 97:145-152. Cloudsley-Thompson, J.L. 1961b. Some aspects of the physiology and behaviour of Galeodes arahs. Entomologia Experimentalis et Applicata 4:257-263. Cloudsley-Thompson, J.L. 1977. Adaptational biology of Solifugae (Solpugida). Bulletin of the British Arachnological Society 4:61-71. Cloudsley-Thompson, J.L. 1991. Ecophysiology of Desert Arthro- pods and Reptiles. Springer-Verlag, Berlin. Dean, W.R.J. & S.J. Milton. 1991. Prey capture by Solpiiga chelicornis Lichenstein (Solifugae: Solpugidae). Journal of the Entomological Society of South Afriea 54:266-267. Gonzalez-Moline, A.L., A. Melie & J.A. Barrientos. 2008. Taxono- mia, distribuci6n geografica e historia natural del endemismo Iberico Gluvia dorsalis (Latreille, 1817) (Solifugae: Daesiidae). Boletin Sociedad Entomologica Aragonesa 42:385-395. Gore, J.A. & B.C. Cushing. 1980. Observations on temporary foraging areas of the sun spider Amniotrechula peninsulana (Banks) (Arachnida: Solpugida). Southwestern Naturalist 25:95-102. Griffin, E. 1990. Seasonal aetivity, habitat selection and species richness of Solifugae (Arachnida) on the gravel plains of the central Namib Desert. Transvaal Museum Monograph 7:77-82. Grosso-Silva, J.M. 1998. Sobre a presenca do solifugo Gluvia dorsalis (Latreille, 1817) (Arachnida, Solifugae, Daesiidae) no eoncelho de Vila Nova de Foz Coa (NE de Portugal). Boletin de la SEA 21:23. Harvey, M.S. 2003. Catalogue of the Smaller Arachnid Orders of the World. CSIRO Publishing, Collingwood, Victoria, Australia. Hernandez, P.A., C.H. Graham, L.L. Master & D.L. Albert. 2006. The effect of sample size and species characteristics on perfor- mance of different species distribution modelling methods. Ecography 29:773-785. Heymons, R. 1902. Biologische Beobachtungen an asiatischen Solifugen. Abhandlungen der Koniglich Preussischen Akademie der Wissenschaftern 1901:1-65. Hijmans, R.J., S.E. Cameron, J.L. Parra, P.G. Jones & A. Jarvis. 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965-1978. Hingston, R.W.G. 1925. Nature at the Desert’s Edge. Studies and Observations in the Bagdad Oasis. Witherby, London. 474 THE JOURNAL OF ARACHNOLOGY Hruskova-Martisova, M., S. Pekar & A. Gromov. 2007. Biology of Galeodes caspiiis subfuscus (Arachnida: Solifugae). Journal of Arachnology 35:546-550. Hruskova-Martisova, M., S. Pekar & T. Bilde. 2010. Coercive copulation in two sexually cannibalistic camel-spider species (Arachnida: Solifugae). Journal of Zoology. In press. DOI: 10.1111/ j.l469-7998.2010.00718.x. Hutton, T. 1843. Observations on the habits of a large species of Galeodes. Annals and Magazine of Natural History 12:81- 85. Junqua, C. 1966. Recherches biologiques et histophysiologiques sur un solifuge saharien Othoes saharae Panouse. Memoires du Museum National d’Histoire Naturelle Paris, Series A 43:1-124. Lawrence, R.F. 1955. Solifugae, scorpions and Pedipalpi, with checklists and keys to South African families, genera and species. South African Animal Life 1:152-262. Lawrence, R.F. 1956. Solifugae (Chelicerata) from Afghanistan. Videnskabelige Meddelelser fra Danske Naturhistorisk Forening i Kjobenhavn 118:115-140. Lawrence, R.F. 1963. The Solifugae of South West Africa. Cimbebasia 8:1-28. Matos de, A. 1978. Nota sobre a existencia do solifugo Gluvia dorsalis Latr. (Arachnida, Solifugae, Daesiidae). Arquivos Museu Bocage 2:319-323. Muma, M.H. 1963. Solpugida of the Nevada test site. Brigham Young University Science Bulletin, Biological Series 3(2): 1-15. Muma, M.H. 1966a. The life cycle of Eremobates durangonus (Arachnida: Solpugida). Florida Entomologist 49:233-242. Muma, M.H. 1966b. Burrowing habits of North American Solpugida (Arachnida). Psyche 73:251-260. Muma, M.H. 1966c. Feeding behavior of North American Solpugida (Arachnida). Florida Entomologist 49:199-216. Muma, M.H. 1967. Basic behavior of North American Solpugida. Florida Entomologist 50:115-123. Pablos, F. 1967. Sobre la presencia de Gluvia dorsalis, var. conquensis en los puertos de Contreras (Cuenca). Publicaciones del Instituto de Biologia Aplicada 43:45M8. Peretti, A.V. & R.H. Willemart. 2007. Sexual coercion does not exclude luring behavior in the climbing camel-spider Oltacola cliacoensis (Arachnida, Solifugae, Ammotrechidae). Journal of Ethology 25:29-39. Phillips, S.J., R.P. Anderson & R.E. Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190:231-259. Phillips, S.J. & M. Dudik. 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161-175. Polis, G.A. & S.J. McCormick. 1986. Scorpions, spiders and solpugids: predation and competition among distantly related taxa. Oecologia 71:111-116. Punzo, F. 1993. Diet and feeding behavior of the solpugid, Eremobates palpisetosus (Solpugida: Eremobatidae). Psyche 100:151-162. Punzo, F. 1994. An analysis of feeding and optimal foraging behaviour in the solpugid Eremobates mormonus (Roewer) (Solpugida, Eremobatidae). Bulletin of the British Arachnological Society 9:293-298. Punzo, F. 1995. Feeding and prey preparation in the solpugid, Eremorhax magnus Hancock (Solpugida: Eremobatidae). Pan- Pacific Entomologist 71:13-17. Punzo, F. 1997. Dispersion, temporal patterns of activity, and the phenology of feeding and mating behaviour in Eremobates palpisetulosus Fichter (Solifugae, Eremobatidae). Bulletin of the British Arachnological Society 10:303-307. Punzo, F. 1998a. Natural history and life cycle of the solifuge Eremobates marathoni Muma & Brookhart (Solifugae, Eremoba- tidae). Bulletin of the British Arachnological Society 11:111-118. Punzo, F. 1998b. The Biology of Camel-Spiders (Arachnida, Solifugae). Kluwer Academic Publishers, Boston. Punzo, F. 1998c. The effects of maternal nest guarding behaviour by Eremobates marathoni Muma & Brookhart on the survivorship of offspring (Solifugae, Eremobatidae). Bulletin of the British Arachnological Society 11:54—56. R Development Core Team. 2009. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. Online at http://www.R-project.org. Rambla, M. & J.A. Barrientos. 1986. Nuevos datos sobre Gluvia dorsalis (Latreille, 1817) (Solifugae, Daesiidae). Pp. 213-218. In Proceedings of the Ninth International Congress of Arachnology. (W.G. Eberhard, Y.D. Lubin & B.C. Robinson, eds.). Smithsonian Institution Press, Washington. Rose, M.R. & L.D. Mueller. 2006. Evolution and Ecology of the Organism. Pearson Prentice Hall, Upper Saddle River, New Jersey. Schenker, R. 1980. Erster Nachweis einer Solifugae-Art {Gluvia dorsalis Latreille 1817) (Arachnida) in Portugal. Mitteilungen der Schweizerischen Entomologischen Gesellschaft 53:283-284. Turner, C.H. 1916. Notes on the feeding behavior and oviposition of a captive American false-spider {Eremobates formicaria Koch). Journal of Animal Behavior 6:160-168. Wharton, R.A. 1987. Biology of the diurnal Metasolpuga picta (Kraepelin) (Solifugae, Solpugidae) compared with that of nocturnal species. Journal of Arachnology 14:363-383. Manuscript received 23 November 2009, revised 2 August 2010. 2010. The Journal of Arachnology 38:475^84 Whip spiders of the genus Sarax in the Papuan region, with description of two new species (Amblypygi: Charinidae) Cahyo Rahmadi: Natural History Laboratory, Faculty of Science, Ibaraki University, Mito, 310-8512 Japan and Museum Zoologicum Bogoriense, Research Center for Biology, Indonesian Institute of Sciences (LIPI) Jl. Raya Jakarta-Bogor Km. 46, Cibinong, 16911 Indonesia. E-mail: cahyo.rahmadi@lipi.go.id Jundchi Kojima: Natural History Laboratory, Faculty of Science, Ibaraki University, Mito, 310-8512 Japan Abstract. Three species of the genus Sarax are recognized in the Papuan region. Among them, two species, Sarax newbritainensis, new species, from New Britain, and S. monodent iculatus, new species, from Waigeo Island are described. Sarax newbritainensis resembles S. willeyi in having the same number of denticles on the pedipalpa! tarsus, but they distinctly differ from each other in body size, form of carapace, length of legs and number and arrangement of the trichobothria of tibia of leg IV. Sarax monodenticulatus is distinguished from the other two Papuan species by possessing a single denticle on the pedipalpal tarsus. The taxonomic status and the natural history of the species are discussed. Keywords: Taxonomy, Waigeo Island, Batanta Island, New Britain, Papua New Guinea, Indonesia Whip spiders, or Amblypygi, are flattened arachnids characterized by their raptorial pedipalps and the first legs modified into extremely elongate antenniform appendages (Weygoldt 2000). They are mostly active during the night and hide under stones and fallen trees during the day. A total of 143 species belonging to 17 genera in five families are currently known (Harvey 2002, 2003; Armas & Gadar 2004; Weygoldt 2002; 2006; Rahmadi & Harvey 2008). Of these, ten species and two subspecies belong to the charinid genus Sarax Simon 1892, which is known from the Oriental and Papuan region (India to the Solomons) (Harvey 2003; see also Weygoldt 2005:12). Kraepelin (1895) recorded Sarax sarawakensis (Thorell 1888) from New Guinea and Solomon Islands, and Pocock (1898) recorded the species from New Britain based on specimens procured by Dr. A. Willey, the director of the Colombo Museum at that time. Gravely (1915) described S. willeyi based on the two specimens “preserved in the Indian Museum” and “collected by Dr. Willey in New Britain.” The other charinid genus occurring in the Papuan region is Charinus Simon 1892. The genus occurs pantropically with 35 species and in the Papuan region only one species, Charinus papuanus Weygoldt 2006, is known, from the type locality near Port Moresby, Papua New Guinea (Weygoldt 2006). During a study on whip spiders collected during the “Ekspedisi Widya Nusantara (e-WIN Expedition)” to Raja Ampat Island, West Papua (organized by the Indonesian Institute of Sciences in 2007 and 2008), and based on the examination of some additional specimens from Papua New Guinea and New Britain and the holotype of Charon sarawakensis (= S. sarawakensis), we recognized three distinct species of Sarax in the Papuan region. Two of them are described here as new to science, and the taxonomic status of S. willeyi is discussed. METHODS Specimens from the e-WIN expedition examined in this study were preserved in 90-95% ethanol and others in 70- 80%. General terminology and pedipalpal spination follow Weygoldt (2000), and pedipalpal terminology follows Harvey and West (1998). We made the measurements (in mm) and drawings using, respectively, an ocular micrometer and a drawing tube mounted on a stereoscopic dissecting microscope (Olympus SZX12). To prepare the digital images of the carapace, we compiled multiple focal planes taken with a digital microscope (KEYENCE VH-5500) using the software Helicon Focus 4.70 (Helicon Soft Ltd. 2009: online at http:// www.heliconsoft.com/heliconfocus.html). In order to examine the structure of genitalia we lifted the genital operculum. The acronyms of the museums/institutions are as follows: MCSG, Museo Civico di Storia “Giacomo Doria,” Genova, Italy; MCZ, Museum of Comparative Zoology, Cambridge, Mas- sachusetts, USA; MZB, Museum Zoologicum Bogoriense, Cibinong, Indonesia; MNHN, Museum National d’Histoire Naturelle, Paris, France. The holotype female of Sarax sarawakensis (Thorell 1888), lodged in MCSG and labeled “Sarax saravakensis Thor. Sarawak, Viag Doria and Beccari” was examined to compare with the three Sarax species treated below. KEY TO THE CHARINID WHIP SPIDERS OF THE PAPUAN REGION 1. Abdominal sternite III without ventral sac covers Charinus papuanus Weygoldt Abdominal sternite III with ventral sac covers Sarax... 1 2. Body reddish-brown. Tibia of leg IV with 17 trichobothria (Fig. 4d); pedipalpal tarsus with one denticle (Fig. 4b) • • • • .Sarax monodenticulatus new species Body pale brown or dark greenish-brown. Tibia of leg IV with 17 or 19 trichobothria; pedipalpal tarsus with two denticles (Figs. 2c, 3b) 3 475 476 THE JOURNAL OF ARACHNOLOGY Figure 1. — Carapace, a, Sarax willeyi from Batanta Island; b, Sarcix mwhritainensis, holotype from New Britain; c, Sarax monodent iculatus, holotype from Waigeo Island. 3. Body pale brown. Legs elongate; tibia of leg IV with 19 trichobothria (Fig. 3e); trichobothrium bt close to distal margin; metatarsus of leg I 1.5 times as long as the first tarsal segment Sarax newbritainensis new species Body dark greenish-brown. Legs not elongate; tibia of leg IV with 17 trichobothria (Fig. 2d); trichobothrium bt locating at about mid-length of fourth basitibial segment; metatarsus of leg I as long as the first tarsal segment Sarax willeyi Gravely TAXONOMY Family Charinidae Quintero 1986 Genus Sarax Simon 1892 Sarax Simon 1892:43, 48; Kraepelin 1895:45; Kraepelin 1899:250; Pocock 1900:131; Gravely 1915:441; Mello-Leitao 1931:55; Werner 1935:471; Weygoldt 2000:25; ITarvey 2003:7. Phrynichosarax Gravely 1915:437; Mello-Leitao 1931:52 (as Phrynicosarax [sic]); Werner 1935:470; Weygoldt 2000:25 (synonymized with Sarax)', Weygoldt 2002:146. Type species. — Sarax: Sarax hrachydactylus Simon 1892, by original designation. Phrynichosarax: Phrynichosarax cochinensis Gravely 1915, by original designation. Diagnosis. — Small to medium-sized whip spiders; adult body length 4-10 mm. Pedipalpal patella with three large primary spines, the distal spine largest, subsequent spines becoming shorter proximally; ventral sac covers present on abdominal sternite III. Lateral eyes close to lateral margin of carapace. Basitibia of leg IV consisting of three or four segments. Remarks. — The family Charinidae currently consists of the following three genera: Sarax, Catageus Thorell 1889 with a single species from Myanmar, and Charinus consisting of 35 species with a circum-tropical distribution (Harvey 2003; Weygoldt & Van Damme 2004; Weygoldt 2006). Sarax is, in general, similar to Charinus, especially in the pedipalpal patella spination, but clearly distinguished by the presence/ absence of the ventral sac cover; in Sarax the ventral sac cover is present, while it is absent in Charinus. Catageus can be distinguished from the other two genera by the second spine of the pedipalpal patella being the largest (in other two genera, the first is the largest) and the proximal spine of the antero- dorsal pedipalpal tibia being larger than the distal one (Weygoldt 2000). Sarax willeyi Gravely 1915 (Figs, la, 2a-g) Sarax willeyi Gravely 1915:441, fig. 7; Mello-Leitao 1931:55; Werner 1935:471; Kraus 1970, figs. 10-11; Harvey 2003:9. Material examined. — INDONESIA: West Papua Province'. 3 females (MZB.Ambl.l 19-120 and 121 [ovigerous]), under stone in forest near Gua Eleg (00°53.51'S, 130°40.06'E, 156 m ask), Wailebet, Raja Ampat Regency, Batanta Island, 1 May 2008, C. Rahmadi; 4 males (MZB.Ambl.l 22, 125-127), 1 female (MZB.Ambl.l 23, ovigerous with 7 eggs), under stone in small limestone forest near Gua Umso (00°49.89'S, 130°53.82'E), Yenanas, Raja Ampat Regency, 27 April 2008, C. Rahmadi; 1 male (MZB.AMbl.128), Solol Village, Salawati Island, 7 May 2008, C. Rahmadi. PAPUA NEW GUINEA: Madang Prov- ince: 2 males, 3 females (MZB.AMbl. 129-133), 2 males, 2 females (MCZ DNA104752), Madang Bitabag Reserve, (05°08'19.3"S, 145°46'28.2"E, 109 m ask), 28 March 2006, R.M. Clouse, Ulai & Nataniek Diagnosis. — Sarax willeyi differs from other congeneric species from the region by its small size (adult body length about 4.0-6. 2 mm) and dark greenish-brown body. The legs are proportionally shorter than other species. The pedipalpal tibia has two spines on the antero-dorsal margin; the proximal spine more than half as long as the distal one. Pedipalpal tarsi with two denticles; proximal denticle about half as long as distal one. Tibia of leg IV with 17 trichobothria; trichoboth- rium be much closer to ft/than to sbf, and bt at about mid-length of the fourth basitibial segment (Fig. 2d). The metatarsal segment of leg I as long as the subsequent two segments together. Description. — Male: Color in alcohol. Carapace dark greenish-brown with yellow marks centrally. Pedipalps and legs green except as follows: major spines of pedipalps light brown; tibia and tarsus of leg I with yellow annulations; distal and proximal margins of basitibiae of Legs II-IV brown; distibiae and tarsi of Legs II-IV light green. Abdomen green with brown spots on each tergite. RAHMADI & KOJIMA— OF THE PAPUAN REGION 477 PI Figure 2. — Sarcix willeyi, from Batanta Island, a. External view of left chelicera; b. Sternal area of carapace, ventral view; c. Antero-dorsal view of left pedipalp; d. Arrangement of trichobothria on the fourth basitibial segment and distibia of leg IV; e, f. Male genitalia (e. ventral view; f dorsal view); g. Dorsal view of female gonopods. Carapace (Fig. la): Width about 1.5 times length; surface finely granulate, without setiferous tubercles; median sulcus deep in posterior one-fourth of the carapace, three sulci running laterally on each lateral half of carapace; flange wide and bend upward; anterior margin rounded, with six fine frontal setae. Median eye tubercle black, without apical setae, slightly emarginated antero-medially to form heart shape; median eyes facing antero-laterally. Lateral eye close to lateral margin of carapace. Chelicera (Fig. 2a): Dorsum smooth, with one fine frontal seta and three fine lateral setae. Basal segment with four teeth: lowermost tooth largest, uppermost tooth bicuspid, with upper cusp larger than lower cusp; inner surface with seven setae arranged in vertical row; outer surface with small blunt tooth opposite bicuspid tooth and four setae near proximal margin. Movable hand with three teeth about equal in size. Sternum (Fig. 2b): First sternite (= tritosternum) elongate, with paired apical, median and strong basal setae; second 478 THE JOURNAL OF ARACHNOLOGY and third sternites rounded and flattened, with paired apical setae. Pedipalp (Fig. 2c): Short and stout. Trochanter with four setiferous tubercles arranged in a row along antero-dorsal margin, one spine medially and four setiferous tubercles on antero-ventral margin; ventro-anterior apophysis equipped with ten setiferous tubercles present on distal margin of trochanter. Femur with four major spines, some setiferous tubercles and small tubercles in the antero-dorsal part, length of spine F1>F2>F3>F4; area without setiferous tubercles or small tubercles forming narrow band running lengthwise from proximal to distal margin; four major spines, several minor spines and small tubercles on antero-ventral margin; one spine present dorsally of FI and as long as 3/4 length of FI, length of spine FI>FII>FIII, three minor spines between FI and FII. Patella with four major spines, several minor spines, several setiferous tubercles and small tubercles on antero-dorsal margin; length of spine P1>P2>P3>P4, two minor spines (one in several specimens) between PI and distal margin of patella; three major spines, several setiferous tubercles and small tubercles on antero-ventral margin, length of spines PI>PII>PIII. Tibia with two major spines on antero-dorsal margin, length of proximal spine more than half that of distal one; one major spine on antero-ventral margin close to distal margin of tibia; outer surface of the tibia with several setiferous tubercles, finely granulate. Tarsus completely divided (claw clearly demarcated by articulation), with two denticles on antero-dorsal margin; proximal denticle about 3/4 as long as distal denticle; cleaning organ ventrally with about 30 modified hairs; apotele present. Legs (Fig. 2d): Femora of Legs I-IV with small tubercles bearing setae. Tibia and tarsus of leg I consisting of 23 and 41 segments, respectively; tibiae of Legs II and III two-segmented; basitibia of leg IV four-segmented, fourth segment with one trichobothrium (value in parentheses: ratio of the distance from the trichobothrium to the proximal margin of the segment against the length of the segment), bt (0.50); distitibiae of Legs II-IV each with 16 trichobothria (Fig. 2d), /?/'(0.07), sbf (0.31), he (0.17), bl at about mid-length of the segment, be close to bf than to sbf. Tarsi of Legs II-IV four-segmented; first segment about as long as length of subsequent three segments combined; second segment with light-yellow transverse line; fourth segment without oblique slit; pulvilli present. Genitalia (Figs. 2e, 0: Covered ventrally by genital opercu- lum; paired apically-pointed small median projections present at posterior margin; two brown marks present near the base of projections (Fig. 2e). In dorsal view, paired anteriorly- rounded brown bands present; paired weakly-sclerotized brown markings present medially (Fig. 2f). Female: Similar to the male. Gonopods with paired finger- like apically pointed projections (Fig. 2g). Measurements. — Male (n = 7), [female (n = 9)]; values for segments of the appendages are their lengths. Body length (excluding chelicera) 4.00-6.20 [4.04-6.95]. Carapace: median length 1.75-2.40 [1.48-2.70]; width 1.50-3.60 [2.08-3.75]; median eyes to anterior margin of carapace 0.04—0.05 [0.04— 0.05]; distance between lateral eyes 0.92-1.72 [0.80-1.75]; lateral eye to anterior margin of carapace 0.25-0.40 [0.20- 0.40]; lateral eye to lateral margin of carapace 0.08-0.25 [0.04-0.25]. Pedipalps: trochanter 0.40-0.80 [0.32-0.75]; femur 1.00-2.28 [0.70-1.90]; patella 1.00-2.60 [1.00-2.40]; tibia 0.40-1.25 [0.28-1.00]; tarsus 0.40-1.00 [0.60-1.25]. Leg I: femur 2.50-4.60 [4.60]; patella 0.35-0.50 [0.32-0.50]. Leg II: femur 1.75-3.20 [1.48-3.25], patella 0.48-0.70 [0.40-0.75]; basitibia 2.08 [0.80-2.25]; distitibia 1.60 [0.84—1.50]; metatar- sus + tarsus 1.60 [0.88-1.50]. Leg III: femur 2.25-3.60 [1.68- I 3.75]; patella 0.48-0.70 [0.40-0.75]; basitibia 1.15-2.80 [1.20- 3.00]; distitibia 1.28-1.80 [1.00-2.00]; metatarsus+ tarsus 1.05-1.60 [1.00-2.00], Leg IV: femur 1.85-3.20 [1.52-3.50]; I patella 0.35-0.70 [0.40-0.70]; basitibia 1.60-3.00 [1.20-3.15]; distitibia 1.00-1.6() [0.80-1.65]; metatarsus + tarsus 0.95-1.60 [0.80-1.75]. Remarks. — Sarax wUleyi was described by Gravely (1915) based on two specimens collected by Dr. A. Willey in New Britain, which were stated to be in the Indian Museum in Calcutta. Pocock (1898) examined the specimens collected by Dr. A. Willey in New Britain, compared them with Sarax specimens from Luzon Island and Andaman Islands, and having agreed with Kraepelin (1895), he concluded that only the single species, S. sarawakensis, was recognized in Sarax. Although Gravely (1915) did not refer to Pocock (1898), the specimens collected by Dr Willey in New Britain that they examined may have been the same. Gravely (1915) distinguished S. wiileyi from S. sarawakensis by the proximal spine of the pedipalpal tarsus being more than ! half as long as the distal one (less than half in S. sarawakensis). The specimens from Batanta Island, Salawati Island, and ' Madang in Papua New Guinea that we examined have the j proximal spine of the pedipalpal tibia longer than half the length of the distal spine. The holotype of Charon sarawa- , kensis (= Sarax sarawakensis), on the other hand, has the pedipalpal tibia with the proximal spine shorter than half the length of distal one. In addition, the holotype of C i sarawakensis has the pedipalpal tarsus armed with two very | small denticles, while all of our specimens identified as S. j wiileyi have the pedipalpal tarsus with two rather long I denticles. All the distribution records of S. sarawakensis in the Papuan regions including Solomon Islands and Bismarck I Archipelago, such as reported by Kraepelin (1899), need ; reconfirmation. At this moment, it is reasonably considered that S. sarawakensis is restricted, in its distribution, to the | Oriental region. Natural history. — We found this whip spider most often under stones, and when specimens were collected during the ! day, they were found in a resting position attached to the \ underside of a stone on forest floor. We also collected some individuals under fallen trees in Salawati Island. According to the collection data, the specimens from Madang, Papua New Guinea were collected under rotten logs. Distribution. — This species occurs in New Britain (Gravely 1915), and is here newly recorded from the West Papua Province of Indonesia, and in the Madang Province of Papua New Guinea (Fig. 5). Sarax newbritainemis new species (Figs, lb, 3a-h) Type material. — Holotype male, Papua New Guinea, East New Britain Province, New Britain Island, Resurgence Lali Bairaman (GPS coord, aprox. 05°39'32.59"S, 15ri2'39.91"E), 17 February 2005, R. Sougeat (Expe Papou 2005) RAHMADI & KOJIMA— OF THE PAPUAN REGION 479 Figure 3. — Sarax newhritainensis, new species, from New Britain, Papua New Guinea, a. External view of left chelicerae; b. Antero-dorsal view of left pedipalp; c. Antero-ventral of left pedipalpal trochanter; d. Left pedipalpal tibia; e. Arrangement of trichobothria on fourth basitibial segment and distibia of left leg IV; f, g. Male genitalia (f. dorsal view; g. ventral view); h. Dorsal view of female gonopods. (MNHN.Am.6). Paratype: 1 female, same locality data as the holotype (MZB.Ambl.134). Etymology. — The specific name refers to the island where the type locality is located. Diagnosis. — Sarax newbritainensis differs from all other congeneric species in the region by the large adult body length (about 8.25-9.5 mm) and the pale brown body. The carapace is proportionally less wide, without distinct lateral sulci, and the eyes are reduced in size. The pedipalpal tarsus has two rather long denticles, separated from each other by about twice the basal diameter of the denticle; the proximal denticle is about half as long as the distal one. The legs are elongate; metatarsi of leg II-IV are longer than the length of the subsequent three tarsal segments combined. Tibia of leg IV with 19 trichobothria, arranged with be situating near the middle of bf and sbf and bt close to the distal margin of the fourth basitibial segment. Description. — Male: Color in alcohol: Carapace light brown with darker marks; pedipalps brown but pedipalpal tarsus light brown. Legs I-IV yellowish brown, without annulations except for tibia and tarsus of leg I having white annulations; patella dark brown. Carapace (Fig. lb): Width about 1.3 times the length; surface finely granulate, sparsely with small tubercles, without setiferous tubercles; sulcus deep and distinct on posterior one- fourth of carapace. Flange present from level of lateral eyes to posterior margin, wide and bent upward along lateral margin, narrow on posterior margin. Anterior margin of carapace 480 rounded, with six frontal setae and four small fine setae close to each antero-lateral corner; slightly concave in part anterior to lateral eyes. Median eye tubercle black, without apical seta, slightly reduced in size, slightly emarginated antero-medially to form heart shape; eyes facing antero-laterally. Lateral eyes close to lateral margin of carapace, distance between them about the diameter of the lateral eye, normal pigmentation and tapetum. Frontal process triangular, visible from above. Chelicera (Fig. 3a): Dorsum smooth, with three setiferous tubercles and two fine frontal setae. Basal segment with four teeth: the lowermost tooth largest, the uppermost tooth bicuspid, with upper cusp larger than lower cusp; inner surface with 14 setae arranged in vertical row near proximal margin; outer surface with small blunt tooth opposite the bicuspid tooth, and with five setae arranged in vertical row. Movable hand with five teeth; second tooth largest. Sternum: First sternite (= tritosternum) elongate, with paired apical and strong median setae, five setae between apical and median setae and 12 small setae at base. Second and third sternites rounded, with paired apical setae; second with six basal setae; third with five basal setae. Pedipalp (Figs. 3b-d): Short and stout, with several setiferous tubercles. Trochanter with a row of seven setiferous tubercles on antero-dorsal margin, six setiferous tubercles dorsally and one median spine on antero-ventral margin, medially with one spine and eight setiferous tubercles; ventral anterior apophysis equipped with several setiferous tubercles in basal part present on distal margin of trochanter (Fig. 3c). Femur with four major spines, several setiferous tubercles and small tubercles on antero-dorsal part, length of spine F1>F2>F3>F4; four major spines, several setiferous tubercles and small tubercles on antero- ventral margin, one spine present dorsally of FI and as long as 3/4 length of FI, length of spine FI>FII>FIII>FIV, one minor spine between FI and FII. Patella with four major spines, several setiferous tubercles and small tubercles on antero-dorsal margin, length of spine P1>P2>P3>P4; one setiferous tubercle and one spine between PI and distal margin of patella, the spine as long as half length of PI; three major and one minor spines, several setiferous tubercles and small tubercles on antero-ventral margin, length of spines PI>PII>PIII. Tibia with two major spines on antero- dorsal margin, length of proximal spine more than half the length of distal one (Fig. 3d); antero-ventral margin with one major spine; outer surface finely granulate, with several setiferous tubercles. Tarsus completely divided (claw clearly demarcated by articulation), with two denticles on antero- dorsal margin: proximal denticle slightly longer than half the length of distal one, distance between them about three times basal diameter of proximal denticle; cleaning organ ventrally with 28-30 modified hairs; several blunt setae on inner surface of tarsus; apotele present. Legs (Fig. 3e): Femora of Legs II V with small setiferous tubercles. Tibia and tarsus of leg I with 23 and 41 segments, respectively; tibiae of leg II and III two-segmented; basitibia of leg IV four-segmented: fourth segment with one trichobothrium (value in parentheses as for S. willeyi), ht (0.57); distitibiae of Legs II-IV with 18 trichobothria, />/ (0,13), 56/(0.32), be (0.23), bt close to distal margin, be at the middle of 6/ and sbf, (Fig. 3e). Tarsi of Legs II-IV four-segmented; first segment slightly THE JOURNAL OF ARACHNOLOGY longer than length of subsequent three segments combined; second segment with light yellow transverse line; fourth segment without oblique slit; pulvilii present. Genitalia (Figs. 3f, g): Covered ventrally by genital oper- culum, of which posterior margin is equipped with paired setae; posteriorly with paired ventral and dorsal lobes, the dorsal lobe smaller than ventral one (Fig. 3f). In dorsal view, submedian brown bands running from anterior margin to the middle; inner margin of median lobe with brown region weakly sclerotized; brown spot present on each median lobe. In ventral view, base of ventral lobe with a narrow brown band (Fig. 3g). Female: Similar to the male but differing as follows: carapace slightly darker; chelicera with three frontal fine setae; meta-sternum slightly more elongate, with two apical setae; tibia and tarsus of leg I respectively with 26 and 42 segments. Gonopods soft, cone-shaped, with several setae on margin of genital operculum (Fig. 3h). Measurements. — Male (holotype, MNHN.Am.6) [female (paratype, MZB.Ambl.134)]; values for segments of the appendages are their lengths: Body length (excluding chelicera) 9.50 [8.25]. Carapace: median length 3.75 [3.50], width 5.00 [5.00]; median eye to anterior margin 0.05 [0.05], distance between lateral eyes 2.60 [2.25], lateral eye to anterior margin 0.60 [0.65], lateral eye lateral margin 0.20 [0.25], Pedipalps: trochanter 1.00 [1.15], femur 3.20 [2.50], patella 3.25 [3.00], tibia 1.50 [1.25], tarsus 1.60 [1.50]. Leg I: femur 10.00 [10.15], patella 0.75 [0.75]. Leg 11: femur 6.00 [5.75], patella 1.00 [1.00], basitibia 5.00 [4.90], distitibia 3.00 [2.90], metatarsus+tarsus 2.55 [2.60]. Leg III: femur 7.25 [6.50], patella 1.10 [1.25], basitibia 6.35 [6.25], distitibia 3.50 [3.25], metatarsus-htarsus 2.95 [3.00], Leg IV: femur 6.50 [5.75], patella 1.00 [1.00], basitibia 7.25 [6.75], distitibia 2.95 [2.75], metatarsus-htarsus 3.00 [3.00]. Remarks. — Sarax newbritainensis is distinguished from the other Sarax species known from New Britain, S. willeyi, by the generally larger body, the distinctly paler coloration, the carapace proportionally less wide (about 1.3 times as wide as long, while about 1.5 times in S. willeyi) and without distinct lateral sulci, the eyes reduced in size, the strongly elongated legs, and the number and arrangement of the trichobothria as given in the key. Natural history. — We collected the specimens of S. new- britainensis from the caves called Resurgence of Lali Bairaman in New Britain. The species has characteristics typical of cave dwellers such as small eyes, elongate legs, and pale body color. Distribution. — This species is known only from the type locality in New Britain (Papua New Guinea) (Fig. 5). Sarax momdenticulatm new species (Figs. Ic, 4a-h) Type material. — Holotype male, Indonesia, West Papua Province, Waigeo Island, Mumes (00°21.23'S, 130°58.93'E), under stone in limestone forest. Raja Ampat Regency, 10 June 2007, C. Rahmadi, E-Win 2007 LIPI (MZB.Ambl.135). Paratypes: 1 female (MZB.Ambl. 136), 1 male (MNHN.Am.7), 2 juveniles (MZB.Ambl. 140), same data as holotype; 4 males (MZB.Ambl. 137, 138, 141, 143), 3 females (MZB.Ambl. 139, 142,144), Air Dingin Monfaya (Resurgence), Wairabiae Lopintol {00°18.13'S, 130°56.09'E), under stone in limestone forest. Raja Ampat Regency, 4 June 2007, C. Rahmadi. 'i If i \ I ^ u I 4 f \ i i i I 1 RAHMADI & KOJIMA— SAi?AA' OF THE PAPUAN REGION 481 Figure 4. — Sarax monodenticulatus, new species, from Waigeo Island, Indonesia, a. External view of left chelicerae; b. Antero-dorsal view of left pedipalp; c. Antero-ventral view of left pedipalpal trochanter; d. Arrangement of trichobothria on fourth basitibial segment and distibia of left leg IV; e. Apical four tibial and basal five tarsal segments of antenniform leg I; f, g. Male genitalia (f. dorsal view; g. ventral view); h. Dorsal view of female gonopods. Etymology. — The specific name refers to the presence of one denticle on the pedipalpal tarsus. Diagnosis. — Sarax monodenticulatus differs from other congeneric species in the region by being small- to medium- sized (adult body length 2. 5-6. 5 mm) and with a reddish- brown carapace. The pedipalpal tarsus has a single denticle. Metatarsus of leg I as long as the subsequent two segments together; short- and long-segment alternating combination present in second to eleventh tarsal segments. Description. — Male: Color in alcohol: Carapace dark red- dish-brown with darker marks; pedipalp dark brown but pedipalpal tarsus light brown. Legs I-IV brown, without annulations, but tibia and tarsus of leg I yellowish brown, with white annulations. Basitibiae of Legs II-IV brown; their distibiae and tarsi greenish-brown. Abdomen dark brown; each tergite with yellow marginal line and light-brown spots. Carapace (Fig. Ic): Width about 1.5 times its length; surface finely granulate; frontal area with dense, small tubercles, without setiferous tubercles; median sulcus present in poste- rior one-fourth of the carapace. Flange present in area posterior to lateral eyes, wide but narrow in posterior margin, bent upward. Anterior margin of carapace rounded, with six 482 I THE JOURNAL OF ARACHNOLOGY frontal setae and some fine setae close to antero-lateral corner. Median eye tubercle black, without apical setae, slightly emarginated antero-medially to form heart shape; eyes facing antero-laterally. Lateral eye large, with normal pigmentation and tapetum, separated from lateral margin of carapace by about its diameter. Frontal process triangular, visible from above. Chelicera (Fig. 4a): Dorsum smooth, with two fine frontal setae and several setae on dorsal and lateral parts of outer margin. Basal segment with four teeth: lowermost tooth largest, uppermost tooth bicuspid, with upper cusp larger than lower one; inner surface with eight setae arranged in a vertical row close to proximal margin; outer surface with a small blunt tooth and five setae near proximal margin. Movable hand with two teeth close to the proximal margin. Sternum: First sternite (= tritosternum) elongate, with paired apical and median setae, paired small setae between apical and median setae, and four small setae basally. Second and third sternites rounded, each with paired apical setae. Pedipalp (Figs. 4b, c): Strong and slender, with several setiferous tubercles and small tubercles. Antero-dorsal margin of trochanter with four setiferous tubercles dorsally, two spines and seven setiferous tubercles ventrally; antero-ventral margin with ventral anterior apophysis basally equipped with several setiferous tubercles; ventral anterior apophysis with one spine medially and one spine dorsally (Fig. 4c). Femur with four major spines, several setiferous tubercles and small scales on the antero-dorsal margin, length of spine F1>F2>F3>F4; five major spines, two minor spines, four setiferous tubercles and small scales present on antero-ventral margin, one spine present dorsally of FI and about as long as 3/4 length of FI, length of spine FI>FII>FIII>FIV, single minor spine present between FII and Fill and between Fill and FIV. Patella with four major spines, several setiferous tubercles and small scales on antero-dorsal margin, length of spine P1>P2>P3>P4, one minor spine present between distal margin and PI and as long as half of the length of PI, two spines and ons setiferous tubercle present between P4 and the proximal margin; three major spines and one minor spine, several setiferous tubercles and small scales on antero-ventral margin, length of spine PI>P1I>PIII, one minor spine between PHI and the proximal margin. Tibia with two major spines on antero-dorsal margin: proximal spine slightly longer than half of distal one; one major spine on antero-ventral margin close to distal margin; outer surface finely granular, with setiferous tubercles arranged in three rows. Tarsus completely divided (claw clearly demarcated by articulation), with one denticle on antero-dorsal margin; blunt setae present on inner surface; cleaning organ ventrally with 27-28 modified hairs; apotele present. Legs (Figs. 4d, e): Femora of Legs I-IV with small scales and setae forming a longitudinal row. Tibia and tarsus of leg I with 23 and 41 segments, respectively (Fig. 4e); tibiae of legs II-III two-segmented; basitibia of leg IV four-segmented (in some specimens left basitibia three-segmented): fourth (third in specimen with three basitibia! segments) segment with one trichobothrium (value in parentheses as for S. willeyi) bt (0.61); distitibiae of legs II-IV each with 16 trichobothria (Fig. 4d), 6/ (0. 10), 56/ (0.30), he (0.17), be about the middle of 6/' and shf, bt close to distal margin. Tarsi of legs II-IV four- segmented; first segment about as long as length of the > subsequent three segments combined; second segment with i light yellow transverse line; fourth segment without oblique I slit; pulvilli present. | Genitalia (Figs. 4f, g): Covered ventrally by genital oper- ^ culum equipped with several setae; distally with paired small I submedian lobes. In dorsal view, with paired large median ^ lobes, of which distal margins are brown; four longitudinal sclerotized bands present posterior to median lobes (Fig. 4f). In ventral view, brown band present near base of distal lobe (Fig. 4g) Female: Similar to the male. Gonopods soft and cone- shaped (Fig. 4h) I Measurements. — Male (n = 5) [female (« = 4)]; values for ! segments of the appendages are their lengths; Body length i! (excluding chelicera) 4.88-6.48 [2.48-5.60]. Carapace: median | length 1.88-2.60 [2.00-2.48], width 2.60-3.72 [3.00-3.60]; , median eyes to anterior margin 0.04-0.08 [0.04-0.08], distance | between lateral eyes 1.20-1.80 [1.40-1.72], lateral eye to ii anterior margin 0.28-0.44 [0.36-0.40], lateral eye to lateral ;; margin 0.12-0.16 [0.08-0.16]. Pedipalps: trochanter 0.40-1.00 j [0.60-0.80], femur 1.20-3.08 [1.40-2.20], patella 1.40-3.52 j [1.60-2.64], tibia 0.60-1.20 [0.68-1.00], tarsus 0.08-1.12 j [0.60-1.20]. Leg I: femur 3.08-5.40 [3.40-4.80], patella 0.52- Ij 0.60 [0.40-0.48]. Leg II: femur 2.00-3.40 [2.40-3.20], patella 0.48-0.64 [0.48-0.72], basitibia 1.3^2.40 [i. 60-2.40], distitibia ! 1.12-1.60 [1.20-1.60], metatarsus+tarsus 1.00-1.60 [1.20-1.36], j Leg III: femur 2.40-4.00 [2.80-3.60], patella 0.48-0.72 [0.56- ! 0.60], basitibia 1.88-3.32 [1.88-3.00], distitibia 1.20-1.80 [1.36-1.68], metatarsus+tarsus 1.20-1.60 [1.20-1.44]. Leg IV: femur 1.20-3.48 [2.40-3.28], patella 0.60-2.80 [0.44-0.60], basitibia 0.48-3.20 [2.16-3.04], distitibia 1.40-2.80 [1.40- 1.60], metatarsus+tarsus 1.04—1.80 [1.20-1.52], ' Remarks. — This species is the only Papuan Sarax with a j J single denticle on the pedipalpal tarsus. The other Sarax j | species that have the single denticle on the pedipalpal tarsus j are S. javensis (Gravely 1915) distributed in West Java and S. [i 1 cochinensis (Gravely 1915) known from the Western Ghats in I Cochin, India. Gravely (1915) distinguished the two species by j the length of the denticle; the denticle of S. cochinesis is long || ^ and distinct, while in S. javensis it is minute (see Gravely j; ^ 1915:figs. 2, 3). Sarax monodenticulatus is distinguished from | these two species by having four-segmented basitibia (single- j or two-segmented in S. cochinensis and three-segmented in S. javensis [see Gravely 1915:437]). Natural history. — We collected specimens of this species | mostly singly under stones in limestone forests. During the 1 exploration of cave fauna in Waigeo Island, it was never found ?; within the caves. Distribution. — Sarax monodenticulatus is known only from r Waigeo Island (Indonesia) (Fig. 5). DISCUSSION The first record of whip spiders of the genus Sarax from \\ New Guinea was by Kraepelin (1895), who reported specimens without giving any precise data under the name S. t sarawakensis, which was