Journal of ARACHNOLOGY OFFICIAL ORGAN OF THE AMERICAN ARACHNOLOGICAL SOCIETY 16th International Congress of Arachnology 2 e ) n VOLUME 33 2005 NUMBER 2 THE JOURNAL OF ARACHNOLOGY EDITOR-IN-CHIEF: Daniel J. Mott, Texas A&M International University MANAGING EDITOR : Paula Cushing, Denver Museum of Nature & Science SUBJECT EDITORS: Ecology — Soren Toft, University of Aarhus; Systematics — Mark Harvey, Western Australian Museum; Behavior and Physiology — Gail Stratton, Univer- sity of Mississippi EDITORIAL BOARD: Alan Cady, Miami University (Ohio); James Carrel, University of Missouri; Jonathan Coddington, Smithsonian Institution; William Eberhard, Univer- sidad 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, Mi- ami University (Ohio); Paul Selden, University of Manchester (UK.); Matthias Schaefer, Universitet Goettingen (Germany); William Shear, Hampden- Sydney College; Petra Si- erwald. Field Museum; 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. Sub- scriptions to The Journal of Arachnology 2ind American Arachnology (the newsletter), and annual meeting notices, are included with membership in the Society. Regular, $40; Stu- dents, $25; Institutional, $125 . Inquiries should be directed to the Membership Secretary (see below). Back Issues: Patricia Miller, P.O. Box 5354, Northwest Mississippi Commu- nity College, Senatobia, Mississippi 38668 USA. Telephone: (601) 562-3382. Undelivered Issues: Allen Press, Inc., 1041 New Hampshire Street, P.O. Box 368, Lawrence, Kansas 66044 USA. THE AMERICAN ARACHNOLOGICAL SOCIETY PRESIDENT: Elizabeth Jacob (2006-2008), Department of Psychology, University of Massachusetts, Amherst, MA 01003 USA. PRESIDENT-ELECT: Paula Cushing (2006-2008), Denver Museum of Nature & Sci- ence, Denver, CO 80205 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: James Carrel (2001-2003), Rosemary G. Gillespie (2002-2004), Brent D. Opell (2003-2005). PAST DIRECTOR AND PARLIAMENTARIAN: H. Don Cameron (appointed), Ann Arbor, Michigan 48105 USA. HONORARY MEMBERS: C.D. Dondale, H. W. Levi, A.R Millidge, W. Whitcomb. Cover design by Jan Bosselaers. Publication date; 13 December 2005 ©This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 2005, The Journal of Arachnology 33:197-204 HORIZONTAL AND VERTICAL DISTRIBUTION OF SPIDERS (ARANEAE) IN SUNFLOWERS Stano Pekar: Research Institute of Crop Production, Drnovska 507, 161 06 Praha 6 — Ruzyne, Czech Republic. E-mail: pekar@vurv.cz ABSTRACT. Sunflowers are an increasingly important crop plant in the Czech Republic. The spider fauna of this crop has not been investigated yet. The aim of this study was to monitor the spider fauna of sunflowers and to study the seasonal change in the spatial and vertical distribution of this fauna. For this purpose a small experimental area was used where spiders on each single leave of 50 sunflower plants were visually checked at monthly intervals from spring until autumn. The density of spiders increased during the season reaching a maximum of seven spiders/plant in the autumn shortly before harvest. The spatial distribution changed accordingly, being random in spring and early summer and normal or aggre- gated toward late summer. Two spider species, Neottiura bimaculata and Theridion impressum (Theridi- idae), dominated (96% of all individuals) throughout the season. These two species exhibited a different microhabitat preference: N. bimaculata individuals were found particularly on the lower sunflower leaves, T. impressum preferred higher leaves. The density of the spiders (per leaf) was independent of the density of two dominant pest species, aphids and leafhoppers. Keywords: Aphids, spatial distribution, agrobiocenosis, stratification, colonization Although sunflowers are considered the second most important oilseed crop in the world (Cobia & Zimmer 1978), in the Czech Republic their importance was not recognized until recently when the current production had not been able to cover the need of our food industry (Jiratko et al. 1996). Since then the planted area has enlarged mainly in the south- eastern part of the country where the warmer climate provides suitable conditions for a high production. In its native region, i.e. North America, the sunflower has many pests (Charlet & Brewer 1998). Thus it thrives better in foreign coun- tries because it has left a multitude of pests and diseases behind. This is particularly true for Europe. In the Czech Republic the sun- flower plants are attacked by only a few pests: aphids, leafhoppers, moths and heteropterans (Jiratko et al. 1996). The fauna of natural enemies of sunflower pests has been so far investigated only outside Europe. It was found to be composed of var- ious heteropterans, lacewings, coccinellids, ants and parasitoids (Lynch & Garner 1980; Boica Junior et al. 1984; Men & Thakre 1998). Spiders were also among the most abundant and important predators (Seiler et al. 1987; Royer & Walgenbach 1991). For ex- ample, an araneid species, Neoscona nautica (L. Koch 1875), was found to prey on aphids and other pests on sunflower (Singla 1999). As the fauna of predators occurring on sun- flowers has not been investigated in Europe, the first aim of this study was to monitor spi- ders, particularly the change in their temporal and spatial distribution in a sunflower plot. Another aim was to observe the vertical dis- tribution of the most abundant species of spi- ders on sunflowers. Very little attention has been paid to the stratification of spider fauna in agroecosystems (exceptions are He et al. 1995; Hao et al. 2000), obviously due to the intensive effort required for such investigation (Holland et al. 2004). METHODS The study was performed in Praha-Ruzyne, the Czech Republic (50°06'N, 14°15T, fau- nistic grid no. 5951). The sunflowers were planted in April 2003 in rows 80 cm apart, at a distance of 30 cm from one seedling to an- other. The total area was about 2,000 m^. In the middle of this area, an experimental plot (4 m X 7 m) including 50 plants was selected. The position (coordinates relative to the left lower corner of the plot) of each plant within the experimental plot was mapped. The investigation began in late May when sunflower plants were 10 cm tall and termi- 197 198 THE JOURNAL OF ARACHNOLOGY Table 1. — List of spiders recorded species on the sunflowers during one season. Numbers are total records and the percentage from the total number. Family/species Number % Araneidae Aculepeira ceropegia (Walckenaer 1802) 11 0.70 Araneus sp. 1 0.06 Araniella sp. 4 0.20 Mangora acalypha (Walckenaer 1802) 1 0.06 Theridiidae Enoplognatha sp. 19 1.20 Neottiura bimaculata (Linnaeus 1767) 282 17.10 Theridion impressum L. Koch 1881 1301 79.20 Theridion varians Hahn 1833 2 0.11 Linyphiidae Microlinyphia pusilla (Sundevall 1830) 15 0.90 Thomisidae Xysticus sp. 6 0.40 Dictynidae Dictyna sp. 1 0.06 Total 1643 nated in September shortly before harvest. Plants in the selected plot were examined at monthly intervals, i.e. altogether five times during the season. On each examination date every single leaf (upper and lower surface) of each of 50 plants was visually inspected to record the number of spiders present. The leaves were gently inspected not to disturb present spiders. The height of each plant was recorded on each date too. The spiders were not sampled, only visually inspected in order to record their change during the seasons. Fur- ther, on each date 25 plants were selected out- side the experimental plot. On each plant one leaf was sampled in order to examine the number of spiders, aphids (unidentified), leaf- hoppers (unidentified) and other insects. All spiders were identified to species, if possible, or to a genus. Juvenile theridiid spiders were identified using Pekar (1999). Statistical analyses were performed using STATISTICA (StatSoft). Distribution of spi- ders at each observation date was tested using Kolmogorov-Smirnov test (KST) for normal- ity. Linear regression models (LM) were used to study the relationship between density and the season and the relationship between prey and spider densities. Since the data did not follow a normal distribution, log-transforma- tion was used prior to analysis. Horizontal dis- tribution of spiders in the study plot was stud- ied using graphical spatial analysis. The analysis projects a three-dimensional dataset that includes two-dimensional coordinates of each plant and the spider density for each plant on a two-dimensional plane. The gradi- ent of density is displayed as shades of gray with white color standing from 0 and black standing for the maximum density. The con- tours of density were modelled using distance weighted least-square method. Numbers rep- resent means ± standard error throughout the text. RESULTS Horizontal distribution, — More than 1600 individual spiders were observed during the study (Table 1). The majority of spiders (99%) were represented by theridiids. Two spider species, T. impressum L. Koch 1881 and Neottiura bimaculata (Linnaeus 1767) (both Theridiidae), accounted for 98% of all spiders, with the former species making up 79% of the spider fauna. The density of spiders on sun- flowers increased over the study season. On 28 May 2003, there were 0.27 ± 0.09 spiders per plant. On 25 June, the density increased to 1.49 ± 0.26 and on 23 July it was 1.12 ± 0.20 individuals/plant. On 20 August, the den- sity further increased to 6.78 ± 0.68 and on PEKAR— HORIZONTAL AND VERTICAL DISTRIBUTION OF SPIDERS 199 Distance [m]x 10"* Figure 1. — Seasonal change in the spatial distribution of spiders on the sunflower plot. The graph represents the study plot. The shades of gray identify spider density (per plant) on individual plants: the darker the shade the higher number of individuals. Contours of densities were modelled using leasUsquare method. Data from 50 plants in one 4 m X 7 m plot, inspected repeatedly. 17 September it was 6.98 ± 0.76 spiders per plant. The overall density thus increased foL lowing a linear model y = -“0.48 + 0.07* X (LM, = 0.91, P < 0.04). The average in- crement was thus 3.5 spiders/plot/day. Although the mean spider density of the en- tire plot increased during the season, detailed analysis of each individual plant revealed that there was a change of density within the plot. The highest increase (5.7 times on average) in the density was recorded from July- August, i.e. in the period following breeding when spi- der density increased on 88% of the plants. A less pronounced increase was found from 200 THE JOURNAL OF ARACHNOLOGY m e o Q 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 5 10 15 20 25 30 35 Leaf no* Figure 2. — Seasonal changes in the vertical distribution of spider density (mean ± SE) on sunflower plants. Leaves are numbered from the bottom to the top of the plants and grouped into height categories of 5 leaves. 17 September 1 1 1 1 May~June (1.2 times) when the spider density increased on 49% of the plants. From June- July the average density decreased (0.37 times) on 41% of the plants. Finally, from Au- gust~September the average density also de- creased (0.2 times) on 53% of the plants. The distribution of spiders changed during the season as follows (Fig. 1): in May, June and July the distribution was rather random (KST, P < 0.01). In August and September it approached a normal distribution (KST, P > 0.10). But the analysis of the spatial distri- bution showed that the distribution was in fact aggregated toward the end of season with two patches of high spider density on the margin of the study plot (Fig. 1), PEKAR— HORIZONTAL AND VERTICAL DISTRIBUTION OF SPIDERS 201 Figure 3. — Mean (± SE) density of N. bimaculata and T. impressum on sunflower leaves (numbered from the bottom to the top of the plants and grouped into height categories of 5 leaves). Data from July and August pooled. Vertical distribution. — The stratification of spiders did not change dramatically during the season. Temporal analysis showed that the spiders were always more abundant on the up- per leaves, except for the terminals, which formed the flower (Fig. 2). The distribution of the two most abundant species, N. bimaculata and r. impressum, differed. While N. bima- culata was mainly found in the lower parts of the plants, T. impressum dominated the upper parts (Fig. 3). The vertical distribution of aphids on sun- flower leaves is shown in Fig. 4. Unlike spi- ders, aphids were more abundant on lower than on upper leaves. The density of spiders (per leaf) was independent of the density of aphids and/or leafhoppers (LM, P > 0.23). DISCUSSION Observed composition of spiders on sun- flowers was similar to the canopy fauna of corn, soybean or rape in Europe (e.g., Alder- weireldt 1989; Nyffeler 1982), i.e., in all these studies it was dominated by theridiid spiders. Some differences were observed in compari- son with other crops, which presumably result from the different plant structure. Large sun- flower leaves do not provide suitable attach- ments for the webs of araneid spiders, which are therefore more abundant on structurally more complex plants, such as soybean or rape. In North America, the sunflower was domi- nated by other spider guilds: thomisid and sal- ticid spiders (Seiler et al. 1987). This is be- cause the spider fauna of agroecosystems in North America is different from that in Eu- rope (Nyffeler & Sunderland 2003). Increase of spider density with the devel- opment of crops was observed in many sea- sonal crops including sunflowers (Royer & Walgenbach 1991; Duffield & Reddy 1997). In this study it was caused by the influx of spiders, mainly theridiids, from neighboring habitats, which is taking place mainly in the spring (Blandenier & Eiirst 1997). The new 202 THE JOURNAL OF ARACHNOLOGY 10 15 20 Leaf no* 25 30 35 Figure 4. — Mean (± SE) density of spiders and aphids on sunflower leaves (numbered from the bottom to the top of the plants and grouped into height categories of 5 leaves). Data from July and August pooled. individuals settled randomly in the plot, in- dicated by their random distribution. In June the spiders reached maturity and mated, fob lowed by a decline in spider density resulting from the death of males. In July they repro- duced (Pekar 1999) and the newborn spider- lings dispersed locally. As a result the distri- bution became aggregated. Such distribution has been rarely documented for spiders in agroecosystems (e.g., Yan 1988; Nyffeler & Breene 1992). Linyphiid spiders showed no evidence of spatial pattern (Thomas et ah 1990; Holland et al. 2004) probably because of their high dispersal ability. A change from random to aggregated distribution as observed in this study was found also by Gang et al. (1989). They recorded that a linyphiid spider Hylyphantes graminicola (Sundevall 1830) had a random distribution at low population densities but aggregated at higher population densities in cotton. Spider densities vary not only temporally but also between different crops. In general, the density is expected to be a function of plant size and complexity, thus smaller plants host fewer spiders than tall ones. In accor- dance with this, Liu et al. (2003) observed a maximum density of four spiders per cotton plant, whereas Zhang et al. (1997) found a maximum of six spiders per corn plant. The two principal species of theridiid spi- ders seem to utilize different strata for their webs. Such dichotomous but syntopic web placement may be a result of competition or site preference. In an experimental work on the competition between two sympatric liny- phiid spiders, Herberstein (1998) found that two species, Frontinellma frutetorum (C.L. Koch 1834) and Neriene radiata (Walckenaer 1842), compete for web space. As a result, they placed their webs in different strata when occurring syntopically. No web displacement was observed for the theridiid spiders. Neot= tiura bimaculata was never found in the upper strata even on plants where T. impressum was absent. Thus it is possible that these two spe- PEKAR— HORIZONTAL AND VERTICAL DISTRIBUTION OF SPIDERS 203 cies have different microhabitat requirements and do not compete mutually for space. Sim- ilar preference for a certain stratum has also been observed in other spider species (e.g. Kim et al. 1989). In total the number of spiders was higher in the upper than in the lower stratum in this study. In contrast to this, Liu et al. (2003) found that there were more spiders in the low- er parts of cotton plants than in the upper parts. Similar results were obtained from a study on rice (Anwaru & Ibrahim 1995). But higher spider densities in the lower strata ob- served in these studies are due to the inclusion of epigeic spiders. Sunflower plants are not used as foraging sites by epigeic spiders as the leaves are high above the ground so it might be difficult for epigeic spiders to climb sun- flower plants. The number of aphids in this study was higher in the lower strata. Similarly, aphids on chili plants were more abundant in the lower than in higher strata (Idris & Mohamad 2002). Rarely has the distribution of predators in ar- able land been observed to be spatially de- pendent on their prey (Wang & Yan 1989). In a study of Holland et al. (2004) linyphiids showed no spatial association with aphids, though being their frequent prey. Pekar (2000) analyzed the diet of T. impressum on sunflow- er. He found it was composed mainly of aphids. Therefore it was expected that the spi- ders would be more abundant in the lower leaves where aphids were more abundant. But it was not, presumably due to counter effect of other factors, either biotic or abiotic. ACKNOWLEDGMENTS The study was supported by the project no. QD 1350 and no. 0002700601 of the Ministry of Agriculture of the Czech Republic. LITERATURE CITED Alderweireldt, M. 1989. Composition and density fluctuations of the invertebrate fauna occurring in a maize field at Melle (Belgium). Acta Phy- topathologica Entomologica Hungarica 24:5-12. Anwaru, I.M. & A.M. Ibrahim. 1995. Vertical and temporal distribution of some arthropod preda- tors in a rice ecosystem. Bangladesh Journal of Zoology 23:119-122. Blandenier, G. & P.A. Fiirst. 1997. Ballooning spi- ders caught by a suction trap in an agricultural landscape in Switzerland. Pp. 177 — 186. In Pro- ceedings of the 17th European Colloquium of Arachnology. (P.A. Selden, ed.). British Arach- nological Society, Burnham Beeches, Bucks. Boica Junior, A.L., A.C. Bolonhezi & N.J. Paccini. 1984. Survey of insect pests and their natural enemies on sunflower (Helianthus annuus L.), grown as first and second crops in the Munici- pality of Selviria, Mato Grosso do Sul. Anais da Sociedade Entomologica do Brasil 13:189-196. Gang, H., Y.J. Zhang & Y.Y. Zhu. 1989. Studies on the spatial distribution pattern and sampling tech- niques of predators in cotton fields. Natural En- emies of Insects 11:70-76. Charlet, L.D. & G.J. Brewer. 1998. Sunflower In- sect pest management. University of Minnesota. Available at http://ipmworld.umn.edu/chapters/ charlet.htm. Cobia, D.W. & D.E. Zimmer. 1978. Sunflowers: production and marketing. N, D. State University Extension Bulletin 25, 73 p. Duffield, S.J. & Y.V.R. Reddy. 1997. Distribution and movement of predators of Helicoverpa ar- migera in intercropped sorghum and short-dura- tion pigeonpea. Crop Research (Hisar) 14:315- 335. Hao, S.G., X.X. Zhang & X.N. Cheng. 2000. Ver- tical distribution and quantitative dynamics of dominant functional groups of arthropod com- munity in rice fields and estimation of natural enemies’ effects. Chinese Journal of Applied Ecology 11:103-107. He, H.J., M.S. Huang, X.X. Fan, S.Y. Liu & P.A. Jia. 1995. Moving pattern of natural enemies in relay interplanting fields of wheat with cotton during wheat harvest. Plant Protection 21:19-22. Herberstein, M.E. 1998. Web placement in sympat- ric linyphiid spiders (Arachnida, Araneae): indi- vidual foraging decisions reveal inter-specific competition. Acta Oecologica 19:67-71. Holland, J.M., L. Winder, C. Woolley, C.J. Alex- ander, & J.N. Perry. 2004. The spatial dynamics of crop and ground active predatory arthropods and their aphid prey in winter wheat. Bulletin of Entomological Research 94(5):4 19-431. Idris, A.B. & R.M.N.V. Mohamad. 2002. Vertical and temporal distribution of Aphis gossypii Glover and coccinellid population on different chilli (Caspicum annuum) varieties. Journal of Asia Pacific Entomology 5(2): 185-1 91. Jiratko, J., K. Veverka & J. Sedivy. 1996. Control of sunflower disease and pests. Studijni infor- mace, Rostlinna vyroba 2, 45 p. [in Czech] Kim, H.S., Y.I. Lee, K.M. Choi & H.P. Lee. 1989. Vertical distribution of spiders and planthoppers on rice plants. Research Reports of the Rural De- velopment Administration, Crop Protection 31: 36-39. Liu, W.X., M.L. Hou, EH. Wan & EL. Wang. 2003. Temporal and spatial niche dynamics of spiders and their control effects on cotton bollworms in 204 THE JOURNAL OF ARACHNOLOGY transgenic Bt cotton fields. Entomological Knowledge 40:160-163, Lynch, R.E. & J.W. Garner. 1980. Insects associated with sunflower in south Georgia. Journal of the Georgia Entomological Society 15:76-182. Men, U.B. & H.S. Thakre. 1998. Parasitoids, pred- ators and pathogens recorded on sunflower sem- ilooper, Thysanoplusia orichalcea F. in Vidarbha. PKV Research Journal 22:218-219, Nyffeler, M. 1982. Field studies on the ecological role of the spiders as insect predators in agroe- cosystems. Ph.D. Dissertation. Swiss Federal In- stitute of Technology, Zurich. 174 pp. Nyffeler, M. & R.G. Breene. 1992. Dominant in- sectivorous polyphagous predators in winter wheat: high colonization power, spatial disper- sion patterns, and probable importance of the soil surface spiders (Araneae). Deutsche Entomolo- gische Zeitschrift 39:177-188, Nyffeler, M. & K.D. Sunderland. 2003. Composi- tion, abundance and pest control potential of spi- der communities in agroecosystems: a compari- son of European and US studies. Agriculture, Ecosystems and Environment 95:579-612. Pekar, S. 1999. Life-histories and the description of developmental stages of Theridion bimaculatum., T. impressum and T. varians (Araneae: Theridi- idae). Acta Societatis Zoologicae Bohemicae 63: 301-309. Pekar, S. 2000. Webs, diet, and fecundity of Ther- idion impressum (Araneae: Theridiidae). Euro- pean Journal of Entomology 97:47-50. Royer, T.A. & D.D. Walgenbach. 1991. Predacious arthropods of cultivated sunflower in eastern South Dakota. Journal of the Kansas Entomolog- ical Society 64:112-116. Seiler, G.J., G. Zolnerowich, N.V. Homer & C.E. Rogers. 1987. Spider fauna of selected wild sun- flower species sites in the southwest United States. Great Basin Naturalist 47:279-281. Singla, N. 1999. Predatory potential of spider, Neoscona nauticus (Koch) on three insect pests of cotton and sunflower. Pest Management and Economic Zoology 5:147-149. Thomas, C.F.G., E.H.A. Hoi & J.W. Everts. 1990. Modelling the diffusion component of dispersal during recovery of a population of linyphiid spi- ders from exposure to an insecticide. Functional Ecology 4:357-368. Wang, H.Q. & H.M. Yan. 1989. Explorations on the community of spiders and insects in the late rice fields. Natural Science Journal of Hunan Normal University 12:73-80. Yan, H.N. 1988, Spatial patterns of spiders in citrus groves and their sampling technique. Natural Science Journal of Hunan Normal University 11: 346-351. Zhang, Y.Q., M.Z. Hou, W. Lu & R.G. Nong. 1997. Study on community structure and dynamics of spiders in maize fields. Journal of Guangxi Ag- ricultural University 16:268-275. Manuscript received 19 August 2004, revised 31 March 2005. 2005. The Journal of Arachnology 33:205-213 LABORATORY METHODS FOR MAINTAINING AND STUDYING WEB BUILDING SPIDERS Samuel Zschokke: Department of Integrative Biology, Section of Conservation Biology (NLU), University of Basel, St. Johanns- Vorstadt 10, CH-4056 Basel, Switzerland. E-mail: samuel.zschokke@ alumni. ethz.ch Marie E. Herberstein: Department of Biological Sciences, Macquarie University, Sydney, NSW 2109, Australia ABSTRACT. Web-building spiders are an important model system to address questions in a variety of biological fields. They are attractive because of their intriguing biology and because they can be fairly easily collected and maintained in the laboratory. However, the only published instructions for working with web-building spiders are somewhat outdated and not easily accessible. This paper aims to provide an up-to-date guide on how to best collect, keep and study web-building spiders. In particular, it describes how to obtain spiders by capturing them or by raising them from cocoons, how to keep and feed spiders in the laboratory and how to encourage them to build webs. Finally it describes how to document and analyze web building and web structure. Keywords: Data collection, laboratory manual, methodology, spider silk, spider web Web-building spiders are a popular model system to address questions in various scien- tific fields such as physiology, ecology, evo- lutionary biology, ethology and chemistry. Silk production, while not unique to this group, is its most characteristic feature (Craig 1997). Physiologists aim to understand how silk is produced while chemists investigate its properties and structure (e.g., Vollrath 1999; Knight & Vollrath 2001). Webs built out of silk are used to catch insects, making web- building spiders important predators, and even biological control agents (e.g., Riechert 1999; Symondson et al. 2002). As prey remains are often retained in the web post-consumption, prey capture can easily be assessed. The evo- lution of the web in itself has been studied extensively (e.g., Eberhard 1982; Coddington & Levi 1991; Benjamin & Zschokke 2004). Similarly, sexual cannibalism, prevalent in several families of web-building spiders, or sperm competition and cryptic female choice have been the focus of many exciting studies (see Elgar 1998; Eberhard 2004 for reviews). Web-building spiders are also attractive to scientists because they can be easily collected and maintained in the laboratory, allowing large sample sizes and large-scale experi- ments. However, many researchers who rec- ognize the value of spiders as model systems may be inexperienced in collecting and main- taining web-building spiders. With the present paper we aim to provide the necessary infor- mation, in the hope to foster cross-disciplinary studies on these fascinating creatures. With towards 40,000 described spider spe- cies (Platnick 2005), we cannot give specific information for each species. Such informa- tion can be obtained either when collecting the spiders in their natural habitat or from re- searchers experienced with that spider species. Here we focus on species we have worked with, i.e. mainly orb-web spiders, but attempt to make our recommendations applicable to all web-building spiders, especially since much research is still needed on webs of most non orb- web spiders. OBTAINING SPIDERS We recommend obtaining spiders by col- lecting them in the wild, thereby gaining a first impression of web structure and physical requirements. As most web-building spiders build webs only under favorable weather con- ditions, they are best found when it is neither raining nor very windy. We do not recom- mend obtaining spiders from dealers, since the source of these spiders is often unclear and they may be inbred. 205 206 THE JOURNAL OF ARACHNOLOGY Figure 1. — “Spi-pot’% a simple device to temporarily immobilize a spider for identifying or measuring (modified after Roberts 1995). It consists of two equally sized, round plastic pots. In one pot (pot A), a round hole is cut into its base, leaving a rim of c. 5 mm for rigidity. This hole is then covered with tightly stretched cling wrap, secured along the sides with tape. On the base of the other pot (pot B) a circular piece of soft foam is glued. The thickness of the foam should correspond to the gap at the base when two pots are stacked inside each other. To examine or measure a spider, place it in pot A, push pot B inside pot A and view the spider through the cling film (C). Capturing spiders* — To capture an orb- web spider sitting in its web, place a small jar around the spider and replace the cover from the other side of the web. If the web should remain undamaged for photos, tap the web op- posite the spider, causing the spider to drop down into a container held below. To capture a spider hiding in a retreat, either collect the entire retreat or lure the spider out of the re- treat by placing a vibrating tuning fork on the web (if no tuning fork is at hand, vibrating forceps sometimes also work; Penney 1995). Theridiid and linyphiid spiders may require a larger jar, lifted up quickly from below around the spider in its web. Do not use butterfly nets to capture web-building spiders as this can damage them. Immature or sub-adults females will have a longer life expectancy than adults and seem to thrive better in the laboratory, whereas adult males do not build webs and adult fe- males may soon start laying eggs, and then build less regular webs. For studying mating behavior, it is essential to control the spiders’ mating histories. Unfortunately, identifying sub-adult, live spiders in the field can be dif- ficult or impossible. A good aid to examine live spiders is the “Spi-pot” (Roberts 1995; Fig. 1). For transport, spiders can be housed in film canisters of 35 mm or APS films. Us- ing the semi-transparent variety, allows check- ing the spider without opening the canister. Leaves or twigs give the spider a substrate to hang onto and provide some humidity. Place spiders singly in containers to prevent canni- balism. Sending live spiders. — To send spiders by courier or ordinary airmail, put them into a fairly airtight container with a small piece of moist cotton or paper towel to prevent desic- cation. The air enclosed in the container is sufficient for many days, and feeding is not necessary during shipment. Legal aspects,- — In certain areas or coun- tries, capturing some or all spider species is not allowed or requires permits from the rel- evant authorities. Similarly, import and export permits and restrictions must be observed when sending or transporting spiders between countries. Rearing spiders from eggsacs. — The eas- iest starting point to rear spiders from eggsacs are gravid, mated females collected in the field. It is virtually impossible to know wheth- er a female has mated, but the likelihood of collecting a mated female increases with the progression of the season. Unmated females will also eventually lay eggs, albeit infertile ZSCHOKKE & HERBERSTEIN— LAB METHODS FOR WEB-BUILDING SPIDERS 207 ones. When the spider has built a cocoon, it must be exposed to appropriate climatic con= ditions, similar to those in its natural habitat. We found it helpful to keep cocoons of vari- ous web-building spiders in a chicken egg in- cubator made of Styrofoam and with a rough temperature control and a water reservoir to maintain humidity levels to prevent desicca- tion of the cocoons. Unfortunately, eggs often fail to hatch, and even if they do, rearing the spiderlings is a real challenge (see below). HUSBANDRY OF SPIDERS Enclosures (frames) to keep spiders. — A variety of frames have been used to study spi- ders and their webs. In the laboratory of Peter Witt, elaborate metal cages were used (Witt 1971). We suggest simpler frames entirely made out of Perspex. The frame’s size should correspond to the web size; initial field mea- surements may therefore be necessary. To house small to medium sized orb-web spiders (e.g., Zilla diodia (Walckenaer 1802), juvenile Araneus diadematus Clerck 1757 or Larinioi- des sclopetarius (Clerck 1757)), we used frames consisting of four pieces of transparent Perspex, 5 cm wide, 30 cm long and 3 mm thick, glued together with industrial strength glue at the corners (Fig. 2). Large orb- web spiders (e.g., diduXi Argiope sp.) require frames made out of 50 cm long Perspex pieces and adult Nephila sp. require even larger frames. Similar frames, but laid horizontally, can be used for sheet- web spiders (Bartels 1929). Spiders building three-dimensional webs (e.g., linyphiid and theridiid spiders) require cube- shaped frames. For some species it can be ad- vantageous to build the frames higher than wide. To facilitate the spider’s grip to the frame’s inside, apply net-like crack-seal tape, painted black beforehand to reduce unwanted reflections when later taking pictures of the web. To allow unobstructed examination of the web, the spiders must be kept in frames where two opposite sides can be removed. To separate the frames, place thin (0.5 mm), large (a few cm larger than the frames), trans- parent and somewhat flexible PVC sheets be- tween them. These sheets are smeared with Vaseline to deter spiders from attaching threads. Alternatively, windowpanes, which are kept very clean, can be used. In addition, puffy foam can be put along the edge of the frame, encouraging spiders to attach to that foam rather than to the glass. The frames are put on a shelve like books with the thin sheets placed between them (Fig. 3). When a spider has built a web, its frame can be easily taken from the shelf and placed in front of a shadow box for examination (see below, taking pic- tures). When handling the frame carefully, the spider usually stays in its web (or retreat). Some freshly caught spiders are likely to leap off the web or leave the hub of the web when their frame is handled for the first time, but will mostly become habituated to being han- dled after a few days. There are many alternatives to the durable Perspex frames described above, which may suit short-term or preliminary experiments, such as using rigid cardboard, wooden frames or 'slices’ of round plastic buckets with cling wrap to prevent spiders from escaping. Spi- ders building webs on rather than between supports can be offered an artificial, standard- ized structure to build their web on (Blackledge & Wenzel 2001), which is then placed inside a larger container with clean, smooth sides. For short-term storage of smaller spiders we use upturned plastic cups from which the bottom has been removed and replaced with a fine mesh. Smaller spiders will build small webs in these cups, which can be misted from the top (with a spray bottle) without lifting the cup. Alternatively, only a small hole is cut into the bottom of the cup and corked with a cotton plug or a tampon piece. Water is then administered by wetting the cotton plug. Keeping spiders in such small cups can also be an experimental procedure; e.g., when studying web-building behavior, larger spiders can be maintained in cups to temporarily pre- vent them from building webs (e.g., Reed et al. 1970; Herberstein et al. 2000). Feeding and watering. — Spiders can be fed with almost all kinds of insects, and most web-building spiders will attack and over- whelm insects trapped in their web in a large range of sizes, up to their own size or even larger. Drosophila flies are often used, as they are easily reared. When rearing Drosophila in bottles with sponge stoppers, spiders can be fed by trapping single flies between the flange of the stopper and the bottle, from where they can be introduced into the web using forceps. Spiders without webs are trickier to feed; some spiders accept live prey held near their mouth with forceps; buzzing flies are more 208 THE JOURNAL OF ARACHNOLOGY Figure 2. — Frame to keep orb-web spiders made out of four Perspex strips, glued together at the corners (not to scale). On the inside of the frame, blackened crack-seal tape has been applied to facilitate the spider’s grip. Similar frames with dimensions accordingly adapted can be used for other web-building spiders. readily accepted than kicking crickets. Sprin- kling water over the offered insect or breaking the insect’s cuticle a bit by snipping a cercus or antenna to release a drop of hemolymph can induce spiders to feed when the liquid touches the spider’s chelicerae. It is sufficient for most spiders to be fed once per week. However, since feeding spiders without web can be tricky, we recommend feeding spiders twice per week. Feed spiders at least so much that they do not loose substantial amounts of weight, causing their abdomens to shrink. Whereas some spiders can be kept for a pro- longed time on such a minimal diet (in our experience e.g., Araneus diadematus, Zygiella x-notata (Clerck 1757)), other species seem to falter when they are not given enough food to grow (e.g., Argiope bruennichi (Scopoli 1772)). Natural prey capture rates may pro- vide helpful starting points when designing feeding regimes in the laboratory. It is impor- tant to either feed the spiders with different insects or to feed the prey insects with high quality food (i.e. supplemented with proteins, vitamin-enriched cereal or pet food), as the spiders may otherwise experience deficiencies (Uetz et al, 1992; Mayntz & Toft 2001). The relatively dry air in most buildings makes spi- ZSCHOKKE & HERBERSTEIN— LAB METHODS FOR WEB-BUILDING SPIDERS 209 Figure 3. — Several frames (cf. Fig. 2) put side to side on a shelf. Thin PVC sheets placed between adjacent frames prevent the spiders from moving between frames. The thin sheets are smeared with Vaseline to prevent the spiders from attaching threads. The first and the last sheet are thicker, more stable ones, held up with bookends (not shown). ders kept inside vulnerable to desiccation. Thus regular misting with a water sprayer or placing a moist sponge at the bottom of the frame is vital. For experiments on the impact of drugs or pesticides on web building consult earlier studies on how to administer drugs (e.g., Witt et ah 1968; Witt 1971; Samu & Vollrath 1992; Hesselberg &, Vollrath 2004). Rearing spiderlings. — This is notoriously difficult and fraught with high levels of mor- tality. To rear spiderlings, place the freshly hatched cocoon into a container with support for the webs such as wood-wool, and add cul- tures of Drosophila or Collembola as food (Dinter 2004). Initially, some spiderlings will consume each other; more established ones then construct webs and capture prey. Do not separate the spiderlings too early as this can lead to almost total mortality. Encouraging web building. — Spiders vary greatly in their propensity to build a web in the laboratory. It is possible to find out which spider species build webs readily by identi- fying species used in earlier laboratory stud- ies. The most popular orb-web species include Araneus sp., Argiope sp., Nephila sp. and ulo- borid spiders. In contrast, Gasteracantha sp., Tetragnatha sp., Meta sp., Metellina sp. and Leucauge sp. are more hesitant to build webs. Feeding a spider that has not yet built a web in the laboratory, or putting a live fly into the cage together with the spider (Pasquet et al. 1994) can help to induce web building. If re- leasing a live fly into the frame with the spider is problematic because of strict feeding re- gimes, flies can be kept in a small jar with some sugar solution and covered by fly mesh. This way the flies buzz and stimulate web building without being captured by the spider (Herberstein et al. 2000). Web building fre- quency is also higher when spiders are ex- posed to natural day — night cycles in light and temperature (Witt 1956). Once the spider has built its first web in the laboratory, feed it soon to encourage the spi- der to build again. Web building frequency 210 THE JOURNAL OF ARACHNOLOGY varies between species. Whereas Araneus dia- dematus, Argiope sp., Larinioides sclopetarius and Zygiella x-notata generally rebuild the capture area of their web every night or every other night, Nephila sp. typically rebuild only sections of the web. Damaging webs. — Some orb-web spiders hesitate to rebuild their web as long as it is intact. To induce web rebuilding, it may there- fore be necessary to damage or destroy the webs. In the field, spiders generally leave the frame and the anchor threads largely intact when rebuilding the web (Carico 1986). Thus, only the capture area should be damaged to induce rebuilding, e.g., by cutting holes into the capture area with a red-hot wire (Fig, 4). Alternatively, cut the lateral anchor threads with scissors to destroy the entire orb-web. However, complete web destruction forces the spider to build the next one from scratch, which can influence aspects of the web (Zschokke & Vollrath 2000). In general, the spider should be allowed to ingest the old web (Peakall 1971). Damaging a single sticky spi- ral segment allows to determine whether the spider rebuilds the web during the next night. Non orb-web spiders do not remove, ingest and rebuild their web as orb-web spiders do, but keep repairing and extending them (Ta- naka 1989; Benjamin & Zschokke 2003, 2004). To study the construction of these webs, remove the old web completely or place the spider in a new frame. DATA COLLECTION Observing web building. — Observing spi- ders during web building is not easy because they are very sensitive to disturbance, espe- cially during the early stages of web building (which are therefore least well known; Zschokke 1996) and because the time of web building is generally during the night but oth- erwise largely unpredictable, likely depending on changes in temperature or light (Spronk 1935; Witt 1956). Observing web building is additionally impeded by the light sensitivity of most web-building spiders. Again, spiders differ in their sensitivity. Araneus diadematus, Nephila plumipes C.L, Koch 1839 and Arg/o- pe keyserlingi Karsch 1878 are fairly tolerant to some light and may even rebuild their web during the day, whereas other species (e.g., Nuctenea umbratica (Clerck 1757), Zygiella x-notata) will only build in absolute darkness. Figure 4. — Tip of a modified soldering gun used to selectively damage parts of a web. The thick wires (b) emerging from the front of the soldering gun (c) are either parts of the original tip or wires as found in 220 V wires. Soldered to these thick wires is a single strand of a 220 V cable (a). To ensure good contact between wire and strand, the strand is wrapped around the wire. These difficulties can be overcome by using automated spider tracking under infrared light (Benjamin & Zschokke 2002). This method additionally records the spider’s time budget, but neither records the position of threads nor all details of the spider’s behavior. Taking pictures. — Since spider web silk is very thin (c. 0.5 p.m-5 p.m), taking pictures of spider webs with all threads clearly visible is difficult. Earlier studies suggested placing the entire web in a box filled with ammonium chloride (Peters et al. 1950) or coating it with white glossy spray paint (Witt & Reed 1965) to increase thread visibility. However, these approaches require removing the spider from the web, they may distort the web and prevent spiders from ingesting the web, as orb-web spiders usually do (Peakall 1971), or to keep using it as non orb-web spiders do. Good pic- tures of spider webs can also be obtained with untreated webs. The main requirements are bright light from the sides and a very dark background, such as a shadow box lined with black velvet (Langer & Eberhard 1969; Zschokke 2002). We obtained satisfactory re- sults using two 15W fluorescent bulbs on ei- ther side of the web with an aperture of 4.5 and an exposure of 1 sec. when using a 55 mm lens on a SLR camera loaded with 100 ISO BAV film (Agfapan). To further improve picture quality, add two bulbs along the top and the bottom of the web. When using a dig- ital camera, a good resolution (at least 3-4 Megapixels) is essential. Since the picture is mostly dark, with fine white lines, use manual ZSCHOKKE & HERBERSTEIN— LAB METHODS FOR WEB-BUILDING SPIDERS 211 settings, as the automatic settings of most cameras will produce inferior to unusable re- sults. Every photograph should be recorded in a lab book and, to avoid any possible confu- sion, include a marker along the edge of the picture for identification, together with a scale and an indicator for the top of the web. Describing webs. — Several approaches have been proposed to estimate the area of orb-webs (Herberstein & Tso 2000 for Argio- pe sp. webs and Blackledge & Gillespie 2002 for webs of Cyclosa sp. and Tetragnatha sp.), as well as the total thread length as a measure of the spider’s investment (Heiling et al. 1998 for Larinoides sclopetarius webs and Venner et al. 2001 for Zygielia x-notata webs). These approaches require measurements of various web parameters, including number of spiral turns, and capture and hub area dimensions. Their suitability depends on the web shape, and whether field or laboratory measurements are made. Field measurements are difficult and it may be wise to select a formula re- quiring only few measurements; a reduced measuring accuracy can be compensated with a larger sample size (Zschokke & Liidin 2001). Even though measurements in the lab- oratory are easier and more precise, these for- mulae only provide estimates. To obtain ac- curate data, take a photograph of the web (see above) and import it into a graphics program that calculates area or thread length digitally. In the past, a multitude of names have been used for the various parts of webs. To avoid confusion, use established names (Zschokke 1999 for orb-webs). Similarly, with names of some spider species changing over the years, make sure to use the current species name (Platnick 2005). Measuring spiders. — Spider size refers to the length or width of a sclerotized body part, such as leg length (typically the tibia-patella length of the first leg is used) or carapace width. As these parts do not grow between molts, they provide information on the growth rate prior to the previous molt; and they can be relevant web parameters (e.g., leg length can influence mesh size in orb- webs; Vollrath 1987). Live spiders need to be immobilized for measuring with a Spi-pot (see above) or with CO2. When using CO2, gently blow CO2 into a sealable jar with the spider until the spider stops moving; taking care not to kill the spiders with too much CO2. Large spiders can be measured with electronic calipers, small ones under a dissecting microscope with an ocular fitted with a reticule. Keeping the exuviae of the spiders allows later size mea- surements. Spider weight is also an informa- tive and fairly easily obtained measure. Weight in addition to size can then be used to estimate recent foraging success by calculat- ing spider condition (weight / size or residuals of weight / size; Jakob et al. 1996; Kotiaho 1999). CONCLUSIONS Web-building spiders provide excellent models to test general and spider-specific hy- potheses. Collection and maintenance of ju- venile and adult spiders is relatively easy, en- suring large sample size and power. While rearing juveniles from eggs is difficult, some research groups have achieved relatively high rates of survival. Manipulation and observa- tion of web-building spiders in the laboratory is simple and can be achieved by non-arach- nologists by following some basic rules set out above. ACKNOWLEDGMENTS We thank (in alphabetical order) Todd Blackledge, Andreas Lang, Viktor Mislin, George Uetz, Fritz Vollrath, Andre Walter and two anonymous reviewers for advice and help on the practical side of spider keeping, for dis- cussion, and for comments on the manuscript. LITERATURE CITED Bartels, M. 1929. Sinnesphysiologische und psy- chologische Untersuchungen an der Trichterspin- ne Agelena labyrinthica (CL). Zeitschrift fiir ver- gleichende Physiologic 10:527-593. Benjamin, S.P. & S. Zschokke. 2002. A computer- ised method to observe spider web building be- haviour in a semi-natural light environment. Pp. 117 — 122. In European Arachnology 2000. (S. Toft & N. Scharff, eds.). Aarhus University Press, Aarhus. Benjamin, S.P. & S. Zschokke. 2003. Webs of ther- idiid spiders: construction, structure and evolu- tion. Biological Journal of the Linnean Society 78:293-305. Benjamin, S.P. & S. Zschokke. 2004. Homology, behaviour and spider webs: web construction be- haviour of Linyphia hortensis and L. triangularis (Araneae: Linyphiidae) and its evolutionary sig- nificance. Journal of Evolutionary Biology 17: 120-130. Blackledge, T.A. & R.G. Gillespie. 2002. Estima- 212 THE JOURNAL OF ARACHNOLOGY tion of capture areas of spider orb webs in rela- tion to asymmetry. Journal of Arachnology 30: 70-77. Blackledge, T.A. & J.W. Wenzel. 2001. State-deter- minate foraging decisions and web architecture in the spider Dictyna volucripes (Araneae Dic- tynidae). Ethology, Ecology and Evolution 13: 105-113. Carico, J.E. 1986. Web removal patterns in orb- weaving spiders. Pp. 306 — 318. In Spiders: Webs, Behavior, and Evolution. (W.A. Shear, ed.). Stanford University Press, Stanford. Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annual Re- view of Ecology and Systematics 22:565-592. Craig, C.L. 1997. Evolution of arthropod silks. An- nual Review of Entomology 42:231-267. Dinter, A. 2004, A mass rearing method for the lin- yphiid spider species Erigone atra (Blackwall) (Araneae: Linyphiidae). Journal of Applied En- tomology 128:200-203. Eberhard, WG. 1982. Behavioral characters for the higher classification of orb-weaving spiders. Evolution 36:1067-1095. Eberhard, W.G. 2004. Why study spider sex: special traits of spiders facilitate studies of sperm com- petition and cryptic female choice. Journal of Ar- acheology 32:545-556. Elgar, M.A, 1998. Sperm competition and sexual selection in spiders and other arachnids. Pp. 307 — 339. In Sperm Competition and Sexual Se- lection. (T.R. Birkhead & A.P. Mpller, eds.). Ac- ademic Press, San Diego. Heiling, A.M., M.E. Herberstein & G. Spitzer. 1998. Calculation of capture thread length in orb webs: evaluation of new formulae. Annals of the Entomological Society of America 91:135-138. Herberstein, M.E., A.C. Gaskett, D. Glencross, S. Hart, S. Jaensch & M.A. Elgar. 2000. Does the presence of potential prey affect web design in Argiope keyserlingi (Araneae, Araneidae)? Jour- nal of Arachnology 28:346-350, Herberstein, M.E. & I. Tso. 2000. Evaluation of for- mulae to estimate the capture area of orb webs (Araneoidea, Araneae). Journal of Arachnology 28:180-184. Hesselberg, T. & E Vollrath. 2004. The effects of neurotoxins on web-geometry and web-building behaviour in Araneus diadematus Cl. Physiology & Behavior 82:519-529. Jakob, E.M., S.D. Marshall & G.W Uetz, 1996. Es- timating fitness: a comparison of body condition indices. Oikos 77:61-67. Knight, D.P. & E Vollrath. 2001. Changes in ele- ment composition along the spinning duct in a Nephila spider. Naturwissenschaften 88:179- 182. Kotiaho, J.S. 1999. Estimating fitness: comparison of body condition indices revisited. Oikos 87: 399-400. Langer, R.M. & W. Eberhard. 1969. Laboratory photography of spider silk. American Zoologist 9:97-101. Mayntz, D. & S. Toft. 2001. Nutrient composition of the prey’s diet affects growth and survivorship of a generalist predator. Oecologia 127:207-213. Pasquet, A., A. Ridwan & R, Leborgne. 1994, Pres- ence of potential prey affects web-building in an orb- weaving spider Zygiella x-notata. Animal Behaviour 47:477-480. Peakall, D.B, 1971. Conservation of web proteins in the spider, Araneus diadematus. Journal of Ex- perimental Zoology 176:257-264. Penney, D. 1995, No tuning-fork? Try forceps! Newsletter of the British Arachnological Society 74:10. Peters, H.M., RN. Witt & D. Wolff. 1950. Die Beeinflussung des Netzbaues der Spinnen durch neurotrope Substanzen. Zeitschrift fiir verglei- chende Physiologic 32:29-45. Platnick, N.I. 2005. The world spider catalog. Ver- sion 5.5. http://research.amnh.org/entomology/ spiders/catalog. American Museum of Natural History, New York. Reed, C.E, RN, Witt, M.B. Scarboro & D.B. Peak- all. 1970. Experience and the orb web. Devel- opmental Psychobiology 3:251-265. Riechert, S.E. 1999. The hows and whys of suc- cessful pest suppression by spiders: insights from case studies. Journal of Arachnology 27:387- 396. Roberts, M.J. 1995. Spiders of Britain & Northern Europe. Collins Field Guide, London. 383 pp. Samu, F. & E Vollrath. 1992. Spider orb web as bioassay for pesticide side effects. Entomologia Experimentalis et Applicata 62:117-124. Spronk, E 1935. Die Abhangigkeit der Nestbau- zeiten der Radnetzspinnen Epeira diademata und Zilla-x-notata von verschiedenen Aussenbedin- gungen. Zeitschrift fiir vergleichende Physiolo- gic 22:604-613. Symondson, WO.C., K.D. Sunderland & M.H. Greenstone. 2002. Can generalist predators be ef- fective biocontrol agents? Annual Review of En- tomology 47:561-594. Tanaka, K. 1989. Energetic cost of web construc- tion and its effect on web relocation in the web- building spider Agelena limbata. Oecologia 81: 459-465. Uetz, G.W, J. Bischoff & J. Raver. 1992. Survi- vorship of wolf spiders (Lycosidae) reared on different diets. Journal of Arachnology 20:207- 211. Venner, S., L. Thevenard, A. Pasquet & R. Leborg- ne. 2001. Estimation of the web’s capture thread length in orb- weaving spiders: determining the ZSCHOKKE & HERBERSTEIN— LAB METHODS FOR WEB-BUILDING SPIDERS 213 most efficient formula. Annals of the Entomo- logical Society of America 94:490-496. Vollrath, E 1987. Altered geometry of webs in spi- ders with regenerated legs. Nature 328:247-248. Vollrath, E 1999. Biology of spider silk. Interna- tional Journal of Biological Macromolecules 24: 81-88. Witt, P.N. 1956. Die Wirkung von Substanzen auf den Netzbau der Spinne als biologischer Test. Springer, Berlin. 79 pp. Witt, P.N. 1971. Instructions for working with web- building spiders in the laboratory. Bioscience 21: 23-25. Witt, P.N. & C.E Reed. 1965. Spider web-building. Measurement of web geometry identifies com- ponents in a complex invertebrate behavior pat- tern. Science 149:1190-1197. Witt, P.N., C.E Reed & D.B. Peakall. 1968. A Spi- der’s Web: Problems in Regulatory Biology. Springer, Berlin. 107 pp. Zschokke, S. 1996. Early stages of web construc- tion in Araneus diadematus Clerck. Revue Suisse de Zoologie hors serie 2:709-720. Zschokke, S. 1999. Nomenclature of the orb-web. Journal of Arachnology 27:542-546. Zschokke, S. 2002. Ultraviolet reflectance of spi- ders and their webs. Journal of Arachnology 30: 246-254. Zschokke, S. & E. Liidin. 2001. Measurement ac- curacy: how much is necessary? Bulletin of the Ecological Society of America 82:237-243. Zschokke, S. & F. Vollrath. 2000. Planarity and size of orb-webs built by Araneus diadematus (Ara- neae: Araneidae) under natural and experimental conditions. Ekologia 19(Supplement 3):307-318. Manuscript received 14 September 2004, revised 8 August 2005. 2005. The Journal of Arachnology 33:214-221 THE LIFE HISTORY OF YLLENUS ARENARIUS (ARANEAE, SALTICIDAE)— EVIDENCE FOR SYMPATRIC POPULATIONS ISOLATED BY THE YEAR OF MATURATION Maciej Bartos: University of Lodz, Department of Teacher Training and Studies of Biological Diversity, Banacha 1/3, 90-237 Lodz, Poland. E-mail: bartos @biol. uni. lodz.pl ABSTRACT. The lifespan of F. arenarius is about 720 days for males and 750 days for females (maximum 770 days), which makes it the longest lived salticid reported from natural conditions. The juvenile spiders emerge at the beginning of June and mature not before the following August. They mate in autumn and hibernate for the second time. For most of the year two cohorts coexist, and at the beginning of June three cohorts can be found simultaneously. The life cycle suggests that in the studied areas there are two groups of individuals, the first of which produces young in odd years, while the other group reproduces in even years. The spider lifespan and phenology suggest no or limited gene flow between the groups. Keywords: Salticidae, life history, sympatric populations, isolation, Yllenus arenarius Spider life histories have received consid- erable attention, which has led to some gen- eral attempts at life cycle classification and understanding the factors responsible for cycle control, and for growth and development of spiders (rev. in Schaefer 1987; Vollrath 1987). The specific knowledge of life cycles is, how- ever, restricted to some groups, while others remain poorly known. The knowledge of salticid life cycles is very scarce. The spiders are relatively small and occur in low densities, so they have at- tracted little interest concerning both general biology and ecology (Wise 1993). Since the works published by the end of the seventies (e.g.: Horner & Starks 1972; Edwards 1975; Jackson 1978) few papers on the topic have appeared (e.g.: Matsumoto & Chikuni 1987). Instead, salticids have recently been the sub- ject of intensive behavioral studies (rev. in Jackson & Pollard 1996) and some species have gradually become models in studies of invertebrate cognition (e.g.: Wilcox & Jackson 1998). The knowledge of the model’s general biology is often, however, essential for the proper interpretation of ecological and behav- ioral data. Jumping spiders are commonly character- ized as possessing stenochronous life cycles (Schaefer 1987). From particular studies car- ried out on species occurring in temperate cli- mate we know that they have annual or bi- ennial life cycles (e.g.: Horner & Starks 1972; Jackson 1978; Matsumoto & Chikuni 1987). Such information can also be deduced from the occurrence of sexually mature individuals, which is sometimes given in local keys (Pro- szynski 1991). The current work aims to describe the life history of Yllenus arenarius Menge 1868, which is a medium-sized jumping spider with an adult body length of about 7 mm. Typically for Salticidae, sexual dimorphism is poorly marked in body size but distinct differences are exhibited in coloration. It is known only from Central and Eastern Europe with the westernmost localities in NW Germany and the easternmost on the river Volga (Proszynski 1991; Logunov & Marusik 2003). The spider inhabits sandy dunes along rivers as well as inland and coastal dunes. In the habitats it is restricted to the initial stage of the Spergulo- Corynephoretum and tends to keep away from dense vegetation (Zabka 1997; Bartos 2000; Merkens 2002). Yllenus arenarius is a polyphagous salticid preying on a wide spectrum of invertebrates (Bartos 2004). It is also known for its condi- tional hunting tactics flexibly adjusted to prey type (Bartos 2000, 2002a). Another interesting 214 BARTOS-LIFE HISTORY OF Y. ARENARIUS fact concerning the spider’s biology is its un= usual sub-sand nests built for various purpos- es: molting, egg-laying, surviving the period of night and hibernating. The nest types differ according to their size, shape and general structure (Bartos 2002b). METHODS Study period and area. — The research carried out from 1998-2003 encompassed 14 populations from Central and Eastern Poland. One of the sites (Kwilno), located in Central Poland about 25 km to the north of Lodz, was visited at least every two weeks during the vegetation season. Data from the other sites were collected more occasionally, but with re- spect to spider phenology and body size they were consistent with the data from Kwilno and therefore they were pooled. Morphometric measurements and phe- nology.— Spiders were collected by means of visual searching through the dune surface be- tween 10:00 and 12:00 hours. During each field visit it was attempted to collect at least 40 individuals (10 juveniles from each cohort, 10 females and 10 males). Spiders from all age groups show the same pattern of variation in activity during the day, therefore the sam- ples seem to be representative for the whole population (Bartos pers. obs.). In anticipation of hatched juveniles or adults after the final molt the field was visited every day starting at least ten days before the expected time. Three measurements of live specimens were taken with a stereomicroscope (preci- sion: 0.01 mm) (n = 1208): abdomen length (AL), abdomen width (AW) and posterior eyes width (PEW). To immobilize the spiders during the measurements they were covered with a transparent kitchen foil and delicately pressed against a piece of sponge. Sex and age of the spiders were also recorded. The char- acteristics allowed to include the spiders to certain age groups (Figs. 1-3). Winter samples. — In winter a two-centi- meter thick layer of sand was collected from the dune surface and dried in the laboratory (temperature at ca 25 °C). Spiders, which emerged from the sand were collected, mea- sured and subsequently reared until they died. The dried sand was also sieved to collect nests and immersed spiders. Air temperature at the level of the sand surface was measured with an alcohol thermometer (precision: 0.2 °C). 215 Rearing. — Females were kept in the labo- ratory to estimate the number of eggs, the pe- riod of egg laying and determine the place where they are laid. Spiders were reared in- dividually in glass containers (1 liter) with a three centimeter-thick layer of dune sand on the bottom. Temperature was maintained at ca 25 °C, light regime 12L:12D and the sand was moistened weekly with 5 ml of water. They were fed ad libitum (10 fruit flies twice a week). Under these conditions 26 females were kept. Six females came from winter sand samples and the rearing started in February, while 20 other females were collected at the beginning of April. The sand from the labo- ratory containers was sieved every two weeks in order to collect nests. The nests were opened and checked to find eggs and exuviae. The individuals in the laboratory survived un- til mid June. Data analysis. — All statistical procedures followed those described by Zar (1984). The significance of the differences in body length parameters was tested with one-way ANOVA and Tukey test with unequal sample sizes. Data are presented as mean ± SD (n) except for Figs. 1-3 which present mean ± 1.96 SE. RESULTS Morphometric measurements and phe- nology.— Taking into account the season in which the spiders were collected, spider size and maturity, four age groups were distin- guished (Figs. 1-3): juveniles in the first sea- son of life (juv-I), juveniles in the second sea- son of life (juv-II), which underwent maturation in August, adults in the second season of life (ad-II) and adults in the third season of life (ad-III), In each of the studied seasons two spider cohorts were observed for the whole season (either juv-II and ad-III or juv-I and juv-II/ad-II) and, in June, spiders from ail three cohorts were observed simul- taneously (juv-I, juv-II and ad-III). Spiders from the same cohort were ob- served for three consecutive years (Figs. 1- 4). Males were found even up to 720 days from leaving their sub-sand nests. The latest recording time for females (n = 3) was 15 July in 1998 and 2003. Their lifespan, calcu- lated from the time of emerging from the sub- sand nest, is about 750 days. The females, however, were numerous only by the begin- 216 THE JOURNAL OF ARACHNOLOGY April May June July August Sept. Oct. Date of measurement Figures 1-3. — Changes of three moiphometric parameters in the life cycle of Yllenus arenarius. 1. Pos- terior eyes width; 2. Abdomen width; 3. Abdomen length; squares, juv-I {n — 573); triangles, juv-II (n = 363); diamonds, males (/? = 124); circles, females {n = 148); symbols are means; error bars are 1.96 SE. BARTOS-LIFE HISTORY OF K ARENARIUS ning of June. Later only scattered individuals were recorded. Spiders of all age groups were active from the first 10 days of April to mid October as long as the air temperature measured on the dune surface was above ca 10 °C. Spiders were never found when the temperature was lower than 10 °C. Therefore, because of over= all harsh autumn weather conditions, in most seasons of study they disappeared before the end of September, rarely at the beginning of October (Figs. 1-3). On several warm winter days in February, however, several females were found hunting on the dune surface (at temperature 11.4 °C), All age groups started their activity at the same time in spring and were also found to burrow almost simulta- neously for hibernation in autumn. Juveniles in the first season of life (juv-I): During three study seasons, when the newly emerged juveniles were searched for, they ap^ peared within almost the same period of time (between 3 and 6 June). Their body length was about 1.87 ± 0.14 mm {n == 30), they had semitransparent carapace and their round cephalothorax and abdomen closely resem- bled the dune sand grains in color, size and shape. Shortly after emergence the spiderlings were found in groups of up to six individuals remaining only a few meters from each other, which suggests that they emerged from the same sac. Later, when the spiders started to disperse, they were rather evenly distributed over the dune. However, even a week after the first juveniles appeared on the surface, some groups consisting of a few spiderlings were also found, which suggests that they had just left the hatching chamber. Juveniles in the first season of life were observed throughout the summer until the beginning of October, when the first hibernation started (Figs. 1-3). Juveniles in the second season of life (juv- II): The spiders finished their hibernation at the beginning of April. In mid June the distal parts of the pedipalps of some juveniles began to swell, which made it easy to determine them as subadult males. They reached the sub- adult stage at different times and swollen ped- ipalps were commonly observed in subadult males only in mid July. Other characteristics of both sexes prior to final molting were in- distinguishable. Spider color and body pattern, so different in adult males and females, were identical in subadults. In the last week of July 217 almost all juveniles in the second season of life burrowed to undergo the final molt in their sub-sand nests. The first to disappear, how- ever, were subadult males. At that period only juv-I were commonly found on the surface. Such pattern of simultaneous disappearing of almost all individuals from a cohort, while the other cohort did not show apparent differences in number, was observed several times in the field. However, such absence was recorded for only a few days, which cannot be presented in the two-week-long periods in the figures. Adult spiders: The first adults to appear were males, observed as early as 8 August. During all three years, when the first mature males were particularly searched for, they ap- peared regularly between 8-10 August. The first adult females were observed at least ten days later. Male coloration and body pattern changed significantly after the last molt. Gen- erally, after the last molt males became much more conspicuous and easy to spot while fe- male cryptic coloration remained unchanged. Female coloration also changed throughout their mature life mainly due to loosing scales in the course of burrowing. In June, i.e. after 10 months from the last molt they were much darker, with patches of black cuticle visible in the areas where the scales were missing. In extreme cases the dorsal area of their abdo- men was black. This makes them unmistak- able from freshly molted females. Not only general appearance, but also some body measurements of adult females changed over their mature life (Figs. 1-3). There were no differences in posterior eyes width (Fig. 1) (^8;i43 ~ 1.38, P > 0.05). Such differences were found, however, for abdomen length (^9;i37 = 4.89, P < 0.001) and abdomen width (^9;i39 = 5.71, P < 0.001). Freshly molted adult females had on average shorter abdo- mens than those in the next spring by the end of May (significant at F < 0.05), but not the females in the second half of April {P > 0.05). A similar tendency was observed for abdomen width, being thinner in females shortly after the final molt in comparison with those in the first half of April and in the first half of May {P < 0.05), but again no differences with fe- males in the second half of April. The latter group had on average shorter abdomens than the females directly before and after the pe- riod {P < 0.05). In the group of adult males the differences 218 THE JOURNAL OF ARACHNOLOGY Figure 4. — Schematic presentation of the life cycle of Yllenus arenahus. grey bars = spiders hatching in odd years; black bars = spiders hatching in even years. were found only between those measured in the second half of August and early next spring. The hrst group was larger than the sec- ond according to posterior eyes width (F’6;i2i = 3.01, P < 0.01) (significant at F < 0.05), abdomen length = 8.38, P < 0.001) {P < 0.001) and abdomen width (F^-iog = 4.17, P < 0.001) (P < 0.005). Winter samples. — In winter sand samples collected in February at ambient temperature of — 8 °C and under ten-centimeter-thick snow cover, 26 live individuals were found. These were juv-II {n = 12), ad-III (n = 14): eight males and six females. Spiders from all age groups hibernated successfully. Spiders were found in only one out of four sand samples. Two spiders were found in elongated nests im- pregnated with organic matter, while the other individuals were being active at the time and were collected from the sand surface. Their winter nests were found after sieving the sand. Rearing. — Females reared in the laboratory (/? = 26) laid on average 6 ± 0.8 eggs {n = 5). The eggs were spherical or slightly oval, on average 1.20 mm in diameter (SD = 0.17 mm, n = 4). In all cases the eggs were laid in one batch only. The nest in which eggs were laid was different from other nests found in the same container. It possessed a specific structure. It was made of dense silk and sand grains and possessed two chambers. Eggs were attached to the wall with a sheet of silk slightly pressing them to the wall. Inside, the eggs and empty chorions were found. The spi- derlings hatched in a small chamber of the nest, and molted for the first time in the big chamber, where their exuviae were found. The actual process of egg-laying was not observed since it takes place inside an opaque, underground nest. For this reason the date of egg-laying in the field is also unknown. How- ever, one of six females found in winter sand samples laid eggs in laboratory conditions af- ter about two weeks from interrupting its hi- bernation. About four weeks later spider nymphs emerged from the underground nest. The spiderlings in laboratory performed bur- rowing behavior and built oval, thin-walled sub-sand nests soon after their first appearing on the surface. DISCUSSION The life history analysis suggests, that the lifespan of Y. arenarius counted as the time from leaving the sub-sand nests to the last in- dividuals observed in the field is about 720 days for males and 750 days for females (maximum 770 days), which makes it the lon- gest-lived salticid reported from natural con- ditions (Jackson 1978; Horner & Starks 1982; Matsumoto & Chikuni 1987). There are reports of Sitticus fasciger, which lived over 800 days in the laboratory, but only up to 428 days in the field (Matsumoto & Chikuni 1987). From the exceptional variability in the rate of development and lifespan of spiders BARTOS-LIFE HISTORY OF Y ARENARIUS (e.g.: Turnbull 1962; Toft 1983) we may ex- pect that the lifespan of the studied spider may also be variable, as local weather conditions and food availability fluctuate. The life cycle of Y. arenarius is character-- ized by an at least potentially long reproduc- tive period, which may last for about two months in autumn and for another two months the next spring. Copulating spiders were, however, found only in autumn. The possible reason is, that as time elapsed, females were getting less receptive (Bartos pers. obs.) and male condition was also getting worse, espe- cially in spring (Figs. 1-3). Therefore most likely copulation occurred in autumn and egg- laying took place the following spring. The spiders most probably lay a few large eggs in one batch. Laying multiple batches in the field cannot be excluded, however. Large egg dimensions in comparison to the largest female’s abdomen size (Figs. 2, 3) suggest that the total number of eggs laid at one time must be close to the number observed in one batch. Semelparity is also suggested by rather uniform size of spiderlings (at least during the first few months). If there were more than one clutch, they would have to be separated by 1- 4 weeks and as a result the spiderlings would differ in condition and size (Horner & Starks 1972; Jackson 1978; Matsumoto & Chikuni 1987), which is not the case here (Figs. 1-3). Even though the exact period of egg laying cannot be directly indicated, it may occur ei- ther in mid April, when female abdomens rap- idly shrink or in mid May. At the beginning of this month female abdomens are the largest in the whole life cycle (Figs. 2, 3). No such tendency was observed in average posterior eyes width (Fig. 1), which suggests that the group of measured females did not differ in overall size but only according to their abdo- men size. Even though there are two periods when female abdomen shrinks, which may suggest egg laying, the process most likely oc- curs in May, which is about a month before a new cohort emerged. This is consistent with average period of egg development and nest residence by spider larvae and nymphs (Hor- ner & Starks 1972; Jackson 1978; Matsumoto & Chikuni 1987). The early period of female abdomen shrinking is possibly due to low prey availability, prey becoming more numer- ous only in late June after juveniles leave their nests (Bartos pers. obs.). Such synchroniza- 219 tion of egg-laying with food availability has been commonly reported (Almquist 1969; Schaefer 1987). Underground nest location with no signs of repeated nest visiting (Bartos 2002b) imply the lack of brood care. Life history traits of T. arenarius such as low fecundity and relatively large eggs, slow development, delayed reproduction, long life span and a degree of territoriality (Bartos pers. obs.) place this species as a K-selected organ- ism, well adapted to the unfavorable environ- ment. As one of major predators, outnum- bered only by ants (Bartos pers. obs.) it seems to be a successful competitor in the environ- ment. Apart from the spiders’ cryptic color- ation, the key adaptation to survival in the cover-free habitat seems utilization of under- ground space, i.e. the burrowing and under- ground nest building, so typical for many an- imals dwelling in arid environments (e.g.: Gwynne & Watkis 1975; Cloudsley-Thomp- son 1983; Henschel 1990). Underground nests provide the spiders with more stable condi- tions, shelter against night active predators, strong wind and periods of inclement weather such as heavy rains, which may be a severe mortality factor. The importance of the nests is also suggested by the number of nests built especially by juveniles. In the laboratory the juveniles built nests daily, which is signifi- cantly more often than in subadults and adults (Bartos 2002b). Such a high rate of nest build- ing in juveniles connected with silk produc- tion and apparently energy-demanding under- ground nest building must be an important expenditure in the energy budget at the ex- pense of other processes, e.g. growth and de- velopment. Morphological and phenological data sug- gest that the spiders lay eggs two years from the time they hatch. However, in the field, newly hatched spiders are found every year. This suggests that in the dunes of Central and Eastern Poland, there are two sympatric pop- ulations of Y. arenarius reproducing in odd and even years (Fig. 4). Phenology of males and females in the field suggests that there is no gene flow between the groups or it occurs accidentally and must be limited. Gene flow may take place if adults from one cohort in spring survive until they meet sexual partners from the other cohort in August. This is very unlikely, though not impossible. Males would have to live for another two months and fe- 220 THE JOURNAL OF ARACHNOLOGY males for one month longer than the most long-lived individuals in the field. This was not observed in the period of studies since very characteristically looking old females were never recorded after mid July. For fe- males such prolonged survival would also mean to live even longer, for another nine months until the next spring (and hibernate for the third time), when eggs are laid and young were found to emerge from egg sacs. The apparent reproductive isolation may be, however, incomplete if at least a small pro- portion of the population reproduces every year or every three years or the immigration from population of an annual cycle (if there is one) occurs. Such phenomena were reported for several spider species (Toft 1976, 1983) and in the recent research on the reproductive isolation of Araneus diadematus were the most probable causes of the lack of genetic differences between markedly separated gen- erations (Johannesen & Toft 2002). Another potential cause of gene flow between succes- sive cohorts may be prolonged hibernation or aestivation. However, it seems very unlikely since it would require surviving several months while being immersed in hot and dry sand. The temperature at the depth the nests are built exceeds 50 °C in hot summer days (Bartos pers. obs.). On the whole, no evidence supporting the alternative scenarios were gath- ered over the period of studies. Interestingly a very similar phenomenon of two sympatric groups isolated by the season of reproduction was described in another salticid, Sitticus fas- ciger (Matsumoto & Chikuni 1987). It is curious how such a pattern evolved in the first place. Certainly not as a result of the appearance of early and late maturing adults (Schaefer 1987), since we would then find slowly and quickly developing individuals. In- stead we observe relatively uniform growth in the whole cohort of individuals. Two hypotheses seem to be most likely: a) if all populations of Y. arenarius require three seasons for development, then a part of them might have been shifted by one season and later mixed with the original group, b) if southeastern populations have shorter cycles, then repeated migrations might have resulted in the pattern observed. Both hypotheses as- sume two allopatric populations, which pos- sessed cycles shifted from each other by one year. After mixing they formed sympatric groups isolated by the year of reproduction. Whichever hypothesis is correct, the life cycle pattern we can observe now is probably in- debted to the well-known variability of spider rate of development depending on local con- ditions (Schaefer 1987) resulting in prolonged or shortened cycles (e.g.: Jackson 1978; Toft 1983). Another interesting question for specula- tions is: when did it happen? Nowadays this stenotopic species inhabits most commonly well isolated dunes. Ballooning has never been observed and seems an unlikely way of reaching another dune, which is a rare habitat not only in Poland, but in all Europe. Sandy areas were, however, more common in the past. So, did the switch happen as long ago as the time after the last glaciation, when bare moraine sands were common in Central Eu- rope? The author hopes that planned studies will help to test these hypotheses. ACKNOWLEDGMENTS I would like to thank Zbigniew Wojcie- chowski and two anonymous referees for helpful comments and suggestions. This re- search was supported by Polish Ministry of Scientific Research and Information Technol- ogy (grant 6P04F 072 15 and 3P04F 058 22). LITERATURE CITED Almquist, S. 1969. Seasonal growth of some dune- living spiders. Oikos 20:392-408. Bartos, M. 2000. Cykl zyciowy i strategia polo- wania pajgika Yllenus arenarius Menge, 1868 (Araneae, Salticidae). Ph.D. Thesis. University of Lodz, Lodz. Bartos, M. 2002a. Distance of approach to prey is adjusted to the prey’s ability to escape in Yllenus arenarius (Araneae, Salticidae). Pp, 33-38. In European Arachnology 2000: Proceedings of the 19* European Colloquium of Arachnology, Aar- hus (S. Toft & N. Scharff, eds). Aarhus Univer- sity Press, Aarhus. Bartos, M. 2002b. The sub-sand nests of Yllenus arenarius (Araneae, Salticidae): structure, func- tion and construction behavior. Journal of Arach- nology 30:275-280. Bartos, M. 2004. The prey of Yllenus arenarius (Araneae, Salticidae). Bulletin of British Arach- nological Society (in press). Cloudsley-Thompson, J.L. 1983. Desert adaptations in spiders. Journal of Arid Environments 6:307— 317. Edwards, G.B. 1975. Biological studies on the jumping spider, Phidippus regius C.L. Koch. M.Sc. Thesis University of Florida, Gainesville. BARTOS-LIFE HISTORY OF K ARENARIUS 221 Gwynne, D.T & J. Watkiss. 1975. Burrow-blocking behaviour in Geolycosa wrightii (Araneae, Ly- cosidae). Animal Behaviour 23:953-956. Henschel, J.R. 1990. Spiders wheel to escape. South African Journal of Science 86:151-152. Horner, N.V. & K.J. Starks. 1972. Bionomics of the jumping spider Metaphidippus galathea. Annales of the Entomological Society of America 65: 602-607. Jackson, R.R. 1978. Life history of Phidippus john- soni (Araneae, Salticidae). Journal of Arachnol- ogy 6:1-29. Jackson, R.R. & S.D. Pollard. 1996. Predatory be- havior of jumping spiders. Annual Review of En- tomology 41:287-308. Johannesen, J. & S. Toft. 2002. A test for repro- ductive separation of alternative generations in a biennial spider, Araneus diadematus (Araneae, Araneidae). Journal of Arachnology 30:65-69. Logunov, D.V. & Y.M. Marusik. 2003. A revision of the genus Yllenus Simon, 1868 (Arachnida, Araneae, Salticidae). KMK Scientific Press Ltd., Moscow. Matsumoto, S. & Y. Chikuni. 1987. Notes on the life history of Sitticus fasciger (Simon 1880) (Araneida, Salticidae). Journal of Arachnology 15:205-212. Merkens, S. 2002. Epigeic spider communities in inland dunes in the lowlands of Northern Ger- many. Pp. 215-222. In European Arachnology 2000: Proceedings of the 19* European Collo- quium of Arachnology, Aarhus (S. Toft & N. Scharff, eds). Aarhus University Press, Aarhus. Proszynski, J. 1991. Salticidae. Pp. 488-523. In Spinnen Mitteleuropas. (S. Heimer & W. Nen- twig, eds.). Parey Verlag, Berlin. Schaefer, M. 1987. Life cycles and diapause. Pp. 331-347. In Ecophysiology of Spiders. (W. Nen- twig, ed.). Springer- Verlag, Berlin, New York. Toft, S. 1976. Life-histories of spiders in a Danish beech wood. Natura Jutlandica 19:5-39. Toft, S. 1983. Life cycle of Meta segmentata (Clerck, 1757) and Meta mengei (Blackwall, 1869) in Western Europe (Arachnida: Araneae: Tetragnathidae). Verhandlung des Naturwissen- schaftlichen Vereins in Hamburg 26:265-276. Turnbull, A.F. 1962. Quantitative studies of the food of Linyphia triangularis (Clerck) (Araneae, Lin- yphiidae). Canadian Entomologist 94:1233- 1249. Vollrath, F. 1987. Growth, foraging and reproduc- tive success. Pp. 357-370. In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer- Verlag, Ber- lin, New York. Wilcox, R.S. & R.R. Jackson. 1998. Cognitive abil- ities of araneophagic jumping spiders. Pp. 411- 433. In Animal Cognition in Nature: the Con- vergence of Psychology and Biology in Laboratory and Field. (R.R Baida, I.M. Pepper- berg & A.C. Kamil, eds.). Academic Press, New York. Wise, D.H. 1995. Spiders in ecological webs. 1st ed. Cambridge University Press, Cambridge. Zar, J.H. 1984. Biostatistical analysis. 2nd ed. Pren- tice-Hall International, Inc., Englewood Cliffs, New Jersey. Zabka, M. 1997. Salticidae. Paj^ki skacz^ce (Arachnida: Araneae). Fauna Poloniae, 16. Mu- zeum i Instytut Zoologii PAN, Warszawa. Manuscript received 16 September 2004, revised 17 June 2005. 2005. The Journal of Arachnology 33:222-229 SPATIAL ASSOCIATION BETWEEN A SPIDER WASP AND ITS HOST IN FRAGMENTED DUNE HABITATS Dries Bonte^ and Jean-Pierre MaelfaiP’^: * Ghent University, Dep. Biology, Research group Terrestrial Ecology, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium. E-mail: dries.bonte@Ugent.be; ^Institute of Nature Conservation, Kliniekstraat 25, B-1070 Brussels, Belgium ABSTRACT, In patchily distributed habitats, species potentially occur wherever conditions are suitable or show a restricted distribution, influenced by patch quality, geometry and configuration. If patch isolation appears to be the main determinant of the species’ distribution then dispersal ability is supposed to be limited. Although only scarce literature is available, dispersal limitation seems to be an important factor in determining the spatial population structure in spiders. In this paper, we document on the spatial population structure of the rare wolf spider Alopecosa fabrilis, restricted to fragmented grey dunes along the Flemish coast (Belgium) and ask whether its distribution appears to be affected by aspects of patch configuration. Simultaneously, we investigated whether the local distribution of its main parasitoid, the spider wasp Arachnospila nifa (Hymenoptera, Pompilidae) was associated with its host. Our results in- dicate that A. fabrilis shows an aggregative population structure, which is determined by the distance to nearest occupied patch, indicating that spatially correlated habitat quality probably determine its occur- rence. Although spider wasps are generally characterized as non-specialists, the almost complete covari- ation between its spatial occurrence and that of A. fabrilis, indicates that spider hunting wasps may, at least temporally and locally, show a restricted host-range. As a result, the presence of a rather generalist parasitoid is a good predictor for the presence of nocturnal and burrowing dune wolf spider. Keywords: Alopecosa fabrilis, Lycosidae, Arachnospila rufa, Pompilidae, metapopulations In fragmented habitats, species live either in patchy populations or in metapopulations (Harrison 1991; Hanski 1999). In patchy pop- ulations, individuals move freely among hab- itat patches, while in metapopulations, most individuals stay in a single patch during their entire life, but dispersing individuals enhance strong colonization-extinction dynamics. This results in a population structure in which some suitable patches remain vacant. Besides the colonization abilities, changes in habitat qual- ity can also attribute to local extinction dy- namics, as demonstrated for specialized but- terflies (e.g. Thomas et al. 1992; Ravenscroft 1994; Moilanen & Hanski 1998; Bergman 1999) and backswimmers of the genus Noto- necta (Briers & Warren 2000). Therefore, studying the spatial structure of populations during successive years by considering as- pects of patch geometry, quality and config- uration enables us to assess actual dispersal limitation and population dynamics in an in- direct way. Snapshots of patch-occupancy in- cidences provide an alternative indirect meth- odology to study dispersal abilities or populations viability with respect to popula- tion size (patch size) resulting from historical habitat fragmentation and population dynam- ics. As shown for a coastal dune wolf spider, living in fragmented grassland habitats, patch occupancy patterns may depend both on as- pects of habitat quality and on different modes of dispersal through the surrounding matrix (Bonte et al. 2003a). For spiders in general, cursorial dispersal and ballooning induce dif- ferent colonization and extinction patterns, de- pendent on the species' mobility, its niche breadth and its propensity for aerial dispersal (Bonte et al. 2003a, 2004b). Especially bal- looning dispersal appears to be important for occupancy patterns at extended temporal and spatial scales (Bonte et al. 2003a, 2004a), while cursorial dispersal induces short-term colonization events at small spatial and tem- poral scales (Bonte et al. 2003a). For spiders, direct estimates of realized dispersal are rare and indicate dispersal distances up to few- hundred meters within suitable habitat (Krei- 222 BONTE & MAELFAIT— ASSOCIATION BETWEEN SPIDER WASP AND HOST 223 ter & Wise 2001; Bonte et aL 2003a; Samu et al. 2003) but restricted dispersal up to few me- ters in unsuitable, often densely vegetated habitat (Bonte et al. 2003a). Indirect estimates of dispersal, estimated by population genetic analyses, are more common (e.g. Boulton et al. 1998; Ramirez & Fandino 1996; Ramirez & Haakonsen 1999; Gurdebeke et al. 2000). In contrast to the latter, studies on occupancy incidences, ideally conducted during several successive years, reveal indirect estimates of dispersal that are independent of historical bottlenecks and selection pressures in spatially structured habitats (Peterson et al. 2001; Bon- te et al. 2003a). In addition to earlier studies on the mobile wolf spider Pardosa monticola (Clerck 1757), we here report on the population structure in coastal dune habitats of the specialized fos- sorial wolf spider Alopecosa fabrilis (Clerck 1757) with limited aerial dispersal abilities (Bonte et al. 2003b) and of its main (winged) parasitoid, Arachnospila rufa (Haupt 1927) (Hymenoptera, Pompilidae), knowing to host on larger Alopecosa wolf spiders (Koomen & Peeters 1993; Peelers et al. 1994). According to theoretical work, host-parasitoid metapo- pulations dynamics strongly interfere and oc- cupancy patterns should strongly overlap in case of limited host interpatch dispersal abil- ities and relatively low infection rates and par- asitoid survival rates (Hassel et al. 1991; Comins et al. 1992; White et al. 1996; Hanski 1999). The population structure of parasitoid and host were investigated in the Flemish coastal dunes, where grey dune habitats, being the op- timal habitat for both A. fabrilis and A. rufa, are truly fragmented since the Second World War (see e.g. Bonte et al. 2003a; Provoost & Bonte 2004). This was done in order to test the following hypotheses: (i) A. fabrilis has limited dispersal abilities, resulting in an iso- lation-dependent population structure, (ii) A. fabrilis, having low population densities only occupies larger habitat patches and (iii) local and temporal host-parasitoid association result in similar distribution patterns, although Pom- pilidae have better dispersal abilities and are believed to be rather generalistic in prey choice according to prey species (Finch 1997). Additionally, as A. fabrilis is nocturnal while its parasitoid is diurnally active, we asked whether the presence of a day-active parasit- oid could be an indicator for the presence of populations of a nocturnal, fossorial spider. METHODS Study Area, Study Species. — Fieldwork was conducted in the Flemish coastal dunes, located between the cities of De Panne (Bel- gium) and Bray-Dunes, France (51°05'N, 2°32'E) consisting of 52 discrete grey dune patches, varying between 0.05 and 27.6 ha (Fig. i). Grey dunes, known as “Fixed coastal dunes” are most readily defined using plant communities. Vegetation includes Atlantic moss dominated dunes (mainly Tortula rur- alis) as well as dune grassland (with a distinct organic soil layer) belonging to the Cladonio- Koelerietalia syntaxon in case of lime rich grey dune, and to the Trifolio-Festucetalia ovi- nae syntaxon in case of decalcified grey dunes (Provoost et al. 2002). In this study, only grey dunes without substantial soil development were surveyed since they are the habitat of both study species (see further). Ecologically it is merely the dry component of the “stressed dune landscape”. The main differ- entiating processes are related to dune fixa- tion, soil formation and vegetation develop- ment (Provoost et al. 2002). At present, rough grass and scrub encroachment result in a se- vere fragmentation within a matrix of dense dune vegetation (shrubs, dense grassland). In an earlier paper, we identified typical spider species for this habitat (Bonte et al. 2002). One of the most habitat-specific species is A/- opecosa fabrilis, although its overall Indicator Value was low, hence, indicating a relative low occupancy rate. Alopecosa fabrilis is used as model organ- ism in this study. It is the largest lycosid spe- cies, living in self-made burrows in dry sandy habitats from temperate regions in Europe (males: 10-12 mm; females: 13-16 mm; Rob- erts 1998). The species has a nocturnal life style in which especially males leave their burrow in search for mating partners during autumn, with a peak activity during Septem- ber (Roberts 1998; Bonte, unpub. data.). Fe- males are active during autumn, but especially during late winter and early spring. The spe- cies occurs in low densities, has a biannual life cycle (Bonte et al. unpub. data) and does not perform ballooning dispersal under labo- ratory conditions (Bonte et al. 2003b). During its period of activity, the spider 224 THE JOURNAL OF ARACHNOLOGY Figure 1. — Map of the surveyed grey dune habitat patches in the coastal dunes of De Panne-Bray Dunes, Belgium. Black: patches occupied by Alopecosa fahrilis', white: unoccupied patches. Circles in- dicate patches not occupied by Arachnospila rufci. hunting wasp (Hymenoptera, Pompilidae) Ar- achnospila rufa has been observed to be the main predator (Koomen & Peeters 1993; Pee- ters et al. 2004; Bonte, pers. obs.; Nieuwen- huijsen, pers. comm.). Alopecosa schmidti (Hahn 1834) has been recorded as prey in Middle-Europe (Schljachtenok 1996). Female pompilids provide each cell of their nest be- low the sand surface with only a single par- alyzed spider, on which one egg is laid. Ar- achnospila rufa is large (body size up to 18 mm) and common in sandy regions of Bel- gium and the Netherlands. It reaches adult- hood from June-October (Peeters et al. 2004; Nieuwenhuijsen in press). More detailed in- formation about its ecology and life history is unfortunately not available. Voucher speci- mens of both Alopecosa fabrilis and Arach- nospila rufa are deposited at the Royal Bel- gian Institute for Natural Sciences. Field Survey & Statistical Analyses. — Oc- cupancy patterns of A. fabrilis were recorded from 25 August-7 October 2003 using pitfall traps (diameter 9 cm, 6% formaldehyde-de- tergent solution). In each of the 52 grey dune patches, at least five traps were randomly placed, depending on the patch area. Patches were digitized from aerial orthophotographs with a Geographic Information System (Arc view 3.1) and discrimination of vegeta- tion types was based on vegetation-specific red (RED) and near-infrared (NIR) reflectance values (Provoost et al. 2002). Because of its high activity during this season, the use of pit- fall traps have been shown to be very useful in catching this typical spider species living within this habitat (Bonte et al. 2004c). The presence of the hunting wasp was recorded in the same pitfall traps, but completed with de- tailed field surveys during sunny days. Al- though not all observed specimens were col- lected to confirm identification, individuals could be identified in the field by their large size and distinct abdominal coloration. From GIS, we measured patch area, the distances to the nearest suitable patch, as a measure of patch isolation and distance to the nearest oc- cupied patch (nearest-neighbor distance) as BONTE & MAELFAIT— ASSOCIATION BETWEEN SPIDER WASP AND HOST 225 measurement of population isolation. For A. rufa, we did not use the latter isolation mea- surement because its presence was only re- corded during September, hence ignoring pos- sible different distribution patterns during the previous summer-period. Patch occupancy patterns for A. fabrilis (0: vacant; 1 : occupied) were analyzed by logistic models for binomial data and logit-link with backward elimination of the non-significant parameters (SAS 9.2). Patch area, patch iso- lation and the interaction between both were included as independent variables. For A. rufa, the occupancy status of the patch by A. fa- brilis was used as an additional categorical variable. RESULTS Patch occupancy of Alopecosa fabrilis was not significantly influenced by patch area (x? = 2.255; P = 0.133). The distance to the near- est suitable habitat (x? = 1.168; P = 0.280) and the interaction between both (x? = 0.208; P = 0.648) did not contribute significantly. If the nearest-neighbor distance was taken into account, a significant negative relationship was found between isolation and patch occu- pation (estimated slope: -0.035 ± 0.014; x? ^ 6.322; P = 0.012; Fig. 2). Other parameters remained non-significant. Although the model describes patch occupancy patterns in a sig- nificant way, 32.7% of the cases were mis- classified. Mainly patches near the inner dune front were occupied (Fig. 1). If only patches at the inner dune front were used for analysis, neither distance to the nearest patch (x? = 0.028; P = 0.865), distance to the nearest oc- cupied patch (xi == 1.144; P — 0.285) or area (Xi = 0.476; P = 0.490) explained distribution patterns. After backward elimination of non-signifi- cant parameters, the occupancy of Arachnos- pila rufa was only determined by the presence of Alopecosa fabrilis (x? — 34.41; P < 0.0001). No other lycosid species were re- corded during the survey period. Landscape geometry appears to be unimportant during this season since the explanatory power of both patch area and patch isolation are ex- tremely low (all variables and interactions: x^ < 1.29; P > 0.256). Patches occupied by A. fabrilis but not by A. rufa {n = 4) during this period were significantly smaller than those (n = 11) occupied by both species (Mann- Whit- ney U-test: Z = 2.74; P = 0.006; Fig. 2). Ar- achnospila rufa was not recorded in patches without A. fabrilis populations. DISCUSSION Population structure of Alopecosa fabril- is,— Our results demonstrate that the distance to the nearest occupied patch contributed sig- nificantly to the patch-occupancy model, whereas the overall habitat structure (i.e. ge- ometry and configuration of all available patches) did not. As a result, the population structure seems not to be affected by local po- tential population sizes (related to local pop- ulation viability; Hanski 1999) and distant dis- persal abilities by ballooning. The latter is in agreement with the absence of aerial dispersal under laboratory conditions (Bonte et al. 2003b). As only the nearest-neighbor distance contributed significantly to the patch-occupan- cy model, the population structure of A. fa- brilis seems to be clearly aggregative, indi- cating low abilities to disperse (and penetrate) cursorially through the hostile matrix or a pre- dominant role of habitat quality. Our data, however, only result from a short term survey. But preliminary data on patch occupancy pat- terns from 2004 show similar distribution pat- terns (Bonte et al. unpub. results), indicating the absence of strong turn-over dynamics at extended, though still rather short time spans. The fact that patch area (related to local pop- ulation size) does not explain occupancy patterns seems to indicate that the spatial pop- ulation structure is more likely to be deter- mined by random extinction-colonization dy- namics within aggregative clusters of closely connected suitable habitat remnants, which appear to be situated near the inner dune front. As the interaction between patch size and patch isolation does not explain occupancy patterns, a rescue effect resulting from source- sink dynamics is unlikely (Dias & Blondel 1996). The aggregated pattern may result from dis- persal limitation after historical population dy- namics or from spatially correlated habitat characteristics, related to habitat quality, as experienced by the species. Although both are possible, the second seems to be more prob- able as especially patches near the inner dune front were occupied. In coastal dunes, char- acterized by a successive ontogenesis, grey dunes, although structurally similar may show 226 THE JOURNAL OF ARACHNOLOGY Figure 2. — Relation between patch area, nearest-neighbor distance and occupancy incidence for A, fabrilis populations. Arrows indicate patches occupied by A. fabrilis, but not by its predator A. rufa. extensive spatial correlation in environmental characteristics related to microclimate, soil properties and aeolic dynamics (sand over- blowing). The fact that considerable non-spa- tial correlated variation in habitat quality af- fects spatial structural patterns has already been demonstrated by e.g. Briers & Warren (2000) and Bonte et al. (2003a). For a fosso- rial species like A. fabrilis, it is not impossible that e.g., grain size properties, soil water re- tention, soil formation and the intensity of sand overblowing strongly determine burrow- ing ability and the sustainability of created burrows. In the same coastal dune landscape, the grassland inhabiting wolf spider P. mon- ticola shows a habitat-quality dependent pop- ulation structure in which both patch area and patch connectivity determined occupancy pat- terns (Bonte et ah 2003a), although not in an aggregative way. As vegetation structure is alike in ail investigated grey dune patches, similarity in habitat quality for A. fabrilis is probably more related to pedological (possibly soil stability or differences in grain size near the inner dune front) or microclimatological (less buffered temperatures far from the sea) conditions and, possibly, the reason for the observed clustered occurrence. Also, the ab- sence of A. fabrilis in large patches more close to the sea, in which large populations should be able to persist, provides more likely evi- dence for a reaction towards spatially corre- lated environmental properties and optimal habitat quality in older grey dunes, situated near the inner dune front. In this case, dis- persal limitation is of inferior importance, as illustrated by the model in which only patches near the inner dune front were included. Host-parasitoid association^^Spatially explicit host-parasitoid metapopulation mod- els generate spatially correlated abundance and, hence, occupancy patterns in host and parasitoid (Hanski, 1999). Arachnopsila rufa and related spider hunting wasps, however, appear to be more generalist parasitoids on (larger) lycosid species (Koomen & Peeters 1993; Endo & Endo 1994; Finch 1997; Pee- ters et al. 2004), The almost complete match between occupancy incidences of host and parasitoid indicate that, at least in the Flemish coastal dunes and during the survey period, A. rufa behaves as a specialized parasitoid on A. BONTE & MAELFAIT— ASSOCIATION BETWEEN SPIDER WASP AND HOST 227 fabrilis. This indicates that at least one species of pompilid wasp may behave as a host-spe- cialist during some periods of its adult life. According to White et aL (1996), spatial pop- ulation patterns in which both host and para- sitoid occur in the same patches are only static in case of low host dispersal and low parasit- oid survival. Certainly the first assumption seems to be fulfilled in our study. However, A. rufa is already adult earlier in the season, and should, as a result, behave more flexibly in prey selection. As predicted by optimal for- aging theory, prey size selection should be op- timized in relation to the species’ load-carry- ing capacity (Evans 1962; Coelho & Ladage 1999). The latter authors indeed found that large female wasps searched for prey that matched their owe body mass and lift capac- ity. As a result, A. rufa should prey on simi- larly large-bodied lycosid species or show be- havioural flexibility according to the available prey spectrum. Taking into account the need for sandy habitats for prey burrowing, A. rufa seems to be restricted to dynamic dune habi- tats in which soil development is poor and bare sand sufficiently available. Female A. fa- brilis disappear from the population in late winter and early spring and are not available as host during the early beginning of the hunt- er wasp’s adulthood (Bonte, unpub. data). Other large-bodied lycosid species, occurring in these habitats include only Alopecosa cu- neata (Clerck 1757) and A. barbipes (Suedev- all 1833) (Bonte et al. 2002), also having a spring activity, and are as a result neither available during summer. Possibly, A. rufa uses smaller lycosids as host during early summer {Arctosa perita (Latreille 1799) or subadult A. fabrilis), potentially resulting in different spatial distribution patterns. The fact that A. rufa was only absent from small patches, inhabited by A. fabrilis seems to provide evidence that higher trophic levels require larger habitats (Holt 1996; Holt et al. 1999), but may also be an artifact in our sam- pling strategy if smaller patches are only ac- cidentally visited by the spider wasp, hence reducing encounter chances in the field. Ac- cording to Holt (1999), higher trophic ranks need larger areas if dispersal is limited, have closed populations and show a high degree of specialization towards resource species. Cer- tainly the latter seems to hold for the (at least temporal and local) A. rufa~A. fabrilis asso- ciation. Assuming that our findings are not due to sampling artifacts, parasitoid absence in small populations indicates low dispersal abilities or at least low dispersal motivation in spider wasp within spatially structured suit- able habitat surrounded by hostile matrix veg- etation. Also, since both juvenile and adult A. fa- brilis stay in their burrow during day, prey selection by A. rufa has probably to rely on olfactory and not on visual cues because bur- row entrances are not visible, as also observed in spider hunting wasps preying on day-active wolf spiders (e.g., Pompilus cinereus (Fabri- cius 1775); Bonte pers. obs.) and in sphecid wasps (although only at short distance after visual detection; Anton & Gnatzy 1998). In fragmented coastal dune habitats, host and parasitoid show, at least temporally and locally, a similar aggregative spatial popula- tion structure, indicating the importance of (lack of) dispersal and possibly spatially cor- related habitat characteristics in structuring patch occupancy patterns. Additionally, the spider hunting wasp A. rufa appears to behave as a specialized parasitoid during the main ac- tivity period of its optimal prey with restricted dispersal motivation. During this period, the presence of an assumed generalist parasitoid is a good predictor for the presence of noc- turnal and burrowing dune wolf spider in larg- er habitat patches in the Flemish coastal dunes. ACKNOWLEDGMENTS The author would like to thank the Ministry of the Flemish Community (Aminal, Afdeling Natuur) and the Conseil General du Nord for granting access to their field sites. The senior author is a Postdoctoral Fellow of the Fund for Scientific Research — Flanders (Belgium) (F.W.O. — Vlaanderen). One anonymous ref- eree gave very valuable suggestions with re- spect to species-area relationships and trophic rank, opening new insights for further re- search. LITERATURE CITED Anton, S. & W. Gnatzy. 1998. Prey specificity and the importance of close-range chemical cues in prey recognition in the Digger Wasp, Liris niger. Journal of Insect Behavior 1 1:671-690, Bergman, K.-O. 1999. Habitat utilization by Lopin- ga achine (Nymphalidae:Satyrinae) larvae and 228 THE JOURNAL OF ARACHNOLOGY ovipositing females: implications for conserva- tion. Biological Conservation 88:69-74. Bonte, D., L. Baert, & J.-R Maelfait. 2002. Spider assemblage structure and stability in a hetero- geneous coastal dune system (Belgium). Journal of Arachnology 30:331-343. Bonte, D., L., Lens, J.-P., Maelfait, M. Hoffmann, & E. Kuijken. 2003a. Patch quality and connec- tivity influence spatial dynamics in a dune wolf- spider. Oecologia 135:227-233. Bonte, D., N., Vandenbroecke, L. Lens, and J.-P. Maelfait, 2003b. Low propensity for aerial dis- persal in specialist spiders from fragmented land- scapes. Proceedings of the Royal Society London Series B 270:1601-1607. Bonte, D., L., Baert, L., Lens, & J.-R Maelfait. 2004a. Effects of aerial dispersal, habitat spe- cialisation, and landscape structure on spider dis- tribution across fragmented grey dunes. Ecogra- phy 27:343-349. Bonte, D., L. Lens, & J.-P. Maelfait. 2004b. Lack of homeward orientation and increased mobility result in high emigration rates from low-quality fragments in a dune wolf spider. Journal of An- imal Ecology 73:643-650. Bonte, D., J.-P. Maelfait, & L. Baert. 2004c. Deter- mining spider species richness in fragmented coastal dune habitats by extrapolation and esti- mation (Araneae). Pp. 189 — 197. In Samu, F. & Szinetar, Cs. (eds.). European Arachnology 2002. Boulton, A.M., M.G. Ramirez, & C.R Blair. 1998. Genetic structure in a coastal dune spider (Geo- lycosa pikei) on Long Island, New York Barrier Islands. Biological Journal of the Linnean Soci- ety 64:69-82. Briers R.A. & PH. Warren. 2000. Population turn- over and habitat dynamics in Notonecta (Hemip- tera:Notonectidae) metapopulations. Oecologia 123:216-222. Coelho, J.R. & L.D. Ladage. 1999. Foraging ca- pacity of the great golden digger wasp Sphex ich~ neumoneus. Ecological Entomology 24:480- 483. Comins, H.N., M.P Hassel, & R.M. May. 1992. The spatial dynamics of host-parasitoid systems. Journal of Animal Ecology 61:735-748. Dias, PC. & J. Blonderl. 1996. Local speciation and maladaptations in the Mediterranean blue tit (Pa- rus caeruleus). Oecologia 107:79-86. Endo, T. & A. Endo, 1994. Prey selection by a spi- der wasp, Batozonellus lacerticida (Hymenop- tera: Pompilidae): effects of seasonal variation in prey species, size and density. Ecological Re- search 9:225-235. Evans, H.E. 1962. The evolution of prey-carrying mechanisms in wasps. Evolution 16:468-483. Finch, O.-D. 1997. Spider wasps (Hymenoptera, Pompilidae) as predators of a spider taxocoenos- is. Proceedings of the 16* European Colloquium of Arachnology, Siedlce. Pp. 83-89. Gurdebeke, S., B. Neirynck, & J.-R Maelfait. 2000. Population genetic effects of forest fragmenta- tion in Flanders (Belgium) on Coelotes terrestris (Araneae:Agelenidae) as revealed by allozymes and RAPD. Ekologia (Bratislava) 19:87-96. Hanski, I. 1999. Metapopulation ecology. Oxford University Press Inc., New York. 313 pp. Harrison, S. 1991. Local extinction in a metapo- pulation context: an empirical evaluation. Bio- logical Journal of the Linnaean Society 42:73- 88. Hassel, M.P, H.N. Comings, & R.M. May. 1991. Spatial structure and chaos in insect population dynamics. Nature 353:255-258. Holt, R.D. 1996. Food webs in space: an island bio- geographical perspective. Pp 313-323. In Polis, G.A. & Martinez, N.D. (eds.): Food Webs-inte- gration of Patterns and Dynamics, Chapman & Hall, New York. Holt, R.D., J.H., Lawton, G.A. Polis, & N.D. Mar- tinez 1999. Trophic rank and the species-area re- lationships. Ecology 80:1495-1504. Koomen, R & TM.J. Peelers. 1993. New prey re- cords for spider hunting wasps (Hymenoptera: Pompilidae). Memoires of the Queeensland Mu- seum 33:571-574. Kreiter, N.A. & D.H. Wise. 2001. Prey availability limits fecundity and influences the movement pattern of female fishing spiders. Oecologia 127: 417-424. Moilanen, A. & 1. Hanski. 1998. Metapopulation dynamics: effect of habitat quality and landscape structure. Ecology 79:2503-2515. Nieuwenhuijsen, H. In press. Determinatietabellen voor de Nederlandse spinnendoders (Hymenop- tera, Pompilidae). Alkmaar. Peeters, TM.J,, C., van Van achterberg, W.R.B, Heitmans, W.E, Klein, V,, Lefeber, A.J., van Loon, A. A,, Mabelis, H., Nieuwenhuijsen, M., Reemer, J., de Rond, J. Smit, & H.H.W. Velthuis. 2004. De wespen en mieren van Nederland (Hy- menoptera, Aculeata). Nederlandse Fauna 6. Na- tionaal Natuurhistorisch Museum Naturalis, Lei- den, K.N.N.V. uitgeverij, Utrecht & European Invertebrate Survey, Nederland, Leiden, 507 pp. Peterson, M.A., R.F Denno, & L. Robinson. 2001. Apparent widespread gene flow in the predomi- nantly flightless planthopper Tumidagena minu- ta. Ecological Entomology 26:629-631 . Provoost, S., C., Ampe, D., Bonte, E. Cosyns, & M. Hoffmann. 2002. Ecology, management and monitoring of dune grasslands in Flanders, Bel- gium. Pp. 11-22. In EUROCOAST (ed,): Littoral 2002, The Changing Coast. Eurocoast/EUCC, Porto, Portugal. Provoost, S, & D. Bonte (eds.). (2004). Levende duinen. Een overzicht van de biodiversiteit in de BONTE & MAELFAIT— ASSOCIATION BETWEEN SPIDER WASP AND HOST 229 Vlaamse kuststreek. Mededelingen van het Insti- tuut voor Natuurbehoud, Brussels, 420 pp. Ramirez, M.G. & L.B. Fandino, 1996. Genetic var- iability and gene flow in Metepeira ventura {At- aneae, Araneidae). The Journal of Arachnology 24:1-8. Ramirez, M.G. & K.E. Haakonsen. 1999. Gene flow among habitat patches on a fragmented land- scape in the spider Argiope trifasciata (Araneae: Araneidae). Heredity 83:580-585. Ravenscroft, N.O.M. 1994. Environmental influenc- es on ro.ate location in male chequered skipper butterflies, Carterocephalus palaemon (Lepidop- tera:Hesperiidae). Animal Behaviour 47:1179- 1187. Roberts, MJ. 1998. Spiders of Britain and Northern Europe. Harper Collins Publishers, London, 383 pp. Same, E, A. Sziranyi, & B. Kiss, 2003. Foraging in agricultural fields: local ‘sit-and-move’ strat- egy scales up to risk-averse habitat use in a wolf spider. Animal behaviour 66:939-947. Schljachtenok, A.S. 1996. Uber die Beutespinnen (Araeei) einiger Wegwespen (Hym,, Pompili- dae). Entomologische Nachrichten und Berichte 40(3): 169-172. Thomas, C.D., J.A. Thomas, & M.S. Warren. 1992. Distribution of occupied and vacant habitats in fragmented landscapes. Oecologia 92:563-567, White, A., M. Begon, & R.G. Bowers. 1996. Host- pathogen systems in a spatially patchy environ- ment. Proceedings of the Royal Society London B 263:325-332. Manuscript received 16 December 2004, revised 17 June 2005. 2005. The Journal of Arachnology 33:230-235 EARLY SUCCESSION OF A BOREAL SPIDER COMMUNITY AFTER FOREST FIRE Seppo Kopoeen: Zoological Museum, University of Turku, FT20014 Turku, Finland. E-mail: sepkopo@utu.fi I ABSTRACT. Ground-living spiders were studied, using pitfall traps, 3-4 months after a wildfire, and then during three post-fire summers. The study area was a pine (Pinus sylvestris) forest in southwestern Finland. Lycosidae dominated in individual numbers at the burned site and Linyphiidae at the control. In species numbers, Linyphiidae dominated at both sites, and Lycosidae, Gnaphosidae and Theridiidae were more species-rich at the burned than control site. The lycosid Xerolycosa nemoralis was dominant at the burned site, and the linyphiid Agyneta cauta at the control. Abundant species found only at the burned site included Xerolycosa nemoralis, Pardosa riparia, Acantholycosa lignaria and Micaria silesiaca. Tap- inocyba pattern and Pardosa lugubris occurred at both sites in large numbers. A slight positive effect of fire on the species richness was found. Species with more or less stable abundance at the burned site during the study period included Pardosa riparia, P. lugubris and Diplostyla concolor. Increasing abun- dance in successive years occurred for Acantholycosa lignaria, Micaria silesiaca, Xerolycosa nemoralis and for the family Lycosidae. Euryopis flavomaculata, Agyneta rurestris, Tapinocyba pattens and the family Linyphiidae showed a decreasing abundance during the study years. The spider community at the burned site remained clearly different compared to the control during three post-fire summers, primarily caused by the abundance of Gnaphosidae and Lycosidae. Keywords: Araneae, ground-living spiders, post-fire succession, pine forest, Finland Intensive and regular fires are a natural part of the ecology in many areas and have an ef- fect on the fauna in those habitats. This is well-known in the Mediterranean-type of eco- systems (e.g., Stamou 1998; Moretti et al. 2004). But in the boreal taiga forest zone of the Holarctic, fires occur normally at long in- tervals, and the fauna living there is less adapted to the fires. Forest fires are rare and small in Finland, mainly due to active fire control. The situation contrasts clearly with that in the boreal conif- erous forest zone both in Russia and Canada where extensive areas of forest are yearly de- stroyed by fire (e.g., Kopoeen 1993). There- fore, little information is available on the ef- fects of forest fire on spiders and their post-fire succession in Finland or in the whole of Fennoscandia. In Finland, Huhta (1971) studied succession after prescribed burning, and Koponen (1988, 1989, 1995) studied the effects of natural fire in a subarctic birch woodland in Finnish Lapland. Some data on the first post-fire summer at the present study site have been given by Kopoeen (2004), Hauge & Kvamme (1983) studied spiders from forest fire areas in Norway. In the pre- sent paper, the post-fire succession of ground- living spiders in a boreal forest in southwest- ern Finland is explored. METHODS The study area is situated in Tammela, Riih- ivalkama, east of the Torronsuo National Park (Finnish Grid 27°E: 6740:323); ca. 60°44'N, 23°45'E. The study site is a dry gentle slope with young pine (Pinus sylvestris) trees, di- ameter 20 cm or less. The forest (about 150 hectares) was burned on 9-10 June 1997. It was totally burned: all moss and lichen as well as vascular plant veg- etation was destroyed. The dead pines were still standing there in autumn 1997, but they were cut down and removed in May 1998. Under the 2-5 mm thick layer of ash and char- coal there was a humus layer, but locally only mineral soil. The black and open site was sun- ny, dry and warm, especially during the first post-fire year. The distance from non-burned forests to the study site was at least 150 m. Ground living spiders were studied 3-4 months after the fire, in order to find coloniz- ers or species which had survived the fire. Twenty-four pitfall traps with ethylene glycol 230 KOPONEN— SPIDER SUCCESSION AFTER FOREST FIRE 231 Table 1. — Ground and field layer vegetation around the traps at burned and control sites in Tammela, Finland. New plant species appearing at different stages of the succession are indicated by +. Burned site Coverage % Plant species 1997 September <1% - 1998 May <1% - July 10% Epilobium angustifolium, Ceratodon moss September 20% 1999 May 30% ELuzula + Calluna vulgaris July 50% September 55% 2000 May 60% +Deschampsia August 80% +Rubus idaeus Control site 100% Pleurozium and Dicranum moss, Linnaea borealis, Trientalis eu- ropea, Vaccinium vitis-idaea, V. myrtillus etc, and Pinus syl- vestris and detergent (mouth diameter 60 mm, with covers) were placed there from 1 2 September- 17 October 1997. During the following three summers, 10 similar traps were placed in the burned site and 10 in a control site about 300 m from the fire. Coverage of the ground and field layer vegetation was estimated visually around the traps (Table 1). There is no infor- mation available of the previous fire history of the study area. The climatic conditions var- ied during the study years; the summer 1998 was cool and rainy, 1999 was warm, and 2000 near the average. The yearly study periods were: 9 May-27 September 1998; 15 May-28 September 1999; 14 May-10 August 2000. We removed the traps in August 2000 due to interference by people visiting the site. The spider mate- rial, deposited on the Zoological Museum, University of Turku, consisted of about 1100 identifiable specimens from the burned and 1540 from the control site. Nomenclature is mainly after Platnick (2004), except Agyneta! Meioneta. RESULTS Altogether 91 species of ground-living spi- ders were found, 70 at the burned site and 59 at the unburned control site. The family Lin- yphiidae clearly dominated in species num- bers. Post-fire autumn. — The spiders were trapped 3-4 months after the fire in autumn 1997. Altogether, 16 species were found dur- ing this short, autumnal collecting period. Tapinocyba pallens (O.R-Cambridge 1872) clearly dominated (25.0%), and Tenuiphantes mengei (Kulczynski 1887) (13.6%) and Agy- neta rurestris (C.L. Koch 1836) (11.4%) were also abundant. Agroeca proxima (O.R-Cam- bridge 1871), Pardosa lugubris (Walckenaer 1802), Trochosa terricola Thorell 1856, Por- rhomma pallidum Jackson 1913, Gnaphosa bicolor (Hahn 1833) and Haplodrassus sig- nifer (C.L. Koch 1839) were represented by at least two specimens. Linyphiids were the dominant spider family in terms of individuals (Table 2) and half of the species caught were linyphiids. For a more detailed description of the results from the autumn 1997, see Kopo- nen (2004). Following summers.— The general com- position of the spider assemblages at the burned and control sites is shown in Table 2. Linyphiids dominated in the number of spe- cies, at both the control site as well as at the burned site, during the study years. Species numbers of Lycosidae, Gnaphosidae and Theridiidae were higher at the burned than the control site. Faunal similarity, as percentage of species found at both sites, in 1998, 1999 and 2000 was 28%, 35% and 25% respective- ly. The situation in 2000 was somewhat biased by the destroyed traps at the burned site (see above). There were no great differences in the yearly species richness between the sites; however, during the 1998 and 1999 summers, more species were found at the burned site: 39 vs. 35 and 51 vs. 46 respectively (Table 2). The family Linyphiidae was dominant in individual numbers at the control site (69.6- 86.7%) during the whole study period, and 232 THE JOURNAL OF ARACHNOLOGY Table 2. — Family composition of individuals (%) and species (no. of species) of the spider fauna at the burned (bu) and control (co) sites, 1997-2000. Trapping periods: 12 September- 17 October 1997, 9 May- 27 September 1998, 15 May-28 September 1999, 14 May- 10 August 2000. 1997 1998 1999 2000 bu bu CO bu CO bu CO Individuals Linyphiidae 75.4 59.4 86.7 29.0 82.9 15.9 69.6 Lycosidae 10.8 32.5 2.3 64.8 8.5 62.0 18.7 Gnaphosidae 6.2 2.2 LI 3.1 0.7 20.9 3.0 Theridiidae 1.5 4.3 1.4 2.1 1.4 0.9 0.4 Others 6.2 1.6 8.5 1.0 6.5 0.3 8.3 Species Linyphiidae 8 21 22 22 29 9 24 Lycosidae 3 7 2 8 4 6 3 Gnaphosidae 2 4 2 7 2 6 4 Theridiidae 1 3 1 4 1 4 2 Others 2 4 8 10 10 5 12 Total 16 39 35 51 46 30 45 also at the burned site during early succession (1997-98) while Lycosidae dominated at the burned site during the two following summers (1999-2000: 62.0-64.8%) (Table 2). No trend was found in catches (#individuals/trap/day) between the sites; in 1998 and 2000 more specimens were caught at the burned and in 1999 at the control site (Table 2). The most abundant species at both sites, 1998-2000, are shown in Tables 3-4. The ly- cosid Xerolycosa nemoralis (Westring 1861) was the dominant species at the burned site during the whole period (Table 3). Pardosa riparia (C.L. Koch 1833), P. lugubris and A/- opecosa pulverulenta (Clerck 1957) (Lycosi- dae), and Diplostyla concolor (Wider 1834) (Linyphiidae) were also abundant, 1998- 2000. Euryopis flavomaculata (C.L. Koch 1836) and Agyneta rurestris were abundant during the first post- fire summer (1998) but later they were less numerous. Tapinocyba pallens had a similar but less clear trend. On the other hand, Micaria silesiaca L. Koch 1875 and Acantholycosa lignaria (Clerck 1757) were trapped in good numbers during the latter summers (1999-2000). At the control site, the composition of abundant linyphiid species was rather stable during the study years (1998-2000), see Table 4. In contrast to the burned site with lycosids dominating, here linyphiids were most nu- merous. They were represented primarily by the species Agyneta cauta (O.R-Cambridge 1902), Tapinocyba pallens, Centromerus ar- canus (O.R-Cambridge 1873) and Agyneta conigera (O.R-Cambridge 1873). These were followed, based on total abundance, by the ly- cosids Pardosa lugubris, which was, however, caught in low numbers during the first sum- mer, and Alopecosa aculeata (Clerck 1757). Typical species at the control site were also the linyphiids Diplocentria bidentata (Emer- ton 1 882), Walckenaeria antica (Wider 1 834), W. cucullata (C.L. Koch 1836), Minyriolus pusillus (Wider 1834), Tenuiphantes alacris (Blackwall 1853), T. tenebricola (Wider 1834), and Bathyphantes parvulus (Westring 1851). From other families, Zora nemoralis (Blackwall 1861), Z. spinimana (Sundevall 1833) , Haplodrassus soerenseni (Strand 1900) and Cryphoeca silvicola (C.L. Koch 1834) can be mentioned. Abundant species found only at the burned site included Xerolycosa nemoralis, Pardosa riparia, Acantholycosa lignaria, Micaria si- lesiaca, Phrurolithus festivus (C.L. Koch 1835) and Agyneta rurestris', and Tapinocyba pallens and Pardosa lugubris were found at both sites. Faunistically interesting species at the burned site included Agyneta gulosa (L. Koch 1869) (a northern species in Finland), and Troxochrota scabra Kulczynski 1 894 and Troxochrus nasutus Schenkel 1925 (rare, mainly southern species). KOPONEN— SPIDER SUCCESSION AFTER FOREST FIRE 233 Table 3. — The most abundant spiders trapped at the burned site, 1998-2000. Number of individuals (n), percentage and rank (only for the 10 most abundant in each year) are given. 1998 1999 2000 n % Rank n % Rank n % Rank Xerolycosa nemoralis Tapinocyba pallens Alopecosa pulverulenta Pardosa riparia Euryopis flavomaculata Agyneta rurestris Pardosa lugubris Centromerus arcanus Diplostyla concolor Tenuiphantes mengei Micaria silesiaca Walckenaeria antica Acantholycosa lignaria Gnaphosa bicolor Phrurolithus festivus 29 11.9 (1) 27 11.2 (2) 25 10.3 (3) 24 9.9 (4) 21 8.7 (5) 16 6.6 (6) 15 6.2 (7) 9 3.7 (8) 8 3.3 (9) 7 2.9 (10) 2 0.8 2 0.8 92 20.4 (1) 36 8.0 (4) 36 8.0 (4) 71 15.8 (2) 8 1.8 9 2.0 (10) 21 4.7 (6) 2 0.4 42 9.3 (3) 10 2.2 (9) 11 2.4 (7) 11 2.4 (7) 4 0.9 1 0.2 2 0.4 103 36.8 (1) 2 0.5 4 1.1 (10) 34 12.1 (2) 1 0.3 3 0.8 19 6.8 (4) 10 3.6 (6) 26 9.3 (3) 9 2.4 (7) 18 6.4 (5) 8 2.1 (8) 7 1.9 (9) DISCUSSION The number of species caught 3-4 months after the fire was rather high, although the trapping period was short (see also Koponen 2004). This can, at least partly, be a result of the number of traps in 1997 (24 vs. 10 in fol- lowing summers). Whether some of the spe- cies had survived the fire is unknown. Some stationary invertebrates (gastropods, milli- pedes, female coccids) were also caught in the traps during the autumn of 1997. These seem to have survived under large stones or in the soil (see also Punttila et al. 1994), the same may be true for some spiders. On the other hand, silk lines were seen in great numbers on the burned ground indicating ballooning. Pioneer species at the burned site are Agy- neta species (subgenus Meioneta), e.g. A. ru- restris, which was caught in highest numbers among the pioneer species, as well as Erigone atra Blackwall 1833 and Oedothorax retusus (Westring 1851) (cf. Merrett 1976; Winter et al. 1983; Koponen & Niemela 1994). None of them was found at the control site. The dominant species at the burned site, Xerolycosa nemoralis, has been found in Fin- land as a colonizer of open, dry and warm areas, often human-influenced. These areas in- clude dried peat bogs (Koponen 1979) and heavily polluted areas (Koponen & Niemela 1994). In a study of burned pine forests in northern Germany, Schaefer (1980) found X. nemoralis in high numbers in young pine plantations but not at the burned sites, where Pardosa lugubris dominated among lycosids. Pardosa lugubris was one of the most abun- dant species both at the burned and control site in Tammela. Pardosa riparia, Acantholycosa lignaria, Micaria silesiaca and Phrurolithus festivus, species found only at the burned site, have often been caught in open areas (e.g., Hanggi et al. 1995; Marusik et al. 2004). Species pre- ferring open and warm areas, like many ly- cosids, have often been caught in high num- bers at burned localities (Brabetz 1978; Schaefer 1980; Koponen 1993, 2004; Buddie et al. 2000). Niwa & Peck (2002) studied the influence of prescribed fire on spiders in co- nifer stands in Oregon and found, in agree- ment with the present study, that Lycosidae and Gnaphosidae were more numerous at burned and Linyphiidae at unburned sites. A slight positive effect of fire on the species richness could be seen (cf. also Moretti et al. 2004). The species-rich fauna at the burned site was a combination of pioneer species (e.g., Agyneta rurestris, Oedothorax retusus), of thermophilous {Xerolycosa nemoralis, Mi- caria silesiaca) and eurytopic {Diplostyla concolor) species often preferring open sites, and of some typical pine forest species {Tap- inocyba pallens, Centromerus arcanus). On the other hand, both species and individual 234 THE JOURNAL OF ARACHNOLOGY Table 4. — The most abundant spiders trapped at the control site, 1998-2000. Number of individuals (n), percentage and rank (only for the 10 most abundant in each year) are given. 1998 1999 2000 n % Rank n % Rank n % Rank Centromerus arcanus 44 16.1 (1) 70 8.8 (4) 21 5.6 (6) Tapinocyba pallens 43 15.8 (2) 77 9.7 (3) 22 5.9 (5) Agyneta cauta 20 7.3 (3) 158 20.0 (1) 56 14.9 (1) A. conigera 17 6.2 (4) 86 10.9 (2) 24 6.4 (4) Diplocentria bidentata 13 4.8 (5) 10 1.3 11 2.3 (8) Walckenaeria antica 13 4.8 (5) 18 2.3 6 1.3 Diplostyla concolor 11 4.0 (7) 20 2.5 (7) 8 1.7 (10) Zora nemoralis 11 4.0 (7) 4 0.5 2 0.4 Bathyphantes parvulus 9 3.3 (9) 9 LI 1 0.2 Tenuiphantes tenebricola 9 3.3 (9) 18 2.3 6 1.3 Pardosa lugubris 4 1.5 44 5.6 (5) 51 13.6 (2) Alopecosa aculeata 8 2.9 15 1.9 29 7.7 (3) Minyriolus pusillus 7 3.6 30 3.8 (6) 8 1.7 (10) Walckenaeria cucullata 8 2.6 19 2.4 (8) 8 1.7 (10) Tenuiphantes alacris 5 1.8 19 2.4 (8) 6 1.3 Cryphoeca silvicola 5 1.8 19 2.4 (8) 4 0.8 Zora spinimana 1 0.4 — 12 2.5 (7) Haplodrassus soerenseni 2 0.7 5 0.6 11 2.3 (8) numbers outside of the four main families (Linyphiidae, Lycosidae, Gnaphosidae and Theridiidae) were higher at the control than at the burned site (see Table 2), indicating a gen- erally more diverse fauna in the unburned for- est. Some general trends in the spider assem- blage at the burned site could be found during the study years. Species with more or less sta- ble abundance at the burned site include Par- dosa riparia, P. lugubris and Diplostyla con- color. Increasing abundance in successive years was true for Acantholycosa lignaria, Mi- caria silesiaca, Xerolycosa nemoralis and for the family Lycosidae. On the other hand, Eur- yopis flavomaculata, Agyneta rurestris, Tapi- nocyba pallens and the whole Linyphiidae showed decreasing numbers during the study years. The spider community at the burned site remained clearly different compared with the control during the study period’s three post-fire summers. This was primarily caused by the species diversity and abundance of Gnaphosidae and Lycosidae. LITERATURE CITED Brabetz, R. 1978. Auswirkungen des kontrollierten Brennes auf Spinnen und Scheecken einer Brachflache bei Rothenbuch im Hochspessart. Ein Beitrag zur Kenntnis der Spinnenfauna des Rhein-Mai-Gebietes. Courier Forschungsinstitut Senckenberg 29:1-124. Buddie, C.M., J.R. Spence & D.W. Langor. 2000. Succession of boreal forest spider assemblages following wildfire and harvesting. Ecography 23: 424-436. Hanggi, A., E. Stockli & W. Nentwig. 1995. Hab- itats of Central European spiders. Miscellanea Faunistica Helvetiae 4: 1-460. Hauge, E. & T. Kvamme. 1983. Spiders from for- est-fire areas in southeast Norway. Fauna Norv- egica, Series B 30:39-45. Huhta, V. 1971. Succession in the spider commu- nities of the forest floor after clear-cutting and prescribed burning. Annales Zoologici Fennici 8: 483-542. Koponen, S. 1979. Differences of spider fauna in natural and man-made habitats in a raised bog. The National Swedish Environment Protection Board, Report PM 1151:104-108. Koponen, S. 1988. Effect of fire on spider fauna in subarctic birch forest, northern Finland, Pp. 148 — 152. In XL Europaisches Arachnologisch- es Colloquium, Berlin 1988 (J. Haupt ed.). Technische Universitat Berlin, Dokumentation Kongresse und Tagungen 38. Koponen, S. 1989. Effect of fire on ground layer invertebrate fauna in birch forest in the Kevo Strict Nature Reserve, Finnish Lapland. Folia Forestalia 763:75-80. Koponen, S. 1993. Ground-living spiders (Araneae) one year after fire in three subarctic forest types. KOPONEN— SPIDER SUCCESSION AFTER FOREST FIRE 235 Quebec (Canada). Memoirs of the Queensland Museum 33:575-578. Koponen, S. 1995. Postfire succession of soil ar= thropod groups in a subarctic birch forest. Acta Zoologica Feenica 196:243-245. Koponen, S. 2004. Effects of intensive fire on the ground-living spider fauna of a pine forest (Ar- aneae). European Arachnology 2003, Arthropoda Selecta, Special Issue 1:133-137. Koponen, S. & P. Niemela. 1994. Ground-living ar- thropods along pollution gradient in boreal pine forest. Entomologica Fennica 6:128-131. Marasik, Yu.M., G.N. Azarkina & S. Koponen. 2004. A survey of East Palaearctic Lycosidae (Aranei). 11. Genus Acantholycosa F. Dahl, 1908 and related new genera. Arthropoda Selecta 12(2):101-148. Merrett, P. 1976. Changes in the ground-living spi- der fauna after heathland fires in Dorset. Bulletin of the British Arachnological Society 3(8):214- 221. Moretti, M., M.K. Obrist & P. Duelli. 2004. Ar- thropod biodiversity after forest fires: winners and losers in the winter fire regime of the south- ern Alps. Ecography 27:173-186. Niwa, C.G. & R.W. Peck. 2002. Influence of pre- scribed fire on carabid beetle (Carabidae) and spider (Araneae) assemblages in forest litter in southwestern Oregon. Environmental Entomolo- gy 31(5):785-796. Plateick, N.L 2004. The world spider catalog, ver- sion 5.0. American Museum of Natural History, online at http://research.amnh.org/entomology/ spiders/cataiog/index.html Punttila, R, S. Koponen & M. Saaristo. 1994. Col- onisation of a burned mountain-birch forest by ants (Hymenoptera, Formicidae) in subarctic Fin- land. Memorabilia Zoologica 48:193-206. Schaefer, M. 1980. Sukzession von Arthropoden in verbrannten Kieferforsten II. Spinnen (Araneida) und Weberknechte (Opilionida). Forstwissen- schaftliches Centralblatt 99:341-356. Stamou, G.P. 1998. Arthropods of Mediterranean- type ecosystems. Springer Verlag. Berlin Heidel- berg. 141 pp. Winter, K., R Diiweke, M. Schaefer & J. Schauer- mann. 1983. Sukzession von Arthropoden in ver- brannten Kieferforsten der Siidheide, Verhadlun- gee der Gesellschaft fur Okologie (Mainz 1981) 10:57-61. Manuscript received 20 December 2004, revised 15 June 2005. 2005. The Journal of Arachnology 33:236-242 ARE SALT MARSH INVASIONS BY THE GRASS ELYMUS ATHERICUS A THREAT FOR TWO DOMINANT HALOPHILIC WOLF SPIDERS? Julien Petillon^’^, Frederic Ysnel\ Jean-Claude Lefeuvre^ and Alain Canard^: ^ E.R.T. “Biodiversite fonctionnelle et Gestion des territoires”, Universite de Rennes 1, 263 Av, du General Leclerc, CS 74205, 35042 Rennes Cedex, France. E-mail: julien. petillon@univ-rennesl.fr; ^ U.M.R. C.N.R.S. “Fonctionnement des Ecosystemes et Biologie de la Conservation”, Universite de Rennes 1, 263 Av. du General Leclerc, CS 74205, 35042 Rennes Cedex, France; ^ Laboratoire d’ Evolution des Systemes Naturels et Modifies, Museum National d’Histoire Naturelle, 36 rue Geoffroy Saint Hilaire 75005 Paris, France ABSTRACT. As a result of the Elymus athericus (Poaceae) invasion in the last ten yeai's, a major change in vegetation cover has occurred in salt marshes of the Mont Saint-Michel bay (France). In this study, we investigated if the high conservation value of invaded salt marshes is preserved. Abundances, densities and flood resistance abilities of the dominant halophilic species Arctosa fulvolineata (nocturnal lycosid) and Pardosa purbeckensis (diurnal lycosid) were compared in both natural and invaded habitats. Elymus invasion involved both positive and negative aspects with respect to the conservation value of the salt marshes invaded: the P. purbeckensis population was clearly reduced in invaded habitats, whereas A. fulvolineata seemed to derive high benefits from the invasion. We supposed that abiotic parameters of the new habitat (mainly vegetation and litter characteristics) affected the two species differently with respect to their aut-ecology and their flood resistance abilities. Furthermore, food resources (estimated by different macrofauna density measurements) were likely to be reduced for P. purbeckensis in invaded habitats and unchanged for A. fulvolineata. Lastly, we hypothesize that individuals of P. purbeckensis are subject to increased interspecific competition (measured as intra-guild densities), whereas spiders from the same guild as A. fulvolineata have not increased in invaded habitats, resulting in an unchanged competition level. Keywords: Conservation value, halophilic species, habitat change, invasive species, food resources Salt marshes are one of the rarest ecosys- tems in the world (Lefeuvre et al. 2000), with a linear but fragmented distribution along coasts, therefore representing a high interest in terms of nature conservation (e.g., Gibbs 2000; Bakker et al. 2002). In fact, these eco- systems host a poorly diversified flora and fauna, which possess a low number of species that are threatened directly or indirectly by hu- man activities such as habitat destruction, dif- fuse soil pollution from adjacent cultivated fields, and overgrazing (Desender & Maelfait 1999; Adam 2002). The high conservation value of these habitats is also due to the high specificity of its fauna, adapted to two main abiotic factors, high soil salinity and regular submergence by seawater. Due to tidal events, natural salt marshes present a specific plant cover (spatial succession from the high to the low marsh) and specific invertebrate commu- nities dominated by some “halophilic arthro- pods” (Foster & Treherne 1976; Irmler et al. 2002; Petillon et al. 2004). In the Mont Saint-Michel bay, France, salt marshes have been invaded by a native spe- cies Elymus athericus (Link) Kerguelen (Va- lery et al. 2004). This high marsh-living grass (Poaceae) started to invade the salt marsh 10 years ago (Bouchard et al. 1995) and now covers some marshes to the mean sea tide lev- el. This progression is characteristic of an in- vasive species that is common in European salt marshes (Bockelmann & Neuhaus 1999). Nevertheless, the effects of E. athericus in- vasions on the high conservation value of salt marshes remain unknown. Spiders constitute an abundant, diversified and well-known tax- onomic group in salt marshes (Fouillet 1986; Desender & Maelfait 1999), including spe- cialist (stenotopic) species, the so-called “hal- 236 PETILLON ET AL.— GRASS INVASION ON SALT MARSH SPIDERS 237 Table L — Synthesis of field experiments carried out in the Mont Saint-Michel bay salt marshes (‘Ferme Foucault' and ‘Vivier-sur-mer’ sites). Parameter Methodology Period Number of replicates Target species abundances Pitfall trap (r = 5 cm) April-November 2002 8 Target species densities Depletion quadrat (1 m^) June-July 2003 32 Target species abundances be- fore and after the tide Pitfall trap (r = 5 cm) April 2004 12 Abiotic parameters WET sensor and manual mea- surements July-August 2002 16 Macrofauna densities Soil cores (r = 5 cm) October 2003 6 Amphipod densities Depletion quadrats (1 m^) June 2003 4 Total spider densities Depletion quadrats (0.25 m^) June 2002 16 ophilic species” (Hanggi et al. 1995) or “salt marsh resident species” (Petillon et al. 2004). In Europe, the numerically dominant hunting spiders in salt marsh ecosystems are the rare lycosid species Pardosa purbeckensis EO. Pickard-Cambridge 1895 and Arctosa fulvolb neata (Lucas 1846) (Fouillet 1986; Baert & Maelfait 1999; Elkaim & Rybarczyck 2000). Pardosa purbeckensis will be considered in this study as different from P. agrestis, dom- inant in Central Europe agricultural land- scapes (e.g., Samu & Szinetar 2002) and in- land salt marshes (e.g., Zulka et al. 1997), mainly because of its morphological charac- teristics (as described in Locket & Millidge 1951) and its osmoregulatory abilities (Hey- demann 1970). Arctosa fulvoUneata is classi- fied as a nationally endangered species in the United Kingdom (Harvey et al. 2002) and be- longs to the “rare species” category in the Ramsar Convention, The aim of the study was to investigate if the newly created vegetation cover induced by the invasion of Elymus athericus still consti- tutes a suitable habitat for these two species of high conservation value. To explain the po- tential impact of the invasive plant on spider species habitat preferences, natural and invad- ed habitats were characterized by abiotic and biotic components. We hypothesized that Ely- mus athericus changed the general microhab- itat characteristics and induced 1) an effect on species density, activity density and flooding resistance due to less appropriate abiotic pa- rameters (e.g., salinity or litter depth), 2) a change in the quality and quantity of food re- sources, and 3) a variation in the predation rate (measured as intra-guild densities) upon the two species Arctosa fulvoUneata and Par- dosa purbeckensis. METHODS Study sites and field surveys. — The Mont Saint-Michel bay, located between Brittany and Normandy (North West France), is unique in Europe for its tidal amplitude, which reach- es 15 meters (the second largest in the world). This exceptional phenomenon results in the extension of salt marshes and mud flats, which together cover 250 km^. Salt marshes are only submerged during spring tides (i.e. monthly strong tides) and are then inundated during two hours per tide (Lefeuvre et al. 2000). The uppermost parts of the salt-marshes are delim- ited by dikes and are not submerged during high tides. Natural stations (dominated by Atriplex portulacoides, Chenopodiaceae) and invaded stations (dominated by Elymus athericus, Po- aceae) were located at the same distance from the dike. Habitat characterization and mea- surements of spider activity and density were carried out at four stations (two natural and two invaded) located at the ‘Ferme Foucault' site’ (Normandy, 48°55'N, 1°52'W) whereas experiments on species reactions to flooding were carried out at six stations (three natural and three invaded) located at the ‘Vivier-sur- mer’ site (Brittany, 48°60^N, 1°78'W). Spider sampling. — Spider abundances: Spiders were sampled with pitfall traps, con- sisting of polypropylene cups (10 cm diame- ter, 17 cm deep) set into the ground so that the lips were flush with the soil surface. Eth- ylene glycol was used as preservative, be- cause of its lack of effects on spider catches 238 THE JOURNAL OF ARACHNOLOGY Table 2. — Mean spider activity abundances (number of individuals/day/meter XlO) and mean densities (number of individuals/m^) in invaded and natural vegetations, and statistical comparisons using ANOVA (Code: n.s. = non significant, * = F < 0.05, ** = p < 0.01). Natural plots Invaded plots F-ratio Code Spider abundances Arctosa fulvolineata 9.31 ± 3.09 9.83 ± 3.15 0.01 n.s. Pardosa purbeckensis 98.5 ± 23.80 41.72 ± 7.70 5.13 * Spider densities Arctosa fulvolineata 0.03 ± 0.03 0.43 ± 0.12 9.71 Pardosa purbeckensis 3.93 ± 1.24 1.37 ± 0.27 4.09 * (Topping & Luff 1995). Traps were covered with a raised wooden roof to keep out rain. Catches in pitfall traps were related to trap- ping duration and pitfall perimeter, which cal- culates an ‘activity trapability density’ (num- ber of individuals per day and per meter: Sunderland et al. 1995). Four pitfall traps were installed at each station and spaced 10 meters apart, considered as the minimum dis- tance for avoiding interference between traps for spider catches (Topping & Sunderland 1992). Activity trapability density of spiders was followed during the entire period of ac- tivity of the two species from April-Novem- ber 2002 except during high tides (Table 1). Spider densities: To compare absolute den- sities of the two lycosids, Im X Im plastic quadrats (1 m height) were used. Quadrats were sampled by regular hand catches and pit- fall traps (one placed in the middle of the quadrat) until there were no more individuals. Eight spatial replicates were sampled in each vegetation type (i.e. invasive and natural), and this technique was used four times in June and July 2003 (i.e. 32 replicates per vegetation: Table 1). Ejfects of flooding: To determine the role of vegetation cover in modifying species ability to resist flooding, spiders were sampled before and after a spring tide (tidal range: 13.35 m) in April 2004 (Table 1). Three natural and three invaded stations were studied at three salt marsh levels ranging from the high to the low marsh (i.e. 50, 150 and 200 meters from the dike). Four pitfall traps per station were acti- vated during three consecutive days before and after the high tide. Catches were completed by hand collecting during activation and collection of pitfall traps for a total of 1 .5 hours per sta- tion before and after the high tide. Characterization of natural and invaded habitats. — Abiotic characteristics: To char- acterize plant communities, vegetation was described four times within a radius of 1 m around each pitfall trap (4 replicates per sta- tion) at the ‘Ferme Foucault’ site: litter depth (to the nearest mm) and vegetation height (to the nearest cm). Soil salinity (estimated by pore water electrical conductivity), soil water content and temperature were measured using a W.E.T. Sensor connected to a moisture meter HH2 (Delta-T Devices Ltd., Cambridge, UK) and made with a specific clay soil calibration. All abiotic variables were assessed in summer 2002 (Table 1). Soil temperatures were mea- sured at llh a.m. to compare this micro-cli- mate parameter between habitats, but not to characterize it along time. Biotic characteristics: Food resources were estimated by vertical sampling using a soil core (12 cm depth and 10 cm for diameter), except for the very mobile amphipod (see be- low). Six cores were collected in each habitat (natural or invaded) during October 2003 (Ta- ble 1). Cores were then sieved through a 250 jjim mesh screen. Because cannibalism and in- traguild predation are a general rule in spiders, especially in structurally simple ecosystems (e.g., Schaefer 1974), spider densities were in- cluded in total arthropod densities when com- paring the food resources between invaded and natural areas. The results do not include an important macrofauna component, the gas- tropod Phytia bidentata because this species is not likely to be consumed by the two ly- cosid species. Because the amphipod Orchestia gammar- ella represents one of the more abundant ar- thropods in west European salt marshes (e.g., Meijer 1980), special attention was paid to PETILLON ET AL.— GRASS INVASION ON SALT MARSH SPIDERS 239 Table 3. — Comparison of mean activity abundances measured by pitfall traps (number of individuals/ day /meter; in parentheses: total number of individuals) before and after the high tide in natural and invaded stations, and statistical comparisons using ANOVA (Code: n.s. = non significant, * = p < 0.05, ** = p < 0.01). Species abundances Before the high tide After the high tide F-ratio Code Natural stations Arctosa fulvoUneata 1.77 ± 0.49 (40) 0.44 ± 0.21 (18) 6.14 * Pardosa purbeckensis 3.36 ± 1.14 (364) 2.12 ± 0.47 (274) 1.00 n.s. Invaded stations Arctosa fulvoUneata 0.97 ± 0.28 (31) 0.80 ± 0.27 (23) 0.21 n.s. Pardosa purbeckensis 4.69 ± 1.06 (214) 2.03 ± 0.51 (163) 5.04 * this species. Densities of O. gammarella were calculated in natural and invaded habitats us- ing a depletion method. Within each plant community, four Im^ quadrats were randomly sampled in June 2003 by hand catches and pitfall traps until there were no more individ- uals. To measure nocturnal and diurnal wanderer densities, spider densities were estimated by the quadrat-flotation technique. Homogeneous 0.25 m^ areas were isolated by iron quadrats (0.5 meter width and 1 meter height) set into the ground to a depth of 0.20 m. All vegeta- tion was removed, stored in sealed bags and analyzed in the laboratory. The spiders re- maining within the quadrat were then hand collected on the bare soil until none were vis- ible. A pitfall trap was placed in the middle of the quadrat and left in place until all re- maining moving individuals were caught, A last hand collection was also carried out dur- ing the time of pitfall trapping. The spider density was calculated by summing individu- als of the initial and final hand collections with the catches of both pitfall trap and veg- etation samplings. Four replicates were per- formed at each station and repeated four times during June 2002 (Table 1). Identification and data analyses.— All the spiders collected were preserved in 70% eth- anol, transported to the laboratory for species identification and kept in the University collec- tion (Rennes, France). In the tables, all means are presented with standard error (mean ± s.e.). Mean environmental and species variables were compared using ANOVA tests after ver- ification of normal distribution according to Kolmogorov-Smimov tests (MINITAB). RESULTS Comparison of habitat preference. — Catches by pitfall traps revealed that Pardosa purbeckensis activity abundances were signif- icantly higher in natural than in invaded plots (Table 2). In accordance with trends found in activity, P. purbeckensis had a significantly higher density in natural plots. However, Arc- tosa fulvoUneata had much higher densities in invaded than in natural plots, whereas its abundance did not differ significantly between the two vegetation types (Table 2). Comparison of flooding effects between natural and invaded vegetation. — Based on their distribution change after the high tide, Arctosa fulvoUneata and Pardosa purbecken- sis presented similar reactions to flooding, with an unmodified distribution (in terms of presence/absence) in natural and invaded sta- tions. Only A. fulvoUneata almost disappeared from one natural station (level 2), where less than 5 individuals were caught after the high tide. In fact, this species presented comparable abundances before and after the high tide in invaded stations (Table 3), whereas its mean abundance significantly decreased in natural stations. Pardosa purbeckensis showed signif- icantly decreased abundances after the high tide in invaded stations and constant abun- dances in natural stations (Table 3). Comparison of habitat characteristics, — Invaded stations {Elymus athericus) differed significantly from natural stations (Atriplex portulacoides), i.e. they had deeper litter and taller plant cover (Table 4). No significant dif- ferences were found regarding soil salinity, soil water content and temperature. Regarding 240 THE JOURNAL OF ARACHNOLOGY potential prey, Acari and Amphipoda {Or- chestia gammarella) densities were signifi- cantly lower in invaded areas, leading to a de- crease in total pedofauna at these stations (Table 4). No significant difference was found for Collembola. Species from the guild of Pardosa pur- beckensis (diurnal wanderers) were found in significantly higher densities in invaded than in natural stations (the more dominant species were the non-coastal Pardosa prativaga, P. proxima and P. pullata), whereas no differ- ence between vegetation types was found for species from the Arctosa fulvolineata guild (nocturnal wanderers, mainly including the non-coastal Agroeca lusatica, Clubiona stag- natilis and Zelotes latreillei). DISCUSSION Changes in habitat characteristics after Elymus invasion. — Lycosids in general and the genus Pardosa in particular are known to prefer open habitats (e.g., Aart 1973), and Harvey et al. (2002) suggested that adults of P. purbeckensis are favored by low vegeta- tion. Kessler & Slings (1980) demonstrated the tendency among juveniles of P. purbeck- ensis to aggregate and to select grass with high shoot densities, probably to avoid can- nibalism. Because more diurnal wandering spiders are hunting in invaded plots, we sug- gest that young specimens of P. purbeckensis are exposed to a higher predation risk than in natural areas, contributing to the lower adult density in the invaded areas. Thus we propose that the habitat structure of Elymus (mainly tall vegetation and deep litter) was not suitable for this species and contributed to its lower occurrence in invaded habitats. In contrast, Arctosa fulvolineata might be favored by the structure of the invaded habitats: a deeper and more complex litter due to a lower rate of Ely- mus litter decomposition (Valery et al. 2004). As a general rule, deep litter, by providing new microhabitats and microclimate condi- tions (Wise 1993), tends to favor nocturnal wanderers (case of Arctosa fulvolineata), am- bush hunters (thomisids) and “litter-sensi- tive” sheet-weavers (Bell et al. 2001). Thus we suggest that A. fulvolineata prefers more heterogeneous litter of invaded areas, where it is often found during the day, at 3-4 cm depth (Petillon pers. obs.). Differences in the responses of dominant salt marsh species to the invasion can be in- fluenced by many other factors than abiotic components of habitats. In particular for sim- ple ecosystems such as salt marshes, interspe- cific competition may reduce spider popula- tions (Schaefer 1972 according to Wise 1993; Marshall & Rypstra 1999). We propose that invaded habitats, by hosting other species from the same guild as P. purbeckensis (es- pecially Pardosa prativaga, P. proxima and P. pullata), increased inter-specific competi- tion. Contrary to P. purbeckensis, A. fulvoli- neata seemed not to be subjected to higher levels of competition in invaded habitats, as nocturnal wanderer densities were the same in natural and invaded habitats. Little is known about salt marsh spider diets, except predation upon Collembola for P. purbeckensis (Schaef- fer 1974) and for the linyphiid Erigone arctica (White 1852) (Legel & Wingerden 1980). So, at the moment there is no evidence for food limitation, even if some potential prey de- creased in invaded habitats (e.g., the amphi- pod Orchestia gammarella). Changes in flood resistance in invaded habitats. — In this study, the two salt marsh species responded differently to flooding in invaded habitats. Pardosa purbeckensis abun- dances decreased only in invaded stations af- ter the high tide. If this species uses under- ground refuges during flooding, as do several intertidal invertebrates (Foster & Treherne 1976; Kneib 1984), then it would be disfa- vored in invaded areas (deep but thin roots of Elymus) compared to natural areas (short and large roots of Atriplex portulacoides). Con- trary to P. purbeckensis, A. fulvolineata seem to derive benefits from the invasive plant be- cause its abundance remained the same after the high tide in the invaded stations and de- creased in the natural stations. This litter-liv- ing species might be favored by Elymus ath- ericus that seems to improve its habitat by increasing the food resources and the amount of air in the litter. Conflicting aspects of the grass invasion: how to manage it. — Habitat suitability of nat- ural and invaded areas clearly differed be- tween the two wolf spiders. Activity trappa- bility density of P. purbeckensis was strongly reduced in invaded areas indicating a decrease in the density and/or the mobility of individ- uals, whereas activity trappability density of A. fulvolineata was enhanced in invaded sta- PETILLON ET AL.— GRASS INVASION ON SALT MARSH SPIDERS 241 Table 4. — Habitat characteristics (mean ± s.e.) of natural and invaded stations, and statistical compar- isons using ANOVA (Code: : n.s. = non significant. * = P < 0.05, ** = = P < 0.01). Units Natural plots Invaded plots F-ratio Code Structural components Vegetation height cm 29 -1- 0.63 78 -i- 1.64 331.43 ** Litter depth Soil water content cm % 60.13 0 -+- 4.21 4 57.72 -H -1- 0.23 1.12 754.60 0.31 ** n.s. Soil salinity mS/m 1009.7 -f- 50.1 1013.8 -+- 45.1 <0.01 n.s. Soil diurnal tempera- ture °C 10.86 0.62 1 1.00 -H 0.38 0.03 n.s. Biotic components Amphipoda Number/m^ 979 311 205 -t- 51 6.02 * Acari Number/m^ 55302 -1- 8688 20202 -1- 3085 14.49 ** Collembola Number/m^ 5008 + 2391 5645 -t- 2574 0.03 n.s. Total pedofauna Number/m^ 60352 -h 7053 26293 -1- 3707 18.27 Diurnal wanderers Number/m^ 6.50 -+- 2.10 29.25 -h 6.97 9.77 * Nocturnal wanderers Number/m^ 0.75 0.48 5.25 -1- 2.78 2.54 n.s. tions. Here, quadrat results confirmed that changes in activity trappability densities also reflected changes in the species’ population densities. More than ten years after the grass inva- sion, the two dominant halophilic species are still present in invaded salt marshes (Fouillet 1986; present study). But our results indicate that in the near future microhabitat changes and (perhaps) competition for space and food between native and non-coastal (immigrated after the Elymus invasion) species could lead to a decline in the absolute density of the dominant native species such as Pardosa pur- beckensis. This is a unique case of an invasion which involves, as it seems at the moment, both positive and negative aspects for domi- nant halophilic spiders, enhancing one species and reducing the other! As similar invasions are affecting more and more salt marshes in Western Europe (Bockelmann & Neuhaus 1999), we suggest that management plans should be undertaken for reducing the para- doxical consequences of Elymus invasion on typical and rare salt marsh biodiversity. The effects of mowing are surveyed at the moment in the Mont Saint-Michel bay. By re-opening the soil, mowing seems to maintain healthy populations of both halophilic species. This is in contrast to sheep-grazing that is likely to only favor P. purbeckensis, and even tends to disfavor both dominant halophilic wolf spi- ders when practiced in a too intensive way (Petillon et al. submitted). ACKNOWLEDGMENTS We would like to thank R Amaya Garcia, Q. Le Mouland, S. Michaud, G. Perrin, E. Pe- tillon and E. Villani for field assistance, Gobi Camberlein and Julien Cucherousset for En- glish improvement. S. Toft and two anony- mous reviewers greatly improved the manu- script by their remarks and suggestions. This study was supported by the French Ministry of Environment (Programme ‘Especes inva- sives’) and the CNRS Zone Atelier ‘La Baie du Mont Saint-Michel et ses bassins versants’. LITERATURE CITED Aart, P.M.J. van der. 1973. Distribution analysis of wolf spider (Araneae, Lycosidae) in a dune area by means of principal component analysis. Neth- erlands Journal of Zoology 23:266-329. Adam, P. 2002. Saltmarshes in a time of change. Environmental Conservation 29:39-61. Baert, L. & J.-R Maelfait. 1999. The spider fauna of the Belgian salt marshes. Bulletin de ITnstitut Royal des Sciences Naturelles de Belgique 69:5- 18. Bakker, J.P., P Esselink, K.S. Dijkema, W.E. van Duiin & D.J. de Jong. 2002. Restoration of salt marshes in the Netherlands. Hydrobiologia 478: 29-51. Bell, J.R., C.R Wheater & W.R Cullen. 2001. The implications of grassland and heathland manage- ment for the conservation of spider communities. Journal of Zoology 25:377-387. Bockelmann, A.C. & R. Neuhaus. 1999. Competi- tive exclusion of Elymus athericus from a high- stress habitat in a European saltmarsh. Journal of Ecology 87:503-513. 242 THE JOURNAL OF ARACHNOLOGY Bouchard, V., E, Digaire, J.-C. Lefeuvre & L.M. Guillon. 1995. Progression des marais sales a Pouest du Mont-Saint-Michel entre 1984 et 1994. Mappemonde 4:28-34. Desender, K. & J.-P. Maelfait. 1999. Diversity and conservation of terrestrial arthropods in tidal marshes along the River Schelde: a gradient analysis. Biological Conservation 87:221-229. Elkaim, B. H. & Rybarczyck. 2000. Structure du peuplement des invertebres des zones halophiles de la Bale de Somme (Manche orientale). Ca- hiers de Biologie Marine 41:295-311. Foster, W.A. & J.E. Treherne. 1976. Insects of ma= rine saltmarshes: problems and adaptations. Pp. 5-42. In Marine Insects. (L. Cheng, ed.). North- Holland Company, Amsterdam. Fouillet, P. 1986. Evolution des peuplements d’Arthropodes des schorres de la Bale du Mont Saint-Michel: influence du paturage ovin et con- sequences de son abandon. Ph.D. thesis. Univer- sity of Rennes 1, Rennes. Gibbs, J.P. 2000. Wetland loss and biodiversity con- servation. Conservation Biology 14:314-317. Hanggi, A., E. Stocklie & W. Nentwig. 1995. Hab- itats of central European spiders. Centre Suisse de cartographie de la faune, Neuchatel. Harvey, P.R., D.R. Nellist & M.G. Telfer. 2002. Pro- visional Atlas of British Spiders (Arachnida, Ar- aneae). Biological Records Centre, Huntington. Heydemann, B. 1970. Okologische untersuchungen zum problem der halophilen und haloresistenten spinnen. Bulletin du Museum National d’Histoire naturelle 41:226-232. Irmler, U., K. Heller, H. Meyer & H.-D. Reinke. 2002. Zonation of ground beetles (Coleoptera: Carabidae) and spiders (Araneida) in salt marsh- es at the North and the Baltic Sea and the impact of the predicted sea level increase. Biodiversity and Conservation 11:1129-1147. Kessler, A. & R. Slings. 1980. Microhabitat selec- tion in adults and juveniles of Pardosa purbeck- ensis EO.P.-Cambridge (Araneae, Lycosidae). Pp. 151-154. In Proceedings of the 8* Interna- tional Congress of Arachnology (J. Gruber, ed.). Verlag H. Egermann, Wien. Kneib, R.T. 1984. Patterns of invertebrate distribu- tion and abundance in the intertidal salt marsh: causes and questions. Estuaries 7:392-412. Lefeuvre, J.-C., V. Bouchard, E. Feunteun, S. Grare, P. Laffaille & A. Radureau. 2000. European salt marshes diversity and functioning: the case study of the Mont Saint-Michel bay, France. Wetland Ecology and Management 8:147-161. Legel, G.J. & W.K.R.E. van Wingerden. 1980. Ex- periments on the influence of food and crowding on the aeronautic dispersal of Erigone artica (White 1852) (Araneae Linyphiidae). Pp. 1-6. In Proceedings of the 8^’’ International Congress of Arachnology (J. Gruber, ed.). Verlag H. Eger- mann, Wien. Locket G.H. & A.F. Millidge. 1951. British Spiders (volume I). The Ray Society, London. Marshall, S.D. & A.L. Rysptra. 1999. Spider com- petition in structurally simple ecosystems. Jour- nal of Arachnology 27:343-350. Meijer, J. 1980. The development of some elements of the arthropod fauna of a new polder. Oecolo- gia 45:220-235. Petillon, J., E Ysnel, S. Le Gleut, J.-C, Lefeuvre & A. Canard. 2004. Responses of spider commu- nities to salinity and flooding in a tidal salt marsh (Mont St-Michel Bay, France). Arthropoda Se- lecta special issue 1:235-248. Samu F. & C. Szinetar. 2002. On the nature of agro- biont spiders. Journal of Arachnology 30:389- 402. Schaefer, M. 1972. Okologische Isolation und die Bedeutung des Konkurrenzfaktors am Beispiel des Verteilungsmusters der Lycosiden einer Kiis- tenlandschaft. Oecologia 9:171-2002. Schaefer, M. 1974. Experimentelle Untersuchungen zur Bedeutung der interspezifischen Konkurrenz bei 3 Wolfspinnen-Arten (Araneida: Lycosidae) einer Salzwiese. Zoologische Jahrbiicher, Abtei- lung Systematik, Okologie und Geographie der Tiere 101:213-235. Sunderland, K.D., G.R. De Snoo, A. Dinter, T. Hance, J. Helenius, P. Jepson, B. Kromp, F. Samu, N.W. Sotherton, S. Toft & B. Ulber. 1995. Density estimation for invertebrate predators in agroecosystems. Pp. 133-162. In Toft, S. & Rie- del, W. (eds.). Arthropod natural enemies in ar- able land,. I. Density, spatial heterogeneity and dispersal. Acta Jutlandica 70:2. Aarhus Universi- ty Press, Arhus. Topping, C.J. & M.L. Luff. 1995. Three factors af- fecting the pitfall trap catch of linyphiid spiders (Araneae: Linyphiidae). Bulletin of the British Arachnological Society 10:35-38. Topping, C.J. & K.D. Sunderland. 1992. Limita- tions to the use of pitfall traps in ecological stud- ies exemplified by a study of spiders in a field of winter wheat. Journal of Applied Ecology 29: 485-491. Valery, L., V. Bouchard & J.-C. Lefeuvre. 2004. Impact of the invasive native species Elymus ath- ericus on carbon pools in a salt marsh. Wetlands 24:268-276. Wise, D.H. 1993. Spiders in Ecological Webs. Cam- bridge University Press, Cambridge. Zulka, K.P., N. Milasowszky & C. Lethmayer. 1997. Spider biodiversity potential of an ungrazed and a grazed inland salt meadow in the National Park ‘Neusielder See-SeewinkeP (Austria): implica- tions for management (Arachnida: Araneae). Bio- diversity and Conservation 6:75-88. Manuscript received 29 December 2004, revised 7 September 2005. 2005. The Journal of Arachnology 33:243-246 THE DIET OF THE CAVE SPIDER META MENARDI (LATREILLE 1804) (ARANEAE, TETRAGNATHIDAE) Peter Smithers: School of Biological Sciences, University of Plymouth. Drake Circus, Plymouth, Devon, PL4 8AA, UK. E-mail: Psmithers@plymouth.ac.uk ABSTRACT. This study investigated the range and number of prey consumed by a population of M. menardi in an abandoned mine drainage adit at Mary Tavy, on the edge of Dartmoor (Devon, UK). The adit was visited each week from October 1997 to November 1998 and any spider found feeding was interrupted and its prey removed and preserved in alcohol. Over the 13 months a total of 69 prey were recovered representing 18 taxa. While a number of flying insects used the adit as a refuge in which to over winter they formed a small percentage of the total prey consumed. Most of the prey were members of the soil or litter fauna (myriapods and slugs) that were observed walking over the surface of the adit walls. Keywords: Meta menardi, prey, myriapods, slugs, litter fauna Meta menardi (Latreille 1804) is a well known member of the twilight zone commu- nity in subterranean systems. This zone is lo- cated just beyond the entrance and provides an environment buffered from the extremes of the outside world while still receiving light from the external environment. As such, this zone forms a series of small habitat islands on the fringes of subterranean systems. While this habitat is connected to the external envi- ronment, the dark entrance zone acts as a sig- nificant barrier to many invertebrates, thus the diversity of potential prey in the twilight zone is low by comparison with the outside. Pred- ators that occupy this zone are isolated from the surrounding habitats and thus have access to a limited range of potential prey. Yet as populations of M. Menardi can be large (in excess of a 100 individuals, pers. obs.), it is clear that an abundance of prey must be avail- able in order to sustain populations of this size. A number of invertebrates have been pre- viously recorded as prey of M. menardi. Yosh- ida & Shinkai (1993) recorded Diplopoda, Diptera, Formicidae and Vespidae from a pop- ulation living between boulders in Japan, while both Eckert & Moritz (1992) and Chap- man (1993) recorded myriapods, Coleoptera and isopods as prey from populations in Ger- many and the UK respectively. Previous stud- ies by the author recorded 2 species of dip- lopods, 1 isopod, carabid beetles, spiders of the genus Meta, plus slugs and oligochaetes (Smithers 1996). Studies of a cellar population in Germany revealed; 9 spp. Coleoptera, 2 spp. Isopoda, 3 spp. Araneae, 2 spp. Diptera, 1 spp. Gastropoda, 2 spp. 1 spp. Opiliones, Chilopoda, 1 spp. Nematophora, 1 spp. Hy- menoptera and 1 spp. Pseudoscorpiones (Potzsch 1966). While previous work has shown that M. menardi consumes a wide range of prey there has not been a systematic study of the relative abundance of these prey in the diet of this species or an investigation of any seasonal variation. This study was designed to inves- tigate the diet of M. menardi, and to determine any seasonal or life stage variations in the prey consumed. METHODS The work was conducted in an abandoned mine drainage adit on the edge of Dartmoor, Devon, UK (SX 513787). A man-made tunnel was chosen due to its linear nature which meant that all members of the population were accessible for observations. The adit was sit- uated in a steep bank, the top of which was covered with deciduous woodland. The site was visited each week from October 1997- November 1998 (no data was gathered be- tween December 97-January 98). At each vis- it the population was examined for spiders with prey in their web or mouthparts. When spiders with prey were disturbed the spider 243 244 THE JOURNAL OF ARACHNOLOGY Table L — The abundance of prey groups recovered from different life stages of Meta menardi. Prey taxon Immature Female Male Total % of total Unidentified prey items 1 2 3 4 Diptera, Nematocera 2 1 2 3 Diptera, Culicidae 1 1 1 Diptera, Eristalis sp. 1 1 1 Coleoptera, Carabidae 2 3 5 7 Trichoptera unidentified 5 4 9 13 Trichoptera, Stenophylax permistus 1 1 1 Trichoptera, pupae 1 1 1 Neuroptera, Sisyridae 1 1 1 Lepidoptera, Scoliopteryx libatrix 1 1 1 Araneae, imm Meta menardi 1 1 1 2 3 Araneae, Meta merianae 2 1 3 4 Myriopoda, unidentified Diplopoda 1 1 1 Myriopoda, Julidae 5 5 10 14 Myriopoda, Cylindroiulus punctatus 1 2 3 4 Myriopoda, Nanagona polydesmoides 4 3 7 10 Myriopoda, Chilopoda Geophilomorpha 3 3 4 Gastropoda (Slugs) 1 12 13 19 Total number of prey recorded 26 39 4 69 would retreat to the top of the web leaving the prey hanging by a silken thread. Any prey dis- covered were removed and taken back to the laboratory were they were preserved in 70% alcohol, then identified to the lowest taxonom- ic unit possible. The prey recovered were al- ways wrapped in silk and in an advanced state of digestion. The exoskeleton of the arthro- pods were broken open and fragmented while the molluscs were usually digested from one end, occasionally leaving a head or rear intact. All of the spiders sampled were assigned to one of three life stage groups, females, males or immatures. Voucher specimens of M. men- ardi collected at the study site have been lodged in the invertebrate collection at the University of Plymouth. RESULTS A total of 69 prey were recovered repre- senting 17 taxa, only 3 of which proved to be unidentifiable (Table 1). The myriapods were the most abundant prey recovered with 24 in- dividuals, followed by slugs with 13 individ- uals, Trichoptera with 11, Araneae with 6 and Carabidae with 5. Other prey were recorded in small numbers (Table 1). It is clear that three taxa dominate the prey recovered over the sampling period. These are the myriapods, the gastropod slugs and the Trichoptera. Few prey were recovered from males while fe- males and immature spiders were recorded consuming approximately equal numbers of most prey groups except slugs, which were primarily collected from females (only one slug was not taken from a female) (Table 1). DISCUSSION The myriapod prey comprised three main taxa, julid millipedes (some of which could be identified as Cylindroiulus punctatus), Nana- gona polydesmoides and the geophilomorph centipedes. The julids were more abundant in May (4 individuals) and in the autumn (Oc- tober & November 2 individuals each) but were occasional prey at other times of the year (February & August 1 individual each). The slight increase in numbers captured in the spring and autumn could be explained by a seasonal vertical migrations in the litter/soil undertaken by julids as reported by Geoffroys (1981). While the autumn migration is down- ward, both the spring and autumn migrations would involve an increase in the activity of individuals. Given the proximity of the adit entrance to the soil litter interface, some in- dividuals becoming active in the spring could reach the surface and follow the rock surface down into the adit entrance. The geophilo- morph centipedes also displayed a small au- tumn peak which could also be explained by seasonal migrations down the soil profile. SMITHERS-DIET OF META MENARDI 245 Nanagona is also an occasional prey over the spring and summer, which is not surprising as this is a well known troglophile that is com- monly encountered in subterranean chambers (Blower 1985; Chapman 1993). The Trichoptera also displayed seasonal patterns of abundance, being abundant in April / May and again in August where they displayed a distinct peak. This was probably the result of an emergence of adults from ei- ther the river outside the adit or the stream within it (a single pupal Trichoptera was re- corded, indicating that individuals were emerging within the adit). The slugs were recovered in small numbers over the late spring through to the autumn in which they displayed a distinct peak. This peak may be a result of their seasonal migra- tion down the soil profile to escape the harsher winter conditions. This would bring them into the mine adit via the micro caverns in the bed rock. Once at a safe depth they are then qui- escent for the winter months, thus explaining their absence from the diet of M. menardi be- tween December and April. At 19% of the prey captured (Table 1) gastropods are an im- portant element in the diet of M. menardi. This is unusual for a spider as a recent review of malocophagy in spiders has shown M. men- ardi to be the only araneomorph spider to in- clude gastropods as a regular part of its diet (Nyffeler & Symondson 2001). While slugs have been observed crawling into water laden webs of Argiope bruennichi (Scopoli 1772) (Quicke 1987), the exact meth- od of prey capture used by M. menardi is as yet unknown. The bias in the number of slugs recovered from females (Table 1) hints that this particular prey may require the larger body size exhibited by most females to suc- cessfully capture this prey. A similar bias has also been observed in some carabid beetles (Nyffeler & Symondson 2001). Further work is required to determine the exact nature of the prey capture method for this species. The remaining prey were captured in low numbers through out the year. A number of flying insects such as the hover fly Eristalis tanax, the Golden caddis fly Stenophylax per- mistus, the herald moth Scoliopteryx libatrix, and mosquitoes of the genus Culex commonly use underground chambers as over-wintering sites (Chapman 1993). These can aggregate in large numbers on the walls of underground chambers but, despite their presence in the adit these species were not common elements of M. menardTs prey spectrum. This may be a reflection of their behavior, as they fly into the chamber and quickly settle on the walls where they become immobile until the follow- ing spring (pers. obs.). Unless they land in a web they are unlikely to attract a spider’s at- tention. Carabidae were recorded in the spring with a single record from the autumn. These are active predators that had probably strayed into the adit via the entrance from the woodland floor above. The spiders Metellina meriane (Scopoli 1763) and M. menardi were occa- sional prey over the spring and autumn, hint- ing that for any individual moving around within the chamber can be hazardous. Predators that occupy the twilight zone have access to a range of potential prey which can be divided into three groups: organisms that move into the subterranean system from the external environment to seek shelter or over winter; those that move down the litter/ soil profile and into the chambers via the mi- cro and meso cavern network that connects the macro chambers with the overlying soil system; and members of the deep cave fauna that may stray into this zone. It appears that M. menardi has specialized in capturing mem- bers of the soil/litter fauna that stray in to un- derground chambers, but will respond oppor- tunistically to any additional prey that walk over the inner surface of the underground chamber. ACKNOWLEDGMENTS The author would like to thank the review- ers for their comments on the first draft of the manuscript. LITERATURE CITED Blower, J.G. 1985. Millipedes. E.J. Brill. London. 242 pp. Chapman, P. 1993. Caves and Cave Life. Harper Collins, London. 224 pp. Ekert, R. & M. Moritz. 1992. Meta menardi (Latr.) and Meta meriane (Scop.): On the biology and habitat of the commonest spiders in caves of the Harz, the Kyffhauser, Thuringa and the Zittau mountains. Mitteilungen aus dem Zoologischen Museum Berlin 68(2):345-350. Geoffroys, J.J. 1981. Modalites de la coexistance 246 THE JOURNAL OF ARACHNOLOGY de deux diplopodes, Cylindroiulus punctatus (Leach) et Cylindroiulus nitidus (Verhoeff) dans un ecosyteme forestier du Bassin Parisien. Acta Oecologia Oecologia Generalis 2:357-372. Nyffeler, M. & W.O.C. Symondson. 2001. Spiders and harvestmen as gastropod predators. Ecology ical Entomology 26:617-628. Potzsch, J. 1966. Notizen zur Ernahrung und Le- bensweise von Meta menardi Latr. (Araneae; Ar- aneidae). Abhandlungen und Berichte des Natur- kundemuseums Gorlitz 41(10): 101-122. QuickC D.L.J. 1987. Orb-weaving spiders preying on slugs. Proceedings and Transactions of the British Entomological and Natural History So- ciety, 20:90. Smithers, P. 1996. Observations on the prey of the cave spider Meta menardi (Latreille, 1804). Newsletter of the British Arachnological Society 77:12-14. Yoshida, M. & A. Shinkai. 1993. Predatory behav- iour and web structure of Meta menardi (Ara- neae: Tetragnathidae). Acta Arachnologica 42(1): 21-25. Manuscript received 4 January 2005, revised 5 September 2005. 2005. The Journal of Arachnology 33:247-255 THE SPIDER FAUNA OF THE IRRIGATED RICE ECOSYSTEM IN CENTRAL KERALA, INDIA ACROSS DIFFERENT ELEVATIONAL RANGES P.A. Sebastian: Division of Arachnology, Dept, of Zoology, Sacred Heart College, Thevara, Cochie-682 013, Kerala, India. E-mail: drpothalil@rediffmaiLcom M*J. Mathew: Division of Arachnology, Dept, of Zoology, Sacred Heart College, Thevara, Cochin-682 013, Kerala, India S. Pathummal Beevi: Dept, of Agricultural Entomology, Biological Control of Crop Pests & Weeds, College of Horticulture, Kerala Agricultural University, Vellanikkara- 680 654, Thrissur, Kerala, India John Joseph: Division of Arachnology, Dept, of Zoology, Sacred Heart College, Thevara, Cochin-682 013, Kerala, India C.R. Biju: Dept, of Agricultural Entomology, Biological Control of Crop Pests & Weeds, College of Horticulture, Kerala Agricultural University, Vellanikkara-680 654, Thrissur, Kerala, India ABSTRACT. A survey of spiders associated with the irrigated rice ecosystem in central Kerala, India was conducted across different elevational ranges. Spiders were collected from rice fields of high ranges, midland and low land areas in two cropping seasons viz., June-September 2002 (Kanni Krishy) and October 2002-Febmary 2003 (Makara Krishy) with a total of 144 hours of sampling time distributed across the two seasons. The sampling areas constituted Adimali and Marayoor of Idukki district (high range), Vannappuram of Idukki district and Kothamangalam of Emakulam district (midland) and Parak- kadavu and Piravom of Emakulam district (lowland). Visual searching methods were used to sample the spider fauna from quadrats. A total of 1130 individuals belonging to 92 species, 47 genera and 16 families were recorded during the study period. Araneidae and Tetragnathidae were the dominant families and Tetragnatha mandibulata Walckenaer 1842 (Family Tetragnathidae) the most abundant species. Various diversity indices, as well as richness and Chao I estimator were used to analyze the possible effect of elevation on species occurrence; the results showed that species richness and diversity were the highest in Parakkadavu, which is a lowland area. In a cluster analysis the localities belonging to the same elevation were found to form separate groups. The species fell into seven feeding guilds. Orb weavers were dominant at all study sites. Keywords: Araneae, central Kerala, diversity, elevational gradient. Spiders are ubiquitous in terrestrial ecosys- tems and abundant in both natural and agri- cultural habitats (Turnbull 1973; Ny feller & Benz 1987). They play an important role in regulating insect pests in agriculture ecosys- tems (Nyffeler & Benz 1987; Nyffeler et al. 1994; Sunderland 1999). Studies of Hama- mura (1969), Sasaba et al. (1973), Gavarra & Raros (1973), Samal & Misra (1975), Kobay- ashi (1977), Chiu (1979), Holt et al. (1987) and Tanaka (1989) clearly described the role of spiders as predators in reducing insect pests in rice fields. In India, studies on the popula- tion and abundance of the spider assemblages in agricultural crops are few. Some basic stud- ies were carried out by Pathak & Saha (1999) and Bhattacharya (2000). Banerji et al. (1993) and Anbalagan & Narayanaswamy (1999) also analyzed the population fluctuation of spiders in paddy fields. However, these studies were mostly limited to the identification of spiders and investigation of the dominant spi- der species. In this paper, we document the araneofauna associated with the irrigated rice 247 248 THE JOURNAL OF ARACHNOLOGY {Oryza sativa L.) ecosystem in central Kerala, India, based on studies conducted during two crop seasons. We also attempted to compare the diversity and richness of spiders across different elevational ranges and to analyze the possible effect of elevation on species occur- rence. METHODS Study Area. — The areas selected for the study belong to Ernakulam and Idukki dis- tricts in Central Kerala, India. Elevation in these two districts ranges from 0-2695 m above MSL. The sampling sites included Ad- imali (latitude 10° 5' N; longitude 77° 44' E; elevation 1 100 m above MSL) and Marayoor (10° 5' N; 77° 4' E; 2100 m above MSL) of Idukki district in the high range; Kothaman- galam of Ernakulam district (9° 58' N; 76° 34' E; 230 m above MSL) and Vannappuram of Idukki district (9° 54' N; 76° 43' E; 270 m above MSL) in midland; and Parakkadavu (11° 15' N; 75° 49' E; 1 1 m above MSL) and Piravom (9° 58' N; 76° 34' E; 30 m above MSL) of Ernakulam district in the lowland. Average annual rainfall in Ernakulam district is 343.2 cm with 139 rainy days. Temperature ranges from 20-35 °C. The western parts of Idukki district comprising the midland area experiences moderate climate, temperature varying between 21-27 °C with minimum sea- sonal variation. The eastern parts of the dis- trict located in the highland have a compara- tively cold climate with temperature varying between minus 1-15 °C in November/January and 5-15 °C during March/Aprik The sam- pling units were selected at random in all the localities. The rice ecosystem in central Kerala con- sists of two physically and morphologically distinct habitats; the rectangular shaped flood- ed fields vegetated mainly by the rice plants, and the surrounding bunds which harbor weeds. This mosaic system is connected with irrigation canals and ditches, while sump ponds, marshes and tanks serve as contiguous aquatic habitats. The rice fields are frequently disturbed by farming practices, i.e. tillage, ir- rigation, fertilization, crop establishment, weeding and pesticide application, and by nat- ural phenomena such as rainfall and flooding. These disturbances result in extreme instabil- ity on a short time-scale during the crop cycle, but relative stability on a long time-scale (Wa- 0- 0.5- (D 1 - Q C 1,5 — iS 2- ‘-a 2.5- IS 3- o 3,5 _ sz O 4_ 4,5 _ 5 ® © @ @ @ ® Figure 1. — Dendrogram for the cluster analysis of six sampling sites. 1 = Parakkadavu, 2 = Pira- vom, 3 = Vannappuram, 4 = Kothamangalam, 5 = Adimali, 6 = Marayoor. tanabe & Roger 1985). Although rice is in- fested by a multitude of insect pests, the most destructive of them in central Kerala are the rice bug, Leptocorisa acuta (Thunberg); green leafhopper Nephotettix virescens (Distant) and brown planthopper, Nilaparvata lugens (Stal). Study Time. — The study was carried out in two cropping seasons, viz. June-September 2002 (Kanni Krishy) and October 2002-Feb- ruary 2003 (Makara Krishy). A total of 144 hours were spent for sampling distributed across both cropping seasons. Sampling. — Visual search was used for sampling in each selected study site. We spent one hour in each sampling unit on a fortnight- ly basis during each cropping season. Sam- pling was done from the same field during both seasons. A total of 24 hours were spent in each site across both cropping seasons, to- taling 144 hours of sampling time. Spiders were collected from four quadrats (1 m x 1 m) placed at the four corners ofalOmX 10 m area. Collections were done during early morning hours since it was observed that spi- der activity was the maximum at that time of day in the rice fields and the morning-dew- covered webs were easy to observe. The area around each plant was searched for possible webs and the plants were thoroughly exam- ined for spiders from the bottom to the top. The spiders collected from each site every fortnight were preserved together in 70% eth- yl alcohol with proper labeling of locality, date of collection and other notes of impor- tance. The preserved specimens were counted under a stereo-zoom microscope (Leica-MS5). Spiders of all life stages were collected during sampling. The spiders were identified to the SEBASTIAN ET AL.-SPIDERS OF RICE ECOSYSTEM IN KERALA, INDIA 249 S Others 1 Space web builders B Ground runners 0 Stalkers S Orb weavers Sites Figure 2. — Percentage of spider species by feeding guilds collected from irrigated rice ecosystem in central Kerala, India. 1 = Parakkadavu, 2 = Piravom, 3 = Vannappuram, 4 = Kothamangalam, 5 = Adimali, 6 = Marayoor, 7 = Total. species level with the help of available liter- ature (Tikader 1987; Barrion & Litsinger 1995) except the immature ones, which could be identified only to the generic level. Vouch- er specimens were deposited in a reference collection housed with the Arachnology Di- vision, Dept, of Zoology, Sacred Heart Col- lege, Cochin, Kerala, India. Data Analysis. — The diversity of spiders across different elevations were analyzed by widely used indices viz., the Shannon- Wiener index, which is sensitive to changes in the abundance of rare species in a community, and the Simpson index, which is sensitive to changes in the most abundant species in a community (Solow, 1993). Shannon-Wiener index is defined as: H' = -y p,iog p, where p,- = n/N is the observed relative abun- dance of a particular species; n,- = the number of individuals of species i, and N = Sn,-. The Simpson index is defined as: 2 - 1) D = N(N - 1) The results are given as 1-D. The Margalef index, a species richness in- dex, was computed based on the relationship between species richness (S) and total number of individuals observed (N) which increases with increasing sample size. The Margalef in- dex: R1 = S-l/ln N. The evenness index is a measure of how evenly species are distributed in a sample. When all species in a sample are equally abundant an evenness index will be at its max- imum, decreasing towards zero as the relative abundance of the species diverge away from evenness. The modified HilTs ratio (E5) is the best evenness index, the least ambiguous, the most easily interpreted and is independent of the number of species in the sample (Ludwig & Reynolds 1988): E5 = (1/D)- l/eH'-l where D = Simpson's index and H' = Shan- non-Wiener index. The Shannon- Wiener, Simpson, Margalef and Evenness (E5) indices were computed us- ing the statistical software, SPDIVERS.BAS of Ludwig & Reynolds (1988). The estimated species richness was calcu- lated to determine whether or not the environ- ment had been sufficiently sampled. The Chao 1 quantitative estimator (Chao 1984; Colwell & Coddington 1994) is calculated as: Schaoi = + (aV2b) 250 THE JOURNAL OF ARACHNOLOGY Table 1. — Total number of families, genera, and species composition of spiders sampled from different localities of central Kerala. 1 Parakkadavu, 2 = Piravom, 3 = Vannappuram, 4 = Kothamangalam, 5 = Adimali, 6 = Marayoor. Family Genera Species Individuals Lowland 1 2 Localities Midland 3 4 Highland 5 6 Araneidae 11 25 132 9 5 13 12 13 9 Clubionidae 1 2 4 1 1 1 1 Eresidae 1 1 1 1 Filistatidae 1 1 1 1 Gnaphosidae 1 1 2 1 1 Hersiliidae 1 1 2 1 1 Linyphiidae 1 1 59 1 1 1 1 1 Lycosidae 3 7 107 6 4 6 6 6 5 Miturgidae 1 1 2 1 Oxyopidae 2 9 171 6 2 8 6 3 Salticidae 10 14 115 10 3 5 10 5 Sparassidae 1 1 1 1 Tetragnathidae 4 16 437 12 4 10 13 13 7 Theridiidae 7 8 85 7 5 3 3 2 3 Thomisidae 1 2 4 2 1 1 Uloboridae 1 2 7 1 Total 47 92 1130 58 27 47 54 44 28 where is the number of species observed; a is the number of singletons and b is the number of doubletons. The Estimates pro- gram (Colwell 2000) was used to calculate the Chao 1. The degree of association or similarity of the sampling sites was investigated using clus- ter analysis. The term “cluster analysis” en- compasses a number of different classification algorithms (Faith 1991). It is a useful data re- duction technique that can be helpful in iden- tifying patterns and groupings of objects. The program CLUSTER. BAS (Ludwig & Reyn- olds 1988) was used for the cluster analysis of the data from different localities using the flexible strategy (Lance & Williams 1967) and chord distance, a measure of dissimilarity. RESULTS Distribution. — Spiders representing 16 families, 47 genera and 92 species were re- corded during the study (Appendix 1). Table 1 is a summary of family composition. The sampling yielded a total of 1130 individuals. Some families were widely distributed throughout the study area while others were restricted to one or a few localities. The wide- ly distributed families were Araneidae, Lycos- idae, Tetragnathidae and Salticidae. Family Araneidae was represented by 25 species be- longing to 1 1 genera. However, the numeri- cally dominant family was Tetragnathidae with a collection of 437 individuals belonging to 16 species and 4 genera. The numerically most abundant species was Tetragnatha man- dibulata Walckenaer 1842 (Family Tetrag- nathidae) with a total of 109 individuals (Ap- pendix 1). T. javana (Thorell 1890), T. cochinensis Gravely 1921, Tetragnatha sp., Pardosa pseudoannulata (Bosenberg & Strand 1906), Pardosa sp., Lycosa tista Ti- kader 1970, Argiope sp., and Chrysso argy- rodiformis (Yaginuma 1952) were present at all study sites. Diversity. — -Diversity measurements did not vary considerably between sites across elevational gradients (Table 2) although the Parakkadavu site in the lowland recorded the highest Shannon-Wiener (3.49), Simpson (0.96), Margalef (6.83) and richness (58) val- ues. The Evenness index E5 was the highest at Marayoor (0.90). The Chao 1 species rich- ness estimator generated species richness val- ues which were higher than the actual richness values. The highest value was observed at Parakkadavu (91.1), whereas the actual rich- ness value at this site was 58. SEBASTIAN ET AL.-SPIDERS OF RICE ECOSYSTEM IN KERALA, INDIA 251 Table 2. — Species diversity measures of spiders in rice ecosystem sampled from different localities in central Kerala. 1 = Parakkadavu, 2 = Piravom, 3 = Vannappuram, 4 = Kothamangalam, 5 = Adimali, 6 = Marayoon Lowland Midland Highland Diversity measures 1 2 3 4 5 6 Shannon- Wiener index (H') 3.49 2.82 3.22 3.46 3.17 3.01 Simpson index (1-D) 0.96 0.92 0.93 0.95 0.93 0.94 Margalef index (Rl) 6.83 3.92 6.00 6.71 5.73 4.21 Evenness index (E5) 0.86 0.85 0.84 0.87 0.84 0.90 Richness (S) 58 27 47 54 44 28 Chao 1 91.1 33.9 61.3 68.3 68.4 41.2 Cluster Analysis.“=“The pattern of cluster- ing for the six localities is summarized in the dendrogram in Fig. 1. The species level anal- ysis revealed two main clusters. The first clus- ter included four sites; Vannappuram and Kothmangalam (mid land) and Adimali and Marayoor (high land). At a chord distance of approximately 2.5, two groups emerged from this cluster. The first group was formed of the two high land sites while the second group included the two mid land sites. The second cluster was occupied by the remaining two low land sites, Parakkadavu and Piravom. Guild Structure Analysis. — The spiders sampled belong to seven different foraging guilds (Uetz et al. 1999). These guilds are orb weavers, stalkers, ground runners, space web builders, sheet web builders, foliage runners, and ambushers (Fig. 2). Even though substrata for anchoring the webs are limited in rice fields compared to other terrestrial habitats, the orb weavers dominated in all the locations constituting 51% of the total collection. Stalk- ers were also seen in abundance (26%). Ground runners and space web builders were represented by 9% and 8%, respectively. Space web builders, foliage runners, and am- bushers were less common in the study area. DISCUSSION The sixteen spider families recorded from the rice fields of Central Kerala represent 37% of the families reported from the country (Ti- kader 1987). A total of 92 spider species were reported from the rice ecosystem of central Kerala using two identification keys. Use of only two keys provided by Tikader (1987) and Barrion & Litsinger (1995) is justified as they are sufficient to identify spiders found in pen- insular India. In a similar study, Bambarade- niya & Edirisinghe (2001) have documented 60 species of spiders from an irrigated rice field ecosystem in Sri Lanka. Other works in Southeast Asia include that of Heong et al. (1991) recording 46 species of predators in- cluding bugs and spiders in Philippine rice fields and Barrion & Litsinger (1995) record- ing about 342 species of spiders from rice fields in the Philippines and other Southeast Asian countries. The observation that Aranei- dae and Tetragnathidae are the largest families is in conformity with the findings of Bambar- adeniya & Edirisinghe (2001) in the rice fields of Sri Lanka. The dominance of tetragnathid spiders in the rice ecosystem of central Kerala might be expected as this wet habitat provides the congenial habitat for this family. An analysis of various diversity indices across different elevations yielded only mini- mal differences in most of the indices used. This suggests that the effect of elevation on the diversity of spiders is not very drastic in the rice ecosystems of central Kerala. None- theless, Parakkadavu, which is a lowland area, exhibited more overall species richness and diversity. There are many environmental fac- tors that affect species diversity. Some of these factors include seasonality, spatial het- erogeneity, competition, predation, habitat type, environmental stability and productivity (Rosenzweig 1995). In terrestrial environ- ments, a decrease in species richness with el- evation and latitude is a common phenome- non. High elevation communities almost invariably occupy smaller areas than lowlands and they will usually be more isolated from similar communities than lowland sites. The effects of area and isolation are certain to con- tribute to the decrease in species richness with 252 THE JOURNAL OF ARACHNOLOGY elevation (Townsend et al. 2002). This ex- plains the high overall species richness and diversity of Parakkadavu in comparison with other sites selected for the study. In addition, at Parakkadavu, rice cultivation is practiced in a cyclical way between polycultures of banan- as and vegetables. This practice will provide enough shelter for spiders in different seasons. However, reasons for the low diversity indices recorded in the other lowland site, Piravom, need to be determined. One reason could be the practice of monoculture prevailing in this locality. From the dendrogram, it is evident that the localities belonging to the same elevation formed one group in the cluster analysis. The midland and high range sites were found to be similar in the occurrence of species. Clus- tering revealed that the two low land sites be- haved as a separate entity from the rest of the sites in species composition. This trend was predictable also because the distance between the mid and high land sites were less than that between the mid and low land sites. The maximum number of species estimated by Chao 1 quantitative estimator showed wide deviations from the observed number of spe- cies. Chao 1 is an abundance-based estimator, so the number of times a species is present in a sample set has a significant effect on the number of species estimated to be present. This explains why Chao 1 gave a larger esti- mate of the overall species richness in the se- lected sites. Also, the presence of singletons and doubletons caused the Chao 1 to behave erratically. High relative percentages of sin- gletons and doubletons during a sampling pe- riod indicate low abundance with Chao 1 (Fassbender 2002). The difference in estimat- ed and observed numbers of species using Chao 1 reveals that the sampling efforts used were inadequate to reveal the true species di- versity at the sites and indicates the necessity to conduct more intensive studies with modi- fication of sampling techniques, viz. including extended sampling time, sampling during dif- ferent time periods of the day, etc. The most common explanation for the ob- served pattern of spider guild structure is the nature of the host crop, including its structural diversity, microenvironment, or the level of disturbance. Ample observations and more re- cent experimental evidence suggest that hab- itat structure maintains a diverse spider as- semblage (Uetz 1991) and may be critical to successful insect pest suppression. The struc- tural complexity may determine the guild composition of a crop’s spider fauna and in- directly influence the level of herbivore dam- age. The rice ecosystem of central Kerala has a diverse spider community and further research is indicated to evolve a better understanding of their ecology. These studies should include exploring other factors which are important in influencing spider diversity and richness in this agroecosystem, viz. effects of insecti- cides, availability of prey species, intra- and interspecific competition, surrounding habitats and climatic factors. ACKNOWLEDGMENTS The authors are thankful to the Indian Council of Agricultural Research (ICAR), New Delhi for financial assistance. The au- thors would also like to thank Rev. Fr. A.J. Saviance C.M.L, Principal, Sacred Heart Col- lege, The vara, Kochi, Kerala, India for pro- viding the laboratory facilities. LITERATURE CITED Anbalagan, G. & R Narayanaswamy. 1999. Popu- lation fluctuation of spiders in the rice ecosystem of Tamil Nadu. Entomon 24(l):91-95. Bambaradeniya, C.N.B. & J.R Edirisinghe. 2001. The ecological role of spiders in rice fields of Sri Lanka. Biodiversity 2(4):3-10. Banerji, D.K., RK. Nanda., P.K. Bera & S.C. Sen. 1993. Seasonal abundance of some important spider groups in rice agro-ecosystem. Records of Zoological Survey of India 93(l-2):275-28L Barrion, A.T & J.A. Litsinger. 1995. Riceland Spi- ders of South and Southeast Asia. CABI Inter- national. 736 pp. Bhattacharya, S. 2000. Biodiversity of spiders in the rice fields of Kalyani, West Bengal, India. Research Journal of Chemistry and Environment 4(2):75. Chao, A. 1984. Non-parametric estimation of the number of classes in a population. Scandinvian Journal of Statistics 4:65-270 Chiu, S.C. 1979. Biological control of brown plant hopper. In Brown Plant Hopper, Threat to Rice Production in Asia. International Rice Research Institute, Los Banos, Laguna, Philippines, pp. 335-355. Colwell, R. 2000. Estimates. Statistical estimation of species richness and shared species from sam- ples. Version 6. Obi. http://viceroy.eeb. uconn. edu/estimates. Colwell, R.K. & J.A. Coddington. 1994. Estimating SEBASTIAN ET AL.-SPIDERS OF RICE ECOSYSTEM IN KERALA, INDIA 253 terrestrial biodiversity through extrapolation. Philosophical Transactions of the Royal Society of London B 345:101-118. Faith, D.R 1991. Effective pattern analysis methods for nature conservation. Pp. 47 — 53. In Nature Conservation: Cost Effective Biological Surveys And Data Analysis (C.R. Margules & M.P. Aus- tin, eds.). CSIRO, Australia. Fassbender, J.L. 2002. Litter and ground dwelling spiders of mixed mesophytic forests in southeast Louisiana. M.S, Thesis submitted to Louisiana State University and Agricultural and Mechani- cal College. 72pp. Gavarra, M. & R.S. Raros. 1973. Studies on the biology of predator wolf spider, Lycosa pseu- doannulata Bos. et Str. (Araneae: Lycosidae). Philippine Entomologist 2(6):427. Hamamura, T 1969. Seasonal fluctuation of spider populations in paddy fields. Acta Arachnologia 22(2):40-50. Heong, K.L., G.B. Aquino & A.T Barrion. 1991. Arthropod community structures of rice ecosys- tems in the Philippines. Bulletin of Entomolog- ical Research 81:407-416. Holt, J., A.J. Cook, TJ. Perfect & G.A. Norton. 1987. Simulation analysis of brown plant hopper (Nilaparvata lugens) population dynamics on rice in the Philippines. Journal of Applied Ecol- ogy 24:87-102. Kobayashi, S. 1977, Changes in population density of the spiders in paddy fields during winter. Acta Arachnologia 27:247-251. Lance, G.N. & W.T. Williams. 1967. A general the- ory of classificatory sorting strategies. 11. Clus- tering Systems. The Computer Journal 10(3): 271-276. Ludwig, J.A. & J.F. Reynolds. 1988. Statistical Ecology: A primer on Methods in Computing. John Wiley & Sons, New York. 337pp. Nyffeler, M. & G. Benz. 1987. Spiders in natural pest control: a review. Journal of Applied Ento- mology 103:321-329. Nyffeler, M., W.D. Sterling & D.A. Dean. 1994. How spiders make a living. Environmental En- tomology 23:1357-1367. Pathak, S. & N.N. Saha. 1999. Spider fauna of rice ecosystem in Barak valley zone of Assam, India. Indian Journal of Entomology 2:21 1-212. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cam- bridge, UK. 458pp. Samal, P. & F.B.C. Misra. 1975. Spiders: the most effective natural enemies of the brown plant hop- pers in rice. Rice Entomology Newsletter 3:31. Sasaba, T, K. Kiritani & S. Kawahara. 1973. Food preference of Lycosa in paddy fields. Bulletin of the Kochi Institute of Agriculture and Forestry Science 5:61-63. Solow, A.R. 1993. A simple test for change in com- munity structure. Journal of Animal Ecology 62: 191-193. Sunderland, K. 1999. Mechanisms underlying the effects of spiders on pest populations. Journal of Arachnology 27:308-316. Tanaka, K. 1989. Movement of spiders in arable land. Plant Protection 43(l):34-39. Tikader, B.K. 1987. Handbook of Indian Spiders. Zoological Survey of India, Calcutta, India. 251pp. Townsend, C.R., M. Begon & J.L. Harper. 2002. Essentials of Ecology (2"^ Ed). Blackwell Pub- lishing, Malden, USA. 544pp. Turnbull, A.L. 1973. Ecology of the true spiders (Araneomorphae). Annual Review of Entomol- ogy 18:305-348. Uetz, G.W. 1991. Habitat structure and spider for- aging. Pp. 325-348. In Habitat Structure; the Physical Arrangement of Objects in Space (E.D. McCoy, S.A. Bell, & H.R. Mushinsky, eds.). Chapman and Hall, London. Uetz, G.W, J. Halaj & A.B. Cady. 1999. Guild structure of spiders in major crops. Journal of Arachnology 27:270-280. Watanabe, I. & PA. Roger. 1985. Ecology of flood- ed rice fields. Pp. 230-242. In Wetland Soils: Characterization, Classification and Utilization. International Rice Research Institute, Philip- pines. Manuscript received 12 January 2005, revised 12 September 2005. 254 THE JOURNAL OF ARACHNOLOGY Appendix 1. — Abundance data (total catches of two seasons) for spiders of rice ecosystem of central Kerala. 1 = Parakkadavu, 2 = Piravom, 3 = Vannappuram, 4 = Kothamangalam, 5 = Adimali, 6 = Marayoor. Species name Lowland 1 2 Localities Midland 3 4 Highland 5 6 Total Family Araneidae Araneus sp. 8 2 3 8 21 Araneus bilunifer Pocock 1900 1 1 Araneus ellipticus (Tikader & Bal 1981) 1 3 1 5 Argiope sp. 2 2 1 3 1 2 11 Argiope aemula (Walckenaer 1842) 3 2 1 1 1 8 Argiope anasuja Thorell 1887 3 3 Argiope catenulata (Doleschall 1859) 3 1 1 3 2 10 Argiope pulchella Thorell 1881 1 2 1 4 Chorizopes sp. 1 1 Cyclosa sp. 1 3 4 Cyclosa bifida (Doleschall 1859) 1 1 Cyclosa fissicauda Simon 1889 1 1 2 Cyrtarachne sp. 1 1 Cyrtophora sp. 3 1 5 2 11 Cyrtophora citricola (Forskal 1775) 1 1 Eriovixia sp. 4 4 Eriovixia laglaizei (Simon 1877) 3 3 Eriovixia excelsa (Simon 1889) 2 1 2 5 Gasteracantha geminata (Fabricius 1798) 2 1 3 6 Gibbaranea bituberculata (Walckenaer 1802) 2 2 Neoscona sp. 6 3 7 4 20 Neoscona bengalensis Tikader & Bal 1981 1 1 Neoscona molemensis Tikader & Bal 1981 1 1 1 1 4 Neoscona vigilans (Blackwall 1865) 1 1 Zygiella sp. 1 1 2 Family Clubionidae Clubiona sp. 1 1 2 Clubiona drassodes O. P. Cambridge 1874 1 1 2 Family Eresidae Stegodyphus sarasinorum Karsch 1891 1 1 Family Filistatidae Pritha sp. 1 1 Family Gnaphosidae Gnaphosa sp. 1 1 2 Family Hersiliidae Hersilia savignyi Lucas 1836 1 1 2 Family Linyphiidae Linyphia sp. 22 16 16 2 3 59 Family Lycosidae Hippasa sp. 1 2 2 2 3 10 Lycosa sp. 6 2 2 8 1 7 26 Lycos a tista Tikader 1970 1 1 1 1 4 Pardosa sp. 2 2 2 2 1 2 11 Pardosa pseudoannulata (Bbsenberg & Strand 1906) 1 1 1 3 1 8 15 Pardosa minuta Tikader & Malhotra 1976 1 1 Pardosa sumatrana (Thorell 1890) 11 1 8 9 11 40 Family Miturgidae Cheiracanthium melanostomum (Thorell 1895) 2 2 Family Oxyopidae Oxyopes sp. 22 6 36 12 3 79 Oxyopes ashae Gajbe 1999 11 8 8 4 31 SEBASTIAN ET AL.-SPIDERS OF RICE ECOSYSTEM IN KERALA, INDIA 255 Appendix 1. — Continued. Localities Lowland Midland Highland Species name 1 2 3 4 5 6 Total Oxyopes bharatae Gajbe 1999 1 1 2 Oxyopes sakuntalae Tikader 1970 3 3 Oxyopes shweta Tikader 1970 2 1 4 7 Oxyopes sitae Tikader 1970 1 1 2 Oxyopes sunandae Tikader 1970 11 13 9 33 Peucetia sp. 2 2 Peucetia viridana (Stoliczka 1869) 2 4 5 1 12 Family Salticidae Bavia sp. 1 1 2 Bianor sp. 6 1 1 8 Hasarius adansoni (Audouin 1826) 4 4 Hyllus sp. 6 1 3 2 12 Myrmarachne orientates Tikader 1973 1 1 2 Phintella sp. 2 2 Phintella vittata (C. L. Koch 1846) 2 2 3 7 Plexippus sp. 2 1 2 4 3 12 Plexippus paykulli (Audouin 1826) 1 4 5 Plexippus petersi (Karsch 1878) 18 10 2 15 6 51 Telamonia sp. 1 1 Telamonia dimidiata (Simon 1899) 1 2 3 Thiania sp. 1 1 Thyene sp. 5 5 Family Sparassidae Heteropoda venatoria (Linnaeus 1767) Family Tetragnathidae • 1 Dyschiriognatha dentata Zhu & Wen 1978 6 1 2 1 10 Leucauge sp. 7 3 2 3 15 Leucauge bituberculata Baert 1987 1 1 1 1 4 Leucauge celebesiana (Walckenaer 1842) 2 1 1 4 Leucauge decorata (Blackwall 1864) 2 6 8 Leucauge pondae Tikader 1970 13 2 11 4 30 Tetragnatha andamanensis Tikader 1977 18 6 8 3 1 36 Tetragnatha ceylonica O. P. Cambridge 1869 4 12 1 4 21 Tetragnatha cochinensis Gravely 1921 12 9 17 4 6 1 49 Tetragnatha fletcheri Gravely 1921 4 4 Tetragnatha javana (Thorell 1890) 19 11 8 12 8 7 65 Tetragnatha mandibulata Walckenaer 1842 32 12 10 36 19 109 Tetragnatha maxillosa Thorell 1895 1 2 1 1 5 Tetragnatha sp. 16 9 14 26 6 71 Tetragnatha vermiformis Emerton 1884 4 4 Tylorida culta (O. P. Cambridge 1869) Family Theridiidae 2 2 Achaearanea sp. 2 2 2 22 28 Achaearanea durgae Tikader 1970 2 2 Argyrodes sp. 1 1 1 3 Chrysso argyrodiformis (Yaginuma 1952) 9 1 7 16 4 37 Dipoena sp. 1 1 2 Phycosoma martinae (Roberts 1983) 1 1 2 Theridion sp. 4 3 1 1 9 Theridula sp. Family Thomisidae 1 2 Thomisus sp. 1 1 1 3 Thomisus andamanensis Tikader 1980 1 1 Family Uloboridae Uloborus sp. 2 1 3 Uloborus danolius Tikader 1969 4 4 Total 325 99 203 238 180 85 1130 2005. The Journal of Arachnology 33:256-268 ECOLOGICAL PROFILES OF HARVESTMEN (ARACHNIDA, OPILIONES) FROM VITOSHA MOUNTAIN (BULGARIA): A MIXED MODELLING APPROACH USING GAMS Plamen G. Mitov: Department of Zoology and Anthropology Faculty of Biology, University of Sofia, 8 Dragan Tsankov Blvd., 1 164-Sofia, Bulgaria. E-mail: pLmitov @ yahoo.com Ivailo L. Stoyanov: Biodiversity Department, Central Laboratory of General Ecology, 2 Yurii Gagarin Street, 1113-Sofia, Bulgaria ABSTRACT. The present study is based on a large-scale sampling program carried out in the area of Vitosha Mountain (Bulgaria). The ecological prohles of the Opiliones inhabiting the investigated area are modelled by a mixed approach, using Generalized Additive Models (GAMs) over a Multiple Correspon- dence Analysis (MCA, performed on the sites by environmental variables matrix) ordination plot. Ac- cording to the literature data describing the harvestmen species from Vitosha Mountain, the most important factor determining the ecological classihcation of the Opiliones is the habitat type. The modelled ecological prohles revealed that the elevation contributes the most to the ecological characterization of the Vitosha harvestmen species, followed by the habitat type and moisture regime of the sampling localities. Few harvestmen species demonstrate preferences to the middle- and high-mountain zones, while the majority of harvestmen species are conhned exclusively to the low-mountain zone. The different species showed different responses (most of them were linear, not unimodal) towards the environmental variables. Keywords: Ecological type, ecological classihcation, Opiliones, Bulgaria, Generalized Additive Models, Multiple Correspondence Analysis Traditionally, studies on Opiliones in Bul- garia have been predominantly faunistic and taxonomic (Star^ga 1976; Martens 1978; Be- ron & Mitov 1996; Mitov 1987, 1994, 1995a, 1997a, 2001, 2002, 2003, 2004; Mitov & Stoyanov 2004), while data about the biology and ecology of the group are comparatively scarce and scattered through the faunistic lit- erature (Star^ga 1976; Martens 1978; Mitov 1986, 1995b,c, 1996, 1997b, 2000, in press; Mitov & Stoyanov 2004). Even on a European scale the ecological research on this animal group remains insufhcient. However, the works of Todd (1949), Pabst (1953), Pfeifer (1956), Williams (1962), Kolosvary (1966b), Tischler (1967), Weiss (1975, 1978, 1984, 1996), Obrtel (1976), Curtis (1978), Weiss & Sarbu (1977), Hiebsch (1978), Bliss & Tietze (1984), Klimes & Spicakova (1984), Klimes (1987, 1990, 2002), Sechterova (1989), Platen (1991, 1996, 2000), Simon (1995), Novak et al. (2004), Komposch & Gruber (1999), Lym- berakis et al. (in press) are specially dedicated to various aspects of the harvestmen ecology. Additional ecological notes may be found in the mostly faunistic studies of Cirdei & Bu- limar (1960), Hiebsch (1972), Meijer (1972), Thaler (1979), Muller (1984), Komposch (1995, 1997a,b,c, 1999, 2000, 2001, 2004), Platen et al. (1991), Platen & Broen (2002), Metzen & Cdlln (1998), Komposch & Gruber (2004), in annotated species lists. Many of these publications contain ecolog- ical classihcations of harvestmen species, based on their affinities towards certain envi- ronmental conditions. These classifications are often based on the subjective evaluation of the author. For example, Lophopilio palpinalis (Herbst 1799) has been described as “steno- topic?” (Komposch 1997a) on one hand, and as “moderately eurytopic” and “vertical-ubi- quistic” (Kolosvary 1965; Komposch 1999) on the other. Further, investigators go into even more detail by classifying this species also as “hemiombrophilous/ombrophilous” (Pfeifer 1956; Weiss 1975; Bliss & Tietze 256 MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 257 1984; Mitov & Stoyanov 2004), “psychro- philous” (Mitov & Stoyanov 2004), “hemihy- grophilous/hygrophilous forest form” (Weiss 1975; Martens 1978; Geyer 1983; Bliss & Tietze 1984; Muller 1984; Platen et al. 1991, 1996, 2000; Komposch 1997a,b, 1999; Metz- en & Colin 1998; Platen & Broen 2002). These categorizations may have a significant empirical background, but often the ecological type of a species is confusing without detailed reference to the analytical procedures that led to them. So the above mentioned discrepan- cies might be a manifestation of Kiihnelt’s principle of regional stenoecy (e. g. Kiihnelt 1965) or due to subjective error. Recently, with the development of more elaborate mod- elling techniques that permit a direct relation between the species and their environment, at- tempts have been made to directly explore the responses of harvestmen species to various environmental variables by employing multi- variate techniques (Klimes 1997; Muster 2001). Subjecting a significant amount of data from the Czech Republic to multivariate an- alytical procedures (such as TWINSPAN and CCA), Klimes (1997) concluded that the main factors explaining the variation of the data were elevation, temperature and human im- pact. Muster (200 1 ) found that the harvestmen from the central part of N Alps were also mostly affected by elevation and light condi- tions. Both these works employed Canonical Correspondence Analysis (CCA; ter Braak 1987), a widely used method for direct (“con- strained”) gradient analysis, that assumes spe- cies to have unimodal distributions along en- vironmental gradients, but none of them tested if this crucial assumption of CCA was met by the data. Consequently, the interpretations may be influenced by potential non-unimodal species responses. Nevertheless, these works may be regarded as first attempts to put the relationships between harvestmen and their habitat on an objective basis. In view of these considerations, the present work will aim at contributing further to the knowledge of aut- ecological features of the opilionid species from the Vitosha Mountains (the region in Bulgaria with the most fully studied opilionid assemblages; see Mitov 2000) by direct mod- elling the response of each species towards an array of environmental factors. Utilizing an extensive data set from a large- scale sampling program, we will focus on: 1) summarizing the main environmental varia- tion across the sampling localities, 2) directly modelling the response of every collected har- vestmen species to the summarized multivar- iate gradient by using the power and flexibility of Generalised Additive Models (GAM), 3) classifying the observed response patterns of the opilionid species and 4) comparing the ecological profiles (obtained in the previous modelling stage) with published ecological data. METHODS Material collected. — The present study is based on the examination of 31,639 specimens (8,314 males, 14,861 females, 8,464 juve- niles) from 22 species and subspecies (see the number of each species caught in legend of Fig. 3), collected by the senior author in the period 28 February 1987-28 April 1989 in the region of Vitosha Mountain (peak Cherni Vrakh: UTM FN81, N 42°33'48.9", E 23°16'45.2", 2290 m). Four species recorded from the area of the mountain (Star^ga 1976; Mitov 2000; Stoyanov & Mitov 2004) are not included in the present analysis. These are: Dicranolasma scabrum (Herbst 1799), Histri- costoma drenskii Kratochvil 1958, Opilio par- ietinus (De Geer 1778), and Rafalskia olym- pica (Kulczynski 1903). The former three species were absent from pitfall trap samples or were represented by only a few {n < 10) individuals, while the latter species has not been recorded from Vitosha Mountain since the original record of Star^ga (1976). The col- lected material is in the opilionid collection of Plamen Mitov. Sampling. — Altogether 653 pitfall traps (plastic buckets with rim diameter 10 cm and 12 cm height), filled with 4% formalin solu- tion were used. The traps were placed 5 m apart, on a zig-zag line through 54 sampling localities, at elevations between 750 and 2290 m (the latter is the maximum elevation for this dome-like mountain), and at intervals between 200 and 500 m (depending on the relief). All major habitat types (approximately 40% of the habitats, according to Dr. Rosen Tsonev, pers. comm.) were sampled during the sampling program that covered the whole area (278 km^) of Vitosha Mountain. Samples were col- lected monthly. For further details on the sam- pling scheme see Mitov (1996). Environmental data. — At each sampling 258 THE JOURNAL OF ARACHNOLOGY Figure 1. — Ordination diagram of the Multiple Correspondence Analysis (MCA) of the full habitat by environmental variables matrix. The first axis summarizes 11.6%, the second = 9.0% of the variability. Dots represent the sampling localities; environmental variables are shown in Table 1, ellipses visualize the spread of environmental variable modalities, CRl and CR2 are the correlation ratios of each variable related to the first and second ordination axis. locality the following environmental variables were measured and recorded (see Table 1): elevational (climatic) zone (3 classes, ordered; classification in Hubenov 1990), geographical exposition (12 classes, ordered) measured with compass, habitat type (8 classes), humid- ity (7 classes, ordered; based on indicator plants according to Nedyalkov 1998), light conditions (2 classes, ordered; based on hab- itat type), vegetation belt (4 classes; classifi- cation in Hubenov 1990), soil type (6 classes; classification in Chucheva 1983), and tilt (6 classes, ordered; classification in Chucheva 1983) measured with an standard plastic angle meter. Data analyses. — The following procedure was used for modelling; 1) a Multiple Corre- spondence Analysis (MCA, Tenenhaus & Young 1985) was performed on the locali- ties by environmental variables matrix to ob- tain a low-dimensional representation of the data structure; 2) the resulting sample ordi- nation space (2 retained axes) was overlayed with the fitted Generalized Additive Model (GAM) surface of opilionid species abun- dance at the ordinated sampling sites, where the poisson error distribution and logarithmic “link function” were used for fitting. The ad- vantage of using GAMs for the modelling is, that it is especially powerful in modelling data with non-normal error distributions (Hastie & Tibshirani 1990; Wood 2000), and that one does not have to assume a particular (unimo- dal or linear) response of species abundance along the environmental gradient, and thus the exploratory phase of the investigation is more flexible. As only 9 harvestmen species were more widespread through the area of Vitosha Moun- tain, after modelling their abundance the site by environmental variables matrix was re- duced to increase the resolution when mod- elling the data for the rest of the opilionid spe- cies. Fourteen high-mountain sites (in the MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 259 Figure 2. — Ordination diagram of the Multiple Correspondence Analysis (MCA) of the reduced habitat by environmental variables matix (the high-mountain sites excluded). The first axis summarizes 12.2%, the second = 10.2% of the varibility. Dots represent the sampling localities. For the environmental variable modalities see Table 1, ellipses visualize the spread of environmental variable modalities, CRl and CR2 are the correlation ratios of each variable related to the first and second ordination axis. right half of Fig. 1) were removed, in order to exclude the sites where the rest of opilionid species (13) were not (or only occasionally) present. The resulting matrix (Fig. 2) was again subjected to MCA to obtain the sam- pling site ordination, over which the GAM surface-fitting procedure for the remaining (i.e. those restricted to the low-mountain zone) species was applied again. All compu- tations were performed in the R statistical lan- guage and environment (Ihaka & Gentleman 1996), using the ade4 (Chessel et al. 2004), mgcv (Wood 2000), and akima (Akima 1978) libraries. RESULTS Environmental gradients. — The MCA on the sampling site by environmental variable matrix shows a strong separation of the low- mountain from the high-mountain zone (in- cluding the a priori defined middle-mountain zone, cf. “METHODS”) along the first ordi- nation axis (see Fig. 1, “elevational zone”). The environmental variables with high corre- lation ratio with this axis (and thus enabling a good separation of the sampling sites along that axis) also include soil type, vegetation belt (both presenting a structure very similar to that of the elevational zone), and exposi- tion. Most strongly associated with the second ordination axis are the following variables: vegetation belt, habitat type, soil type, and light conditions (Fig. 1). The similar patterns of soil type, elevational zone and vegetation belt are due to strong interdependence be- tween these factors. When analyzing the reduced sites by envi- ronmental variables matrix (Fig. 2), a not so abrupt (and hence more complex) gradient is apparent. Its first axis is mainly determined by the moisture gradient, the habitat type, and ex- position, while the exposition, tilt, habitat and moisture (both similarily important) summa- rize the main variation along the second or- 260 THE JOURNAL OF ARACHNOLOGY Table 1. — Environmental variables measured at each sampling locality. Variable Classes Elevational zone 1) low- (up to 1450 m), 2) middle- (1450-1850 m), 3) high-moun- tain zone (above 1850 m, max. 2290 m) Geographical exposition 1) N, 2) NNE, 3) NE, 4) ENE, 5) E, 6) SE, 7) SSE, 8) S, 9) SW, 10) WSW, 11) W, 12) NNW Habitat type 1) coniferous forests, 2) deciduous forests, 3) rivulet-bank in conif- erous forests, 4) rivulet-bank in deciduous forests, 5) rivulet-bank through meadows, 6) meadows, 7) peat moss bogs, 8) forest- glades Humidity 1) dry, 2) dry-mesophilous, 3) dry-fresh, 4) mesophilous-fresh, 5) fresh, 6) fresh-moist, 7) moist Light conditions 1) dark, 2) light Vegetation belt 1) Quercus-Carpinus, 2) Fagus, 3) coniferous, 4) subalpine Soil type 1) Chromic Luvisols, 2) Distric Cambisols, 3) Humic Cambisols, 4) Orthic Umbrosols, 5) Rendzic Leptosols, 6) Histic Umbrosols Tilt 1) 0-5°, 2) 6-10°, 3) 11-20°, 4) 21-30°, 5) 31-40°, 6) 41-50° dination axis (see the correlation ratios in Fig. 2). Modelled ecological profiles* — As evident from the distribution plots (Fig. 3), the distri- bution-patterns of the Opiliones from Vitosha Mountain may be classified in two groups. The first one contains species with region- wide distribution (indicated by the spread of lines that connect the sampling sites where a species has been sampled): Pyza bosnica (Roewer 1919), Paranemastoma radewi (Roewer 1926), Paranemastoma aurigerum ryla (Roewer 1951), Phalangium opilio Lin- naeus 1758, Rilaena cf. serbica Karaman 1992, Lophopilio palpinalis (Herbst 1799), Lacinius horridus (Panzer 1794), Mitopus mo- rio (Fabricius 1779), and Leiobunum rumeli- cum Silhavy 1965. The second group include opilionid species restricted more or less to the low-mountain zone (the compact cluster, lo- cated left of the main vertical axis on Fig. 3). These are Mitostoma chrysomelas (Hermann 1804), Carinostoma ornatum (Hadzi 1940), Trogulus tricarinatus (Linnaeus 1767), T. do- sanicus Avram 1971, Opilio saxatilis C, L. Koch 1839, O. ruzickai Silhavy 1938, O. di- naricus Silhavy 1938, Rilaena balcanica SiL havy 1965, Zachaeus crista (Brulle 1832), Z. anatolicus (Kulczynski, 1903), Lacinius den- tiger (C.L. Koch 1847), L. ephippiatus (C.L. Koch 1835), Odiellus lendli (S0rensen 1894). The fitted GAM surfaces for some of the members of the first group mentioned above do not show any prominent optimum within the study area, as for example Pyza bosnica. Paranemastoma radewi, Lophopilio palpin- alis, and Mitopus morio (Fig. 4). These spe- cies increase their abundance towards the margin of the scatterplot more or less linearly. The first mentioned species has its maximum abundance in the low-mountain zone as well as in the coniferous forest habitats (in the mid- dle-mountain zone); the second species tends to occur more massively in deciduous forests (in the low-mountain zone), and the latter two reach highest numbers in the middle- and high-mountain zones respectively. Leiobunum rumelicum is mainly distributed in forest hab- itats (predominantly in the low-mountain zone and several occupied localities in the middle- mountain zone). Rilaena cf. serbica and Phal- angium opilio seem to prefer middle-mountain open habitats (where a well defined peak may be observed); a similar pattern is also dis- played by Paranemastoma aurigerum ryla, but the peak is not so prominent. Finally La- cinius horridus shows a clearly bimodal dis- tribution pattern, showing a prominent peak in forests of the low-mountain zone and increas- ing at the same time its abundance towards open habitats in the middle-mountain zone. From the predominantly low-mountain har- vestmen species, Zachaeus crista (Fig. 5) and Trogulus tricarinatus (not shown) are more or less evenly distributed within the zone. Rilae- na balcanica (Fig. 5), O. dinaricus, and Opilio ruzickai (both not shown) are clearly associ- ated with forests locations in the oak-horn- beam vegetation zone. The modelled respons- es of Carinostoma ornatum and Opilio MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 261 saxatilis (Fig. 5) show a clear preferendum (peak) towards relatively dry and open habi- tats, the peak of the latter species is more to- wards open and dryer (and not so slanted) stations (cf. the habitat characteristics distri- bution on Fig. 2). In contrast to the previously mentioned spe- cies, the following harvestmen do not show a pronounced optimum in their response. Lacin- ius dentiger (Fig. 5) and Lacinius ephippiatus (not shown) demonstrate a slightly bimodal response, being strongly associated with fresh to moist slanted forest habitats in both the beech and oak-hornbeam vegetation belt (the latter species being more dependent on mois- ture conditions, than the former). A somewhat bimodal, but not easy interpretable response pattern may be observed on the GAM plot for Trogulus closanicus (Fig. 5). This species seems to be associated with fresh to moist riv- erside habitats in forests and fresh meadows, but due to the relatively low number of indi- viduals collected, this pattern is not very well supported. Finally, the abundances of three of the op- ilionid species: Mitostoma chrysomelas, Za- chaeus anatolicus and Odiellus lendli, could not be modelled because of their very restrict- ed occurrence (i.e. very low frequency and abundance of catches) on Vitosha Mountain. The last mentioned species were collected mainly on a few meadows, and while Z. an- atolicus could be regarded as relatively rare throughout Bulgaria, O. lendli was locally very abundant (911 specimens come from a fresh beech forest meadow). When focusing on the response types of congeneric opilioeid species, we may observe that these species tend to display opposite trends, as for example the species of the gen- era Paranemastoma Redikorzev, 1936 (Figs. 3, 4), Lacinius Thorell, 1876 and Rilaena Sil- havy 1965. The differences are not so prom- inent in the responses of the Trogulus Latreille 1802 and Zachaeus C.L. Koch, 1839 species, while in species of the genus Opilio Herbst, 1798 only the response of O. saxatilis shows a trend opposite to the responses of the other species from this genus (Fig. 3). Ecological profiles from literature data*— When examining the published eco- logical profiles of harvestmen species, four groups can be delimited. 1. In the first one we include species that have repeatedly been reported to prefer moist habitats in forests. These are Paranemastoma radewi (Star^ga 1976; Mitov 1986, 1996), Pyza bosnica (Star^ga 1976; Mitov & Stoy- aeov 2004), Paranemastoma aurigerum ryla (see Star^ga 1976), Lophopilio palpinalis (Pfeifer 1956; Cirdei & Bulimar 1960; Hiebsch 1972; Weiss 1975; Star^ga 1976; Martens 1978; Geyer 1983; Bliss & Tietze 1984; Muller 1984; Platen et al. 1991; Platen 1996, 2000; Platen & Broen 2002; Komposch 1997a, b, 1999; Metzen & Colin 1998; Kom- posch & Gruber 2004; Mitov & Stoyaeov 2004; but see above for alternative opinions) and L. ephippiatus (Mitov & Stoyaeov 2004). Nevertheless, many European harvestmen re- searchers have described the latter as eurytop- ic (Platen et al. 1991; Platen 1996, 2000; Plat- en & Broen 2002; Komposch 1997a, 1999), hygrophilous (Martens 1978; Hiebsch 1978; Muller 1984; Platen et al. 1991; Karaman 1995; Komposch 1997a, 1999, 2001; Kom- posch & Gruber 2004), thermophilous (Pfeifer 1956), or as a montane forest species (Star^ga 1976). Despite the scarce information in the literature about the ecological status of Leio- bunum rumelicum, which only Star^ga (1976) reported as a species inhabiting montane for- ests, we add this species to the above men- tioned group. 2. According to the examined literature sources, most of the species found in the Vi- tosha Mountain seem to generally prefer ther- mophilous forests in the low-mountain zone. This group include Rilaena cf. serbica (only recently reported from Bulgaria by Mitov & Stoyanov 2004 who described it as thermoph- ilous forest-dweller), Lacinius horridus (Pfei- fer 1956; Stargga 1976; Martens 1978; Thaler 1979; Muller 1984; Platen et al. 1991; Weiss 1996; Karaman 1995; Metzen & Colin 1998; Komposch 1999; Platen & Broen 2002; Kom- posch & Gruber 2004; Mitov & Stoyaeov 2004), and Trogulus tricarinatus (Kolosvary 1965; Stargga 1976; Martens 1978; Platen et al. 1991; Karaman 1995; Weiss 1996; Kom- posch 1997a, 1999; Metzen & Colin 1998; Platen 2000; Muster 2001; Platen & Broen 2002; Komposch & Gruber 2004; Mitov & Stoyanov 2004; but Komposch & Gruber (2004) question its thermophily). As the rep- resentatives of genus Zachaeus C.L. Koch 1839 have been repeatedly classified as ther- mophilous (Martens 1978), it is understand- 262 THE JOURNAL OF ARACHNOLOGY Figure 3. — Distribution plot of the Opiliones from Vitosha Mountain. The space of sampling localities (dots) is the same as in Fig. 1; lines connect samples where each species is present with the centroid of the distribution; ellipses visualize the spread of individual species occurences. Species name abbreviations: Mit.chr (Mitostoma crysomelas, n = \1 sampled individuals), Car.orna (Carinostoma ornatum, n — 45), Pyz.bos (Pyza bosnica, n = 1844), Par.rad (Paranemastoma radewi, n = 774), Par.aur.ryl (P.aurigerum ryla, n = 318), Tro.tri (Trogulus tricarinatus, n = 89), Tro.clo {T. closanicus, n = 155), Pha.opi {Phal- angium opilio, n = 2875), Opi.sax (Opilio saxatilis, n = 103), Opi.ruz {O. ruzickai, n = 76), Opi.din {O. dinaricus, n = 318), Ril.bal (Rilaena balcanica, n = 996), Ril.sp {R. cf. serbica, n = 533), Lop. pal (Lophopilio palpinalis, n = 1881), Zac.cri (Zachaeus crista, n = 1431), Zac. ana (Z. anatolicus, n — 26), Lac.hor {Lacinius horridus, n — 12164), Lac. den (L. dentiger, n = 1950), Lac.eph (L. ephippiatus, n — 689), Odi.len {Odiellus lendli, n = 1002), Mit.mor (Mitopus morio, n = 4021), Lei. rum (Leiobunum rumelicum, n = 342). able that Zachaeus crista also falls into this group (Star^ga 1976; Weiss & Sarbu 1977; Martens 1978; Weiss 1975, 1996; Karaman 1995; Mitov 2003; Mitov & Stoyanov 2004). Here we include also Opilio ruzickai (Star^ga 1976; Komposch & Gruber 2004; Mitov & Stoyanov 2004), Opilio dinaricus (Komposch 1997a, 1999; Mitov & Stoyanov 2004), Rilae- na balcanica (Star^ga 1976; Mitov & Stoy- anov 2004) and Lacinius dentiger (Cirdei & Bulimar 1960; Star^ga 1976; Martens 1978; Thaler 1979; Karaman 1995; Komposch 1995, 1997a, 1999; Platen & Broen 2002; Kom- posch & Gruber 2004; Mitov & Stoyanov 2004). 3. The third group includes harvestmen that occur in forests, as well as in open habitats. These species have been frequently described as eurytopic, such as Mitopus morio (e. g. Cir- dei & Bulimar 1960; Tischler 1967; Star^ga 1976; Martens 1978; Geyer 1983; Muller 1984; Platen et al. 1991; Karaman 1995; Komposch 1997a,b, 1999; Metzen & Colin 1998; Zingerle 1999, 2000; Platen 2000, Mus- ter 2001; Platen & Broen 2002; Komposch & Gruber 2004), Mitostoma chrysomelas (Mar- tens 1978; Weiss 1984, 1996; Karaman 1995; Komposch 1997a,b; Metzen & Colin 1998; Zingerle 1999, 2000; Muster 2001; Komposch & Gruber 2004); the latter has also been de- scribed as euryphotic-hygrophilous (Hiebsch 1972) and forest hygrobiont/philic species, also inhabiting open habitats (Meijer 1972; Star^ga 1976; Platen et al. 1991; Platen 1996, MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 263 4k dp 'dsl Pyza bosnica m Paranemastoma aun gerum ryla [ Lophopilio palpinalis Plpy m A om A, m- 1 Mitopus morio Paranemastoma radt mi Ay. Leiobunum rumelicur n f 100 d id "V'i Rilaena cf. serbica Wi Phalangium opilio i Lacinius horridus ^ 200 — \cJ Figure 4.—GAM surface plots for the modelled abundance of harvestman species that occurred at all elevations: Pyza bosnica, Paranemastoma aurigerum ryla, Lophopilio palpinalis, Mitopus morio, Para- nemastoma radewi, Leiobunum rumelicum, Rilaena cf. serbica, Phalangium opilio, Lacinius horridus. The space of sampling localities (dots) is the same as in Fig. 1; the isolines show the modelled abundance of each species. 2000; Platen & Broen 2002). Carinostoma or- natum (Stargga 1976; Mitov 1986; Karamae 1995; Mitov & Stoyanov 2004), Trogulus do- sariicus (Weiss 1978, 1996; Komposch 1997a, 1999; Metzeri & Colin 1998; Mitov & Stoy- anov 2004) and the thermophilous, photophi- lous and xerophilous Odiellus lendli (Star^ga 1976; Weiss & Sarbu 1977; Mitov & Stoy- anov 2004) are also included in this group. 4. The last three species may be listed to- gether as harvestmen characteristic of open habitats. This group contains the relatively well known, ecologically widely adapted, photopMlous, and thermophilous species such as Phalangium opilio and Opilio saxatilis (Pfeifer 1956; Kolos vary 1965, 1966a,b; Star- ^ga 1976; Weiss & Sarbu 1977; Hiebsch 1978; Czechowski et ah 1981; Klimes 1987; Kuschka 1991; Platen et aJ 1991; Platen & Broen 2002; Karaman 1995; Weiss 1996; Komposch 1997a, 1999, 2001, 2004; Metzee & Colin 1998; Mitov 2003; Komposch & Gruber 2004; Mitov & Stoyanov 2004), as well as Zachaeus anatoUcus, a Balkan sub- endemic (Mitov 2004), that may also be in- cluded here, based on a single report about its thermophilc nature (Mitov 2001). DISCUSSION According to the summarized literature data, habitat type and moisture are the most important factors for the ecological classifi- cation of the opilioeid species. Our data, gath- ered from a study at a Bulgarian mountain, demonstrated that the main factor responsible for determining the ecological profiles of the harvestmen is elevation. But since elevation could not be regarded as a physiologically ac- tive factor per se, it may be suggested that elevational biotic (e.g. the decrease of pro- ductivity) or abiotic (e.g. low amount of avail- able microhabitats, harsher climatic condi- tions) correlates, or even an unmeasured environmental parameter, would rather be the immediate ecological component acting upon the harvestmen. However, the habitat type and 264 THE JOURNAL OF ARACHNOLOGY Figure 5. — GAM surface plots for the modelled abundance of harvestman species that occurred only at low-elevations localities: Zachaeus crista, Rilaena balcanica, Carinostoma ornatum, Opilio saxatilis, Lacinius dentiger, Trogulus closanicus. The space of sampling localities (dots) is the same as in Fig. 2; the isolines show the modelled abundance of each species. moisture turned out to be of some importance for the species in the low-mountain zone of Vitosha, mainly because the habitat type de- pends on moisture on one hand and regulates it on the other. This observation is in concor- dance with the observations of Platen (pers. comm.) that in Germany most harvestmen species occur in shady and somewhat moist habitats. It might be suggested that elevation was found to be an important factor mainly because of the “mountaineous” character of this study, but two further works, one on a mountain (Muster 2001; Alps) and a region- wide one (Klimes 1997, data from the entire Czech Republic) also demonstrated the pri- mary role of elevation for shaping the opi- lionid assemblages. One particular reason for the failure of this investigation to show any strong association of the Opiliones with factors other than ele- vation, could be a result of the very complex environmental matrix obtained in this study. The habitat parameter that showed the best spread among the sampled localities was the elevation (and its correlates such as vegetation belt and soil type; see Figs. 1, 2). The other measured environmental parameters do not show such a broad variation among sampling units, and thus may not contribute signifi- cantly to their discrimination. Consequently, when modelled over the ordination plane, the response of individual species could not be clearly associated with environmental factors that do not demonstrate large variation across the investigated area, especially when these responses do not show any pronounced opti- ma at factor centroids. Another reason could be the dependence of harvestmen on various structures or conditions occurring within a specific habitat (and not on the habitat itself). Since we have not investigated microhabitat structures, this question should remain open until a study focused on within-habitat (mi- crohabitat) structures is conducted. In contrast to the mostly unnuanced and un- diversified classifications found in the litera- ture, is the finding that different species show quite different response-types towards the en- vironmental parameters. This fact could not be discovered by the modelling studies cited above (Klimes 1997; Muster 2001), since these have employed the modelling technique of choice without verifying its basic asump- tions (i. e. the unimodal response of species). We found that in fact the minority of the har- vestmen species from Vitosha Mountain had a unimodal distribution with a clear optimum (or preferendum) throughout the studied area. This linear response may be due to the inves- tigations following a gradient of elevation, and because some species have made their niche at a certain elevation to avoid compe- MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 265 tition and/or unfavourable environmental con- ditions» As Platen pointed out (pers. comm.), even in a lowland opilioeids may display un- imodal responses along gradients of moisture and light exposure, respectively. Whether there is a bimodal response in some species can be precisely decided in laboratory exper- iments, but it may be suggested that it would be of rare occurrence in Opiliones because of their strong dependence on humidity. In this situation it may be argued that an even more variable environmental matrix (i.e. with broader amplitude of environmental condi- tions, or including more measured environ- mental variables) and/or specially designed laboratory experiments should be used for re- fining the delineated ecological profiles, as well as for allowing the observation of poten- tial uni- and bimodal responses of harvestmen species. Another important information that emerged from this study is that congeneric species are quite different in their responses towards the environmental variables. This has been repeatedly postulated by theoretical ecol- ogists as a mechanism for minimizing the po- tential competition (e.g. Begoe et al. 1996; Giller 1984), and we suggest that this could be valid also for the Opiliones from Vitosha Mountain. ACKNOWLEDGMENTS We would like to express our gratitude to Dr; Christian Komposch (Graz, Austria), Dr. Spren Toft (Aarhus, Denmark), Dr. Ralf Platen (Berlin, Germany) and an anonymous review- er for their valuable comments on the manu- script. We would also like to thank Chr. Kom- posch, R. Platen, and Dr. Vito Ziegerle (Bolzano, Italy) for the kindly provided liter- ature. LITERATURE CITED Akima, H. 1978. A method of bivariate interpola- tion and smooth surface fitting for irregularly distributed data points. ACM Transactions on Mathematical Software 4(2): 148=159. Begon M., C. Harper & C. Townsend. 1996. Ecol- ogy: Individuals, Populations and Communities. Blackwell Science, Oxford. Beron, P. & P. Mitov. 1996. Cave Opilionida in Bul- garia. Historia naturalis bulgarica 6:17=23. Bliss, P. & F. Tietze, 1984. Die Straktur der epe- daphischen Weberknechtfauna (Arachnida, Opi- lioees) in unterschiedlich immissionsbelasteten Kiefernforsten der Diibener Heide. Pedobiologia 26:25-35. Chessel D., A.B. Dufour & J, Thioulouse, 2004. The ade4 package — I: One-table methods. Rnews 4:5-10. Chucheva, N. 1983, Landshaftno-ekologicheski an- aliz. Bioklimatichna karta. Tipowe mestoraste- niya. Rastitelna pokriwka. In Naroden Park “Vi- tosha’h Sofia, IPPGSS “Agrolesproekt” (in Bulgarian). Cirdei, F. & F. Bulimar. 1960. Date sistematice §i ecologice asupra subfam.: Gyantinae §i Oligolo- phinae (Ord. Opiliones) din Moldova. Analele §tiin|ifice ale Universita,|ii “AL L Cuza” din Ia§i (SN), Secflunea II 6(2):301~306 (in Russian, with Romanian and French summary). Curtis, D.J. 1978. Community parameters of the ground layer araneid-opilionid taxocene of a Scottish Island. Symp. Zoological Society of London 42:149-159. Czechowski, W., A. Kubicka & W. Star^ga. 1981. Harvestmen (Arachnoidea, Opiliones) of Warsaw and Mazo via. Memorabilia Zoologica 34: 1 1 1- 118. Geyer, K.A, 1983. Weberknechte aus dem Westerz- gebirge (Arachnida, Opiliones), Teil 1. Veroffee- tlichungen des Museums fiir Naturkunde Karl- Marx-Stadt 12:21-24. Giller, P.S. 1984. Community Structure and the Niche. Chapman and Hall, London. Hastie, T. & R. Tibshirani. 1990. Generalized Ad- ditive Models. Chapman and Hall, London. Hiebsch, H. 1972. Beitrage zur Spinnen- und We- berknechtfauna des Neissetales bei Ostritz. Ab- handlungen und Berichte des Naturkundemu- seums Gorlitz, Leipzig 47:1-32. Hiebsch, H. 1978. Beitrage zur Weberknechtfauna des Naturschutzgebietes Tiefeetal. Veroffentli- chungen des Museums der Westlausitz 2:85-91. Hubenov, Z. 1990. Physicogeographical character- istics of Vitosa. Pp. 7-18. In Botev, B. (ed.) Fau- na of Southwestern Bulgara, Part 3. Publishing house of the BAS, Sofia.(in Bulgarian with En- glish summary), Ihaka, R. & R. Gentleman. 1996. R; a language for data analysis and graphics. Journal of Computa- tional and Graphical Statistics 5(3):299-314. Karaman, 1. 1995. Fauna of harvestmans (Arach- nida, Opiliones) from Durmitor region. MSc. Thesis, Faculty of Natural Sciences and Mate- maticws, University of Novi Sad, Novi Sad, Yu- goslavia, 73 pp. (in Serbian, with English ab- stract). Klimes, L. 1987. Poznamky k pohybove aktivite sekacou (Opiliones). Sbornik Jihoceskeho muzea V CeskWh Budejovicech, Prirodni Vedy 27:51- 54. Klimes, L. 1990. Impact of floods on the life cycle in Rilaena triangularis (Herbst) (Opiliones), 266 THE JOURNAL OF ARACHNOLOGY Sbornik Jihoceskeho Muzea v Ceskych Budejov- icech, Pnrodm Vedy 30:37-45 (with English summary). Klimes, L. 1997. Harvestman (Phalangida) assem- blages in the Czech Republic. Acta Societatis Zoologicae Bohemicae 61:297-309. Klimes, L. 2002. Effect of floods on ground beetles (Carabidae) and harvestmen (Opiliones). Pp. 361-371, In Freshwater Wetlands and their Sus- tainable Future: Evidence from the Trebon Basin Biosphere Reserve (J. Kvet & J. Jenik, eds.). Man and the Biosphere series: v. 28, UNESCO. Klimes, L. & E. Spicakova. 1984. Beitrag zum Er- kennen der Dynamik vom Arachnofauna an dem Waldokoton. Acta Universitatis Palackianae Olo- mucensis Facultas Rerum Naturalium, Biologica 81:167-190 (in Czech, with German and Russian summary). Kolos vary, G. 1965. Opilioniden des Gebietes der ungarischen Volksrepublik. Acta Biologica, Sze- ged N.S. 11(1-2):165-168. Kolosvary, G. 1966a. Die Opilioniden Reservates “Saser” bei der Theiss. Tiscia 2:119. Kolosvary, G. 1966b. Uber Okologie der Weber- knechte der Inundationsraume der Theiss. Tiscia 2:121-125. Komposch, Ch. 1995. Weberknechte (Opiliones). Pp. 69 — 74. In Naturfiihrer Sablatnigmoor (Ch. Wieser, A. Kofler & P. Mildner, eds.). Verlag des Naturwissenschaftlichen Vereins fiir Karnten, Klagenfurt. Komposch, Ch. 1997a. Kommentierte Chekliste der Weberknechte (Opiliones) Karntens. Carinthia II, 187(107):597-608. Komposch, Ch. 1997b. Die Weberknechtfauna (Op- iliones) des Nationalparks Hohe Tauern. Faun- istisch-okologische Untersuchungen von der Montan- bis zur Nivalstufe unter besonderer Be- riicksichtigung des Gossnitztales. Wissenschaftli- che Mitteilungen aus dem Nationalpark Hohe Tauern, 3:73-96. Komposch, Ch. 1997c. The arachnid fauna of dif- ferent stages of succession in the Schtitt rockslip area, Dobratsch, southern Austria (Arachnida: Scorpiones, Opiliones, Araneae). Proceedings of the 16th European Colloquium of Arachnology, Siedlce, Poland: 139-149. Komposch, Ch. 1999. Rote Liste der Weberknechte Karntens (Arachnida: Opiliones). Pp 547-565, In Rote Listen gefahrdeter Tiere Karntens (T. Rot- tenburg, C. Wieser, P. Mildner & WE. Holzinger, eds.). Naturschutz in Karnten 15, Klagenfurt. Komposch, Ch. 2000. Harvestmen and spiders in the Austrian wetland “Horfeld-moor” (Arach- nida: Opiliones, Araneae). Ekologia (Bratislava) 19, Supplement 4:66-77. Komposch, Ch. 2001. Weberknechte (Opiliones). Pp. 314-317, 747. In Landesamt fiir Umwelts- chutz Sachsen-Anhalt (Hrsg.): Arten- und Bio- topschutzprogramm Sachsen-Anhalt. Land- schaftsraum Elbe, Teil 1-3, 781 pp. Berichte des Landesamtes fiir Umweltschutz Sachsen-Anhalt, SH 3. Komposch, Ch. 2004. The harvestmen fauna of Hungary (Arachnida, Opiliones). Pp. 227-242, In European Arachnology 2002 (E Samu & Cs. Szinetar, eds.) (Proceedings of the 20th European Colloquium of Arachnology, Szombathely, 22- 26 July 2002), Plant Protection Institute & Berz- senyi College, Budapest. Komposch, Ch. & J. Gruber. 1999. Vertical distri- bution of harvestmen in the Eastern Alps (Arach- nida: Opiliones). Bulletin of the British Arach- nological Society 1 1(4): 131-135. Komposch, Ch. & J. Gruber. 2004. Die Weberkne- chte Osterreichs (Arachnida, Opiliones). Denisia 12, zugl. Kataloge der OO. Landesmuseen Neue Serie 114:485-534. Kiihnelt, W 1965. Grundriss der Okologie. Gustav Fischer Verlag, Jena. Kuschka, V. 1991. Beitrage zur Okologie der epi- gaischen Arthropoden in Feucht-Okosystemen des Naturschutzgebiets (NSG) “Giilper See”. Teil I: Spinnen (Araneae) und Weberknechte (Opiliones). Zoologische Jahrbiiher ftir Syste- matik 1 18(2):2 17-246. Lymberakis, R, P. Mitov, M. Chatzaki & M. My- lonas. In press. Phenology of Opiliones on an altitudinal gradient at Lefka Ori mountains, Cre- te, Greece. Ecologia Mediterranea. Martens, J. 1978. Die Tierwelt Deutschlands, 64. Teil: Spinnentiere, Arachnida: Weberknechte, Opiliones. Gustav Fischer Verlag, Jena. Meijer, J. 1972. Some data on the phenology and the activity-patterns of Nemastoma lugubre (Muller) and Mitostoma chrysomelas (Hermann). Netherlands Journal of Zoology 22(1):105-1 18. Metzen, A. & K. Colin. 1998. Landasseln und We- berknechte (Crustacea, Isopoda: Oniscidea; Arachnida: Opiliones) aus dem Ortsbereich und der Umgebung von Gdnnersdorf (Kr. Daun, Ei- fel). Dendrocopos 25:238-269. Muller, H.G. 1984. Beitrag zur Opilionidenfauna von Hessen. Hessische faunistische Briefe 4(1): 2-6. Muster, C. 2001. Biogeographie von Spinnentieren der mittleren Nordalpen (Arachnida: Araneae, Opiliones, Pseudoscorpiones). Verhandlungen d. naturwissenschaftlichen Vereins Hamburg NF 39:5-196. Mitov, P. 1986. On some species of the family Ne- mastomatidae (Opiliones) from the Rodopa Mt. Travaux scientifiques Universite de Plovdiv “P.Hilendarski”, Biologie, 24, 1:297-299. (in Bulgarian, with German abstract). Mitov, P. 1987. New species of Opiliones for the fauna of Vitosha Mountain. Travaux scientifiques Universite de Plovdiv “Hilendarski”, Biologie. MITOV & STOYANOV— ECOLOGICAL PROFILES OF BULGARIAN OPILIONES 267 25(6):59-6L (In Bulgarian, with English ab- stract). Mitov, P. 1994. Siro beschkovi, spec, nov, aus Bul- garien (Arachnida, Opiliones, Cyphophthalmi). Spixiana 17(3):275-282. Mitov, R 1995a. New faunistic and chorologic data about Opiliones (Arachnida) from Bulgaria. An- nuaire de FUniversite de Sofia “St. Kliment Ohr- idski”, Faculte de Biologic, Livre 1, Zoologie 86/87(l):63-65. Mitov, P. 1995b. Artenbestand und Tagesaktivitat von Opiliones (Arachnida) aus einigen Geroll- halden im Vitoscha-Gebirge, Bulgarien. Arach- nologisches Magazin 3(12): 1-6. (English sum- mary). Mitov, P. 1995c. Teratological data about Opiliones (Arachnida) from Bulgaria. Pp. 147-158, In Pro- ceedings of the 15th European Colloquium of Arachnology. (V. Ruzicka ed.), Institute of En- tomology, Ceske Budejovice, Czech Republic. Mitov, P. 1996. Vorlaufige Beobachtungen zur Tag- und Nacht-Bewegungsaktivitat der epigaischen Weberknechnte (Opiliones, Arachnida) aus ver- schiedenartigen Habitaten im Vitosha-Gebirge (Nordteil), SW-Bulgarien. Revue suisse de Zool- ogie, vol. hors serie:479-489. (English summa- ry). Mitov, P. 1997a. Ein neuer Nelima Roewer aus Bul- garien (Arachnida, Opiliones, Phalangiidae). Spixiana 20(2):97-105. Mitov, P. 1997b, Preliminary observation on the spatial distribution of the harvestmen (Opiliones, Arachnida) from Vitosha Mt, (SW Bulgaria). Pp. 249—258. In Proceedings of the 16th European Colloquium of Arachnology. (M. Zabka ed.), Wydawnictwo Wyzszej Szkoly Rolniczo-Peda- gogicznej, Siedlce, Poland. Mitov, P. 2000. Faunistic, Biological and Ecological Investigation on the Opiliones from Vitosha Mt. with Some Zoogeographical Notes. PhD Thesis, University of Plovdiv “P. Hilendarski”, Plovdiv, Bulgaria, 274+131 pp. (in Bulgarian). Mitov, P. 2001. Harvestm.eri (Opiliones, Arachnida) of Kresna Gorge (SW Bulgaria). Pp. 75-83, In Biodiversity of Kresna Gorge (SW Bulgaria) (R Beron, ed.). National Museum of Natural History & Institute of Zoology, BAS, Sofia. (In Bulgar- ian, v/ith English summary). Mitov, P. 2002. Rare and endemic harvestmen (Op- iliones, Arachnida) species from the Balkan Pen- insula. 1. On Mediostoma stussineri (Simon 1885) (Nemastomatidae) — new species and ge- nus for the Bulgarian fauna. Linzer biologische Beitrage 34(2): 1639-1648. Mitov, P. 2003. Rare and endemic harvestmen (Op- iliones, Arachnida) species from the Balkan Pen- insula. 11. Three species new for the Bulgarian fauna with zoogeographical notes. Linzer biolo- gische Beitrage 35(l):273-288. Mitov, P. 2004. Harvestmen (Arachnida: Opiliones) of the Eastern Rhodopes (Bulgaria). Pp. 167- 179, In Biodiversity of Bulgaria. 2. Biodiversity of Eastern Rhodopes (Bulgaria and Greece) (P. Beron & A. Popov, eds). Pensoft & National Mu- seum of Natural History, Sofia. Mitov, P. In press. Spatial niches of Opiliones (Arachnida) from Vitosha Mts., Bulgaria. In Bio- geography and Ecology of Bulgaria. (V. Fet & A, Popov, eds.). Kluwer Academic Publishers, Netherlands, Mitov, EG. & I.L. Stoyanov. 2004. The Harvest- men Fauna (Opiliones, Arachnida) of the City of Sofia (Bulgaria) and its Adjacent Regions. Pp. 319-354, In Ecology of the City of Sofia. Spe- cies and Communities in an Urban Environment (L. Penev, J. Niemela, D.J. Kotze & N. Chipev, eds.), PENSOFT, Sofia-Moscow. Nedyalkov, S. 1998. Teoriya na Ekologiyata. PSSA, Sofia, 408 pp. (in Bulgarian). Novak, T, S. Lipovsek, L. Sencic, M. A. Pabst & F. Janzekovic. 2004. Adaptations in phalangiid harvestmen Gyas annulatus and G. titanus to their preferred water current adjacent habitats. Acta Oecologica 26:45-53. Obrtel, R. 1976. Soil surface harvestmen (Opilion- idea) in a lowland forest. Acta scientiamm na- turalium Academiae scientiarum bohemoslova- cae Brno 10(12): 1-34. Pabst, W. 1953. Zur Biologic der mitteleuropaisch- en Troguliden. Zoologische Jahrbiicher (Syste- matik) 82(1/2): 1-156. Platen, R. 1991. Okotoxikologische Untersuchungen im Staatswald Burgholz. Die Wirkung von Na- PCP auf die Spinnentierfauna (Araneida, Opilion- ida) mit einer Diskussion fiber die Indikatoreigung von Spinnentieren ftir Umweltchemikalien. Jah- resberichte des Naturwisseeschaftlichen Vereins in Wuppertal 44:115-132. Platen, R. 1996. Statistisch-okologische Analyse der Spinnenzonosen (Araneida) in exotischen und einheimischen Geholzanbauten im Staat- swald Burgholz. Jahresberichte des Naturwissen- schaftlichen Vereins in Wuppertal 49:145-168. Platen, R. 2000. Spinnen und Weberknechte im Staatswald Burgholz — Historic, Forschungspro- gramme, Ausblick. Jahresberichte des Naturwis- senschaftlichen Vereins in Wuppertal 53:206- 239. Platen, R. & B. v. Broee, 2002. Chekliste und Rote Liste der Webspinnen und Weberknechte (Arach- nida: Araneae, Opiliones) des Landes Berlin mit Angaben zur okologie. Markische Entomologis- che Nachrichten, 2:1-69. Platen, R., M. Moritz & B.v. Broen. 1991. Liste der Webspinnen- und Weberknechtarten (Arach.: Ar- aneida, Opilionida) des Berliner Raumes und ihre Auswertung fur Naturschutzzwecke (Rote Liste). pp. 169-205, In Rote Listen der gefahr- 268 THE JOURNAL OF ARACHNOLOGY deten Pflanzen und Tiere in Berlin (A. Auhagen, R, Platen & H. Sukopp, eds.), Landschaftsen- twicklung und Umweltforschung S6. Pfeifer, H. 1956. Zur Okologie und Larvalsystem= atik der Weberknechte. Mitteilungen d. Zoolo- gischen Museum Berlin 32(1):59--104. Rafalski, J. 1962. Opilio dinaricus Silhavy a little known Species of Harvestmen (Opiliones). Stu= dia Societatis scientiarum Torunensis E6:l-12. Sechterova, E. 1989. The spatial horizontal distri^ bution of epigeic spider and harvestman popu- latioes living in a spruce forest in the Beskydy Mountains (Araneae; Opiliones). Studia OecoL ogica (Ostrava) 2:17-27. Simon, U. 1995. Untersuchung der Stratozonosen von Spinnen und Weberknechten (Arachn.: Ar- aneae, Opilionida) an der Waldkiefer (Pinus syP vestris L.), Wissenschaft und Technik Verlag, Berlin. Stargga, W. 1976. Die Weberknechte (Opiliones, excl. Sironidae) Bulgariens. Annales Zoologici, Warszawa 33(18):287-433. Tenenhaus, M. & F.W. Young. 1985. An analysis and synthesis of multiple correspondence analy- sis, optimal scaling, dual scaling, homogeneity analysis and other methods for quantifying cat- egorical multivariate data. Psychometrika 50(1): 91-119. ter Braak, C.J.F 1987. The analysis of vegetation- environment relationships by canonical corre- spondence analysis. Vegetatio 69:69-77. Thaler, K. 1979. Fragmenta faunistica Tirolensia, IV (Arachnida: Acari: Caeculidae; Pseudoscor- piones; Scorpiones; Opiliones; Aranei; Insecta: Dermaptera; Thysanoptera; Diptera Nematocera: Mycetophilidae, Psychodidae, Limoniidae und Tipulidae). Veroffentlichungen des Museums Ferdinandeum (Innsbruck) 59:49-83. Tischler, W. 1967. Zur Biologic und Okologie des Opilioniden Mitopus morio F. Biologisches Zen- tralblatt 86(4):473-484. Todd, V. 1949. The habits and ecology of the british harvestmen (Arachnida, Opiliones), with special reference to those of the Oxford district. Journal of Animal Ecology 18(2):209-229. Weiss, I. 1975. Untersuchungen iiber die Arthro- podenfauna xerothermer Standorte im Sudsie- benbiirgischen Hiigelland. II. Weberknechte (Op- iliones, Arachnida). Studii §i Comunicari 19: 263-271. Weiss, 1. 1978. Biometrische und okologische Un- tersuchung der Gattung Trogulus am Konglom- erat von Podu Olt in Sudsiebenbiirgen (Arach- nida, Opiliones). Studii §i Comunicari 22:213- 228. Weiss, I. 1984. Okofaunistische Untersuchung der Spinnen und Weberknechte eines Hangprofils be! Seica Mare im siebenbiirgischen Hiigelland. Stu- dii §i Comunicari 26:243-277. Weiss, I, 1996. Die Weberknechtfauna Siebenbiir- gens (Arachnida: Opiliones). Stapfia 45:259- 280. Weiss, L & S.I. Sarbu. 1977. Zur Kenntnis der Spinnen und Weberknechte des botanischen Gar- tens Ia§i. Studii §i Comunicari 21:225-243. Williams, G. 1962. Seasonal and diurnal activity of harvestmen (Phalangida) and spiders (Araneida) in contrasted habitats. Journal of Animal Ecolo- gy 31:1-13. Wood, S.N. 2000. Modelling and smoothing param- eter estimation with multiple quadratic penalties. Journal of the Royal Statistical Society, series B 62(l):413-428. Zingerle, V. 1999. Epigaische Spinnen und Weber- knechte im Naturpark Sextner Dolomiten und am Sellajoch (Siidtirol, Italien). Berichte des natur- wissenschaftlich-medizinischen Vereins in Inns- bruck 86:165-200. Zingerle, V. 2000. Zoocenosi di Ragni e Opilioni nelle Dolomiti sudoccidentali (Parco Naturale Paneveggio-Plae di S. Martino, Italia) (Araneae, Opiliones). Studii Trentini di Scienze Naturali Acta Biologica 75:87-107. Manuscript received 27 January 2005, revised 13 September 2005. 2005. The Journal of Arachnology 33:269-279 INFLUENCE OF GRAZING BY LARGE MAMMALS ON THE SPIDER COMMUNITY OF A KENYAN SAVANNA BIOME Charles M. Warui^’^, Martin H. VilleC, Truman P. Young^ and Rudy Jocque"*: ^Department of Invertebrate Zoology, National Museums of Kenya, RO. Box 40658- 00100 GPO, Nairobi, Kenya, E-mail: cmwarui@yahoo.com; ^Department of Zoology and Entomology, Rhodes University, 6140 Grahamstown, South Africa; ^Department of Plant Sciences, University of California, Davis, 95616 USA; invertebrate Section, Department of African Zoology, Royal Museum for Central Africa, Leuvensesteenweg 13, B-3080 Tervuren, Belgium ABSTRACT. Pitfall trap and sweep net samples were taken over a period of fifteen months (2002- 2003) in the Kenya Long-term Exclosure Experiment (KLEE), in which the presence of domestic and wild herbivores have been independently manipulated since 1995. ANOVA and ANCOVA showed that the exclosure treatments significantly affected plant cover, with the presence of cattle significantly reducing the relative vegetation cover and spider diversity. Herbivory by indigenous mega- and meso-herbivores did not have a significant influence on the diversity of the spider fauna, but abundance of three dominant species (Cyclosa insulana Costa (Araneidae), Argiope trifasciata ForskM (Araneidae) and Runcinia flavida Simon (Thomisidae)) decreased in cattle-grazed plots. In contrast, Aelurillus sp. became more prevalent where cattle have been grazing. Multivariate analyses revealed that the spider community responded to grazing pressure by aggregating into three groups that reflected control, cattle grazing and non-cattle grazing clusters. It was probable that the direct effects on vegetation mediated an indirect influence of herbivores on spider diversity. The relative vegetation cover was a positive predictor of spider diversity. Spider communities were found to be an indicator of the activity of mammals and could be used as indicators of land use changes and for bio-monitoring. Keywords; Grazing, mammals, savanna, Kenya, spiders Savanna inventories. — Little ecological work has been done on spiders of African sa- vannas and inventories from this habitat are rare. For example, the only inventory work in Kenya was carried out by Russell-Smith et al. (1987), who reported 68 species from Kora Game Reserve. Recently, Warui et al. (2004) reported a checklist of 132 species from a black cotton soil ecosystem in Laikipia. In Tanzania, a checklist of 508 species from Mkomazi Game Reserve was published by RusselLSmith (1999). In South Africa, several surveys of spiders were undertaken in the Sa- vanna Biome. Dippenaar-Schoeman et al. (1989) reported 98 species from Roodeplaat Dam Nature Reserve while Dippenaar-Schoe- man and Leroy (2003) reported another 152 species from the Kruger National Park and Foord et al. (2002) recorded 127 species from the western Soutpansberg. Another 55 species were recorded from Rietondale, Pretoria (van den Berg & Dippenaar-Schoeman 1991), and 268 species from Makalali Game Reserve in the Limpopo Province (Whitmore et al. 2001). Lastly Lotz et al. (1991) working on grassland biome reported 31 families of spiders from Bloemfontein. The only other works on sa- vanna spiders apart from check-lists are those of Russell-Smith (1981), who reported 135 species from Botswana; and Blandin & CeL erier (1981), who studied savanna spiders in Ivory Coast. Current study. — This study was part of the Kenya Long-term Exclosure Experiment (KLEE), a long-term multi-species vertebrate herbivore exclusion experiment in a semi-arid savanna ecosystem in Laikipia, Kenya (Young et al. 1998). KLEE is aimed at comparing the impacts of cattle and wildlife (elephants, gi- raffes, buffaloes, antelopes and other savanna ungulates) on various components of the sa- vanna biome including biodiversity. Refer- 269 270 THE JOURNAL OF ARACHNOLOGY ence is made to spiders because they inhabit a large array of microhabitats ranging from the ground layer, to the tree layer and makes them particularly suitable to integrate and evaluate activity by the different guilds of her= bivores. Since the response of spiders to the particular structure of the habitat is very fine- grained (Gunnarsson 1988; Uetz 1991; Ryp- stra et al. 1999), it was expected that changes caused by the different guilds of herbivores, would be reflected in the spider fauna. The influence of abiotic environmental variables was also investigated for a few individual spe- cies. Most studies on the influence of grazing and trampling concentrate on the effects on the fauna or vegetation as a whole. Outside Africa and in different ecosystems, such gen- eral investigations were carried out by Gibson et al. (1982, 1992) and Curtis et al. (1990) who found that communities of spiders were negatively affected by grazing and trampling. Abensperg-Traun et al. (1996) studied the grazing impact of mammals on invertebrates in Australian woodland and found that the abundance of the spider families Idiopidae and Lycosidae was highest in moderately dis- turbed woodlands. Rambo & Faeth (1998) looked at influence of grazing on plant insect communities. In Africa, Woldu & Saleem (2000) focused on plant biodiversity in Ethi- opia, while Rivers-Moore & Samway s (1996), Fabricius (1997), Seymour (1998), Seymour & Dean (1999) and Fabricius et al. (2002) demonstrated that grazing or trampling has ef- fects on various groups of invertebrates in South Africa. Earlier African studies were re- viewed in Skarpe (1991). Few studies are available that report the influence of grazing on spiders in particular: Churchill (1998) re- ported a variation in the abundance of domi- nant spider families along grazing and rainfall gradients in Australian tropics. Abrous-Kher- bouche et al. (1997) investigated the effects of grazing in mountain grassland in North Afri- ca. The present study is the first that studies the subject in tropical Africa and uses a large- scale experimental set-up for the purpose. This is the second paper on Kenyan savanna spiders by the author and more reference can be made to Warui et al. (2004). METHODS Study area. — The study was conducted at Mpala Research Centre (MRC) (00°17'N 037°52'E, 1750-1800 m asl), a 1200 ha piece of land adjacent to Mpala Ranch in the Lai- kipia District of central Kenya. The study site is characterized by black cotton soil (Chromic vertisols), which are heavily textured cracking clays with impeded drainage (Ahn & Geiger 1987; Taiti 1992). Its vegetation is Acacia bushed grassland (Young et al. 1998) domi- nated by A. drepanolobium (Harms) Sjostedt, accounting for over 95% of the woody vege- tation. Rainfall averages 500-600 mm per year (Young et al. 1995, 1998). Data were col- lected from May 2001 to July 2002. The KLEE study design. — The Kenya Long-term Exclosure Experiment is a set up in which the presence of domestic and wild herbivores has been independently manipulat- ed since 1995. KLEE allows herbivory (graz- ing and browsing) in six combinations of three categories of herbivores. These three catego- ries are (1) meso-wildlife (W) (or meso-her- bivores: buffalo and other smaller ungulates), referred to as ‘wildlife’ in Young et al. (1998); (2) mega-wildlife (M) (or mega-herbivores: giraffes and elephants); and (3) cattle (C). The grazing by cattle was moderate, with one live- stock unit per 5-8 ha (Young et al. 1998). The details of this design are shown in Fig. 1. The three categories of the large mammalian her- bivores were managed such that (i) only cattle (C); (ii) only meso-herbivores (W); (iii) only mega-herbivores and meso-herbivores (MW); (iv) mega-herbivores, meso-herbivores and cattle (MWC); (v) only meso-herbivores and cattle (WC); and (vi) no large mammalian her- bivores (control, O) were allowed to graze/ browse. Each treatment plot is 200 X 200 m and is replicated three times, once in each of three blocks (north, central and south), total- ling 18 plots. Spider collection. — Spiders were collected with pitfall traps and by sweep-netting. Much has been published about advantages and lim- itations of pitfall traps (e.g., Greenslade 1964; Uetz & Unzicker 1976; Spence & Niemela 1994; Green 1999; New 1999) and this study employed them to allow comparison with data from published studies. The pitfall traps con- sisted of two cone-shaped plastic (polyethyl- ene) cups 9 cm wide at the mouth and 14 cm deep, one inside the other, buried to their rim. Three pitfalls per plot for each of the 1 8 sam- pling plots were used, making a total of 54 traps. The three pitfall traps were laid on a WARUI ET AL.-EFFECTS OF MAMMAL GRAZING ON SPIDERS IN KENYA 271 Mpala Farm - — Dirt road Graded firebreak Ungraded track ............. Game fence Mega-herbivore fence ■••••• Cattle posts piqi boundaries Tall connecting line • Location of electrics Glades 400 ™J metres Figure 1. — Schematic representation of the experimental design of the KLEE study plots at Laikipia, Kenya. Letters in each plot represent the herbivores allowed in: C == cattle, W = meso-herbivores, M = mega-herbivores, O = control (all large mammalian herbivores excluded). N, C and S represent north, central and south blocks respectively. Each plot measures 200 X 200 m. The distance between the furthest placed plots (between north and south block) is approximately 2 km. Adapted from Young et al. (1998). line transect every 3 m. The inner cup of each trap was filled to a third of its volume with a 2% formaldehyde solution as a preservative. Traps were left open and emptied every sec- ond week. Sweep-netting was done by walk- ing through the herb layer swinging a sweep net (40 cm in diameter) through the vegetation for a standard number of times (Coddington et al. 1996; Scharff & Griswold 1996; Dip- penaar-Schoeman et al. 1999). Sweeping was done on a randomly selected 50 m transect in each of the 18 plots. A hundred sweeps (emp- THE JOURNAL OF ARACHNOLOGY 272 Z Herbivores Herbivores J Figures 2-3. — 2. Effects of 'cattle’ (levels: absent [treatmentsW and MW] vs. present [WC and MWC]) and ‘megaherbivores’ (levels: absent [W and WC] vs. present [MW and MWC]) on relative vegetation cover (mean + SE). Two treatments (O and C) were omitted from the data set so that the analysis was fully crossed. The interaction term was not significant (P = 0.28). 3. Effects of ‘cattle’ (levels: absent [O and W] vs. present [C and WC]) and ‘mesoherbivores’ (levels: absent [O and C] vs. present [W and WC]) on relative vegetation cover (mean T SE). Two treatments (MW and MWC) were omitted from the data set so that the analysis was fully crossed. The interaction term was not significant (P = 0.79). tied after every 10 sweeps with an aspirator) were made along each transect. The process was repeated every fortnight throughout the study period. Vegetation sampling* — The vegetation cover was sampled once every month in all the study plots using a ten-point pin frame and quadrat methods where samples were collect- ed on sweep-netting and pitfall-trapping tran- sects. The percentage relative vegetation cover was calculated by deducting the total number of bare hits from pin totals to give the plant cover hits, which were then expressed as a percentage. Weather measurements. — Monthly rain- fall was recorded using three rain gauges placed in each of the three study blocks (north, central and south). The mean maxi- mum temperature is between 24 and 27 oC (Ahn & Geiger 1987). Statistical analyses. — Four diversity indi- ces [Shannon- Wiener (H), Margalef (d), Pie- lou (J) and total species (S)] were computed using PRIMER (Clarke & Gorley 2001). Oth- er statistical tests were performed using STA- TISTICA (StatSoft 1999). In this study, ordi- nations by non-metric multidimensional scaling (MDS) were computed in the MDS module of PRIMER, where the original abun- dance data matrix was first converted into a Bray-Curtis similarity matrix using the SIM- PLER module of PRIMER (Clarke & War- wick 1994). This is the most commonly used similarity coefficient in ecological work and accounts well for rare species. It down- weights the contributions of rare species in an entirely natural way such that the rarer the species, the less it contributes (Clarke & War- wick 1994). MDS only considers that an or- dination is a reasonable representation of sim- ilarity by looking at stress values which range from 0-1 and increase with reduced dimen- sionality of the ordination. Low stress values (< 0.1) are the best two-dimensional presen- tation of data points. In the current study only ten iterations were used. Normality and transformation of data. — Levene’s test was used to test the homosce- dacity of the data while data on percentage relative vegetation cover were arcsine-trans- formed before being subjected to ANOVA. Square root transformation was performed on all spider abundance data in order to make the underlying distribution normal before any ANOVA or analyses of covariance (ANCO- VA) were performed. ANOVA and ANCOVA results were done only where Levene’s test was not significant or there were no serious violation of the assumptions of ANOVA. RESULTS A total of 10,487 specimens, representing 132 species in 30 families, were collected WARUI ET AL.-EFFECTS OF MAMMAL GRAZING ON SPIDERS IN KENYA 273 Table L — Results of ANOVA on effects of the factors ‘cattle’ (levels: absent [treatments O, W and MW] vs. present [C, WC and MWC]) and ‘herbivores’ (levels: herbivores absent [O and C], only meso- herbivores present [W and CW], and both meso- and mega-herbivores present [MW and MWC]) on relative vegetation cover. The codes for the treatment abbreviations are (cf. Fig. 1): O = control (no large mammalian herbivores); W = meso-herbivores; M = mega-herbivores and C = cattle. No treatments were omitted from the data set. * = Significant at a = 0.05. Factor Mean relative Absent cover ± SE Present df MS F P Cattle 59.24 ± L74 53.43 ± 1.18 1 151.90 8.77 0.012* Herbivores 58.06 ± 2.08 57.61 ± 2.67 2 40.86 2.36 0.137 Cattle & Herbivores 56.35 ± 1.22 52.94 ± 2.60 2 14.61 0.84 0.454 Error 12 17.31 from the study area (Warui et aL 2004). Newly recorded species appeared throughout the sampling period for both sweep-netting and pitfall (see Warui et aL 2004). The sweeping method accounted for 67 species and pitfall- trapping accounted for approximately 110 species. Vegetation co¥er. — The first analysis used all six cattle treatments with two levels for the factor 'cattle’ (present/abseet), and three lev- els for the factor ’herbivores’ (absent/only meso-herbivores present/both meso- and mega-herbivores present). Only the presence of cattle had a significant, negative effect on vegetation cover (Table 1). Similarly, a second analysis tested the effects of the factors ‘cat- tle’ (with levels present vs. absent) and ‘mega- herbivores’ (with levels present vs. absent), using all treatments containing herbivores (W, WC, MW, MWC). Two treatments (O and C) were omitted because the KLEE experimental layout was not fully crossed. This analysis re- vealed that only the presence of cattle had a significant, negative effect on vegetation cov- er (Lj^g = 1231, P = 0.008, Fig. 2). Mega- herbivores had an almost significant negative effect on relative vegetation cover (Lj^ g = 4.59, P = 0.065, Fig. 2), A third analysis test- ed the effects of the factors ‘cattle’ (with lev- els present vs. absent) and ‘meso-herbivores’ (with levels present vs. absent) in the four treatments that excluded mega-herbivores (O, C, W, WC). The mega-herbivore treatments (MW and MWC) were omitted because the KLEE experimental layout was not fully crossed. The results showed that there was no significant effect of cattle or meso-herbivores on relative vegetation cover and the resulting interaction was not significant (Fig. 3). How- ever the mesoherbivores had a near significant negative effect on relative vegetation cover (Fig. 3). Spiders* — -Only the presence of cattle had a negative effect on spider abundance from sweep-netting samples (Fj^ 500 ™ 5.84, P = 0.016). The presence of mesoherbivores had no significant effect on abundance of spiders from sweep-netting samples (^1^500 = 5.84, P = 0.177). Similarly, an ANOVA to test the effects of cattle and mega- and meso-herbi- vores on spider richness (total number of spe- cies) revealed that only the presence of cattle had a significant negative effect on sweep-net- ting samples (^^^332 ~ 6.05, P = 0.014), (Fig. 4). Only the presence cattle had a significant negative effect on Shannon-Wiener diversity from sweep-netting samples (Fj 332 = 4.68, P - 0.031). There was a positive, significant correlation between relative vegetation cover and Pielou’s evenness index and the Shannon-Wiener di- versity index for sweep-netting samples (Ta- ble 2). Diversity indices from pitfall-trapping samples were not significantly related to rel- ative vegetation cover (Table 2). Four study species were chosen for individ- ual analysis based on the fact that they were the most numerically dominant and represent- ed a number of different functional groups: Cyclosa insulana (Costa 1834), Argiope tri- fasciata (Forskal 1775) (both Araeeidae), Runcinia flavida (Simon 1881) (Thomisidae), and Aelurillus sp. (Salticidae). A series of analyses of covariance (ANCOVA) were per- formed to establish their response to some bi- otic and abiotic factors, namely relative veg- etation cover, total monthly rainfall and presence of large mammalian herbivores. The 274 THE JOURNAL OF ARACHNOLOGY A P A P Herbivores Figure 4. — Effects of 'cattle' (levels: absent [O, W and MW] vs. present [C, WC and MWC]) and ‘herbivores’ (levels: herbivores absent [O and C], only mesoherbivores present [W and CW], and both meso- and megaherbivores present [MW and MWC]) on total number of spider species from sweep-netting samples (mean + SE). No treatments were omitted from the data set. The interaction term was not significant (P — 0.81). summarized results are shown in Table 3. The presence of cattle and meso-herbivores had significant, negative effects on the abundance of all of the species except Aelurillus sp., where the presence of cattle was related to an increase in the species’ abundance. Only K flavida and Aelurillus sp. were significantly affected by the amount of rainfall (Table 3). Finally, the stress values of multidimen- sional scaling (MDS) ordinations for the sweep-netting (Fig. 5) and pitfall-trapping data sets were 0.15 and 0.01, respectively, which implies that the plots were reliable two- dimensional representations of the n-dimen- sional similarities of the samples and therefore worth interpreting (Clarke & Warwick 1994). The aim of this analysis was to show whether the spider community organised itself in pat- tern that reflected the intensity of grazing by different herbivore groups. The MDS ordina- tions for sweep-netting samples have a clearer separation into three clusters of control, cattle and non-cattle grazing, (Fig. 5) when com- pared to pitfall-trapping samples (not shown) which did not separate by herbivore grazing group. For sweep-netting samples, only the southern control plot was peculiar (Fig. 5) and appeared to be in the same position as the cat- tle grazing plots. The other two control plots are in their own well-separated cluster. Graz- ing and control plots are separated by meso- Table 2. — Correlations between relative vegeta- tion cover and four measures of diversity (Shannon- Wiener diversity index [H'], Margalef’s richness in- dex [d], Pielou’s evenness index [J'] and total spider species [S]) for data sets generated at Laikipia, Kenya in 2001-2002 using sweep-netting and pit- fall-trapping samples, df = 18. * = Significant at a = 0.05. Method Diversity index r~value F-value Sweep-netting samples S 0.35 0.160 d 3.14 0.204 J' 0.54 0.020* H' 0.61 0.007* Pitfall-trapping samples S 0.29 0.244 d 0.26 0.304 J' 0.06 0.809 H' 0.23 0.356 herbivores (W) and mega-herbivore (M) treat- ment plots. For the pitfall-trapping data most cattle-grazing and non-cattle grazing plots overlapped, thus no interpretation could be made. DISCUSSION There is considerable evidence that grazing and trampling have an influence, and in vir- tually all cases a negative one, on spider di- versity (Gibson et al. 1982, 1992; Curtis et ak 1990; Abensperg-Traun et al. 1996; Rivers- Moore & Samways 1996; Abrous-Kherbou- che et al. 1997; Fabricius 1997; Churchill 1998; Fabricius et al. 2002). Yet, this is the first paper that compares the influence of do- mesticated animals on spiders with that of wildlife. Our analyses (Table 1 and Figs. 2- 4) support the conclusion that the presence of cattle, much more than that of other large mammalian herbivores, reduces relative veg- etation cover and spider diversity and abun- dance, while other results (Table 2) demon- strate that diversity and species richness are correlated with relative vegetation cover. As expected, the presence of herbivores had an indirect effect on spiders, presumably by re- ducing the relative vegetation cover and hence the complexity of the habitat. Spiders were significantly scarcer in the treatments with cattle compared to those with other large mammalian herbivores. However, some of the effects by mega- and meso-her- bivores were close to significance suggesting WARUI ET AL.-EFFECTS OF MAMMAL GRAZING ON SPIDERS IN KENYA 275 NMWC Stress 0.15 Figure 5. — Multidimensional scaling (MDS) ordination of the spider community in the sweep-netting samples of spiders collected at Laikipia, Kenya in 2001-2002, with convex hulls superimposed to enclose regions characteristic of control, cattle and non-cattle treatments. In all cases the first letter of any code represents the three study blocks, namely north (N), central (C) and south (S). All other letters represent the animals present, where O = control, C = cattle, W = meso-herbivores, and M = mega-herbivores. that this group also had effects on spiders. Earlier research in the KLEE experiment has shown that exclosure of ungulates (control plots) resulted in a 60% increase in the total number of small mammals (Keesing 2000). In most cases, mega-herbivores (elephant, gi- raffe) influence the type of habitat under study by browsing its shrub and tree layer (Dublin 1995). Perhaps both mega-herbivores and meso-herbivores have little effect in the cur- rent study because they have low densities compared to cattle. It is already documented that most wildlife in Laikipia lives outside na- tional parks (Western 1989; Mbugua 1986; LWF 1996). However, the densities of wildlife on ranches are considerably lower than that of livestock. This may be why only cattle den- sities were high enough to cause a statistically significant effect on the relative vegetation cover and, by extension, on the spider com- munity. The diversity indices from pitfall-trapping samples were not significantly related to rel- ative vegetation cover unlike those from sweep-netting samples. Such difference be- tween the two methods may be caused by the difference in biology of the species targeted by the two methods. It was possible that sweep-netting mainly caught foliage dwelling spiders, which were likely to be affected by changes in vegetation cover more than ground living spiders that dominated the pitfall trap samples. The influence of experimental treatments or abiotic environmental variables could be test- ed for only a few abundant species. Cyclosa insulana reacted to changes in relative vege- tation cover, while R. flavida and Aelurillus sp. were more sensitive to seasonal changes. All four species including A. trifasciata, were significantly affected by the presence of cattle but in different ways. Aelurillus sp. was more abundant in plots grazed by cattle, while the reverse was true for the other three species. The specific behavior of each species (e.g., its way of acquiring food), or the kind of habitat where it lives may explain this difference. Ae- lurillus is a ground-active jumping spider that 276 THE JOURNAL OF ARACHNOLOGY Table 3. — Analysis of covariance (ANCOVA) to establish the effects of the factors 'meso-herbivores’ (levels: absent [O and C] vs. present [W and WC]) and ‘cattle’ (levels: absent [O and W] vs. present [C and WC]) and two co variates, relative vegetation cover and total monthly rainfall, on the abundance of Cyclosa insulana, Argiope trifasciata, Runcinia flavida and Aelurillus sp recorded at Laikipia in 2001- 2002. The codes for the above abbreviations are such that O = control (no large mammalian herbivores); (W) = meso-herbivores; (M) = mega-herbivores and (C) = cattle. * = Significant at a = 0.05. Mean abundance ± SE Effect Absent Present df MS F-value F-value Cyclosa insulana Intercept Relative vegetation cover Total monthly rainfall Cattle Meso-herbivores Cattle*Meso-herbivores Error Argiope trifasciata Intercept Relative vegetation cover Total monthly rainfall Cattle Meso-herbivores Cattle*Meso-herbivores Error Runcinia flavida Intercept Relative vegetation cover Total monthly rainfall Cattle Meso-herbivores Cattle*MesO“herbivores Error Aelurillus sp Intercept Relative vegetation cover Total monthly rainfall Cattle Meso-herbivores Cattle* Meso-herbivores Error 1 1 1 1.73 ± 0.06 1.94 ± 0.05 1 1.82 ± 0.06 1.99 ± 0.05 1 1.89 ± 0.09 1.96 ± 0.07 1 498 1 1 1 1.01 ± 0.02 0.88 ± 0.02 1 0.99 ± 0.03 0.92 ± 0.01 1 1.00 ± 0.03 1.04 ± 0.06 1 498 1 1 1 1.16 ± 0.03 1.00 ± 0.02 1 1.07 ± 0.03 1.08 ± 0.02 1 1.02 ± 0.04 1.11 ± 0.05 1 498 1 1 1 1.05 ± 0.03 1.21 ± 0.03 1 1.08 ± 0.03 1.15 ± 0.02 1 0.98 ± 0.04 1.18 ± 0.05 1 498 107.23 128.15 <0.01* 41.46 49.55 <0.01* 2.39 2.86 0.09 3.52 4.21 0.04* 0.42 0.51 0.48 2.82 3.36 0.07 0.84 5.09 32.64 <0.01* 0.00 0.01 0.92 0.54 3.46 0.06 1.47 9.44 0.02* 0.49 3.15 0.08 0.04 0.28 0.60 0.16 6.06 25.29 <0.01* 0.58 2.43 0.12 3.75 15.64 <0.01* 1.27 5.28 0.02* 0.09 0.38 0.54 0.25 1.04 0.31 0.23 8.54 37.44 <0.01* 0.02 0.09 0.77 0.89 3.89 0.04* 2.84 12.46 <0.01* 0.63 2.75 0.09 0.11 0.49 0.48 0.23 does not build webs to catch prey but chases and jumps onto prey. It seems likely then that it thrived well where there was more grazing and more open ground, compared to a web- builder like Argiope that preferred a complex habitat where it could find vegetation to an- chor its web. Since Aelurillus is known to feed on ants, perhaps grazing makes ants more abundant and this in turn makes Aelurillus in- crease in abundance. Other related studies on individual species have shown that species level of resolution has a limitation when used for such analysis since a single species toler- ant of a perturbation might strongly influence the results (Caro and O' Doherty 1999). This was noted in the current study, where C in- sulana was found to be very dominant. The pattern shown by MDS analysis (Fig. 5) seems to correspond with the relative veg- etation cover distribution pattern, which is found to be lov/er in grazing plots and higher in control plots. This could mean that the spi- WARUI ET AL.-EFFECTS OF MAMMAL GRAZING ON SPIDERS IN KENYA 277 der community was responding to habitat complexity, including the factor “vegetation cover.” As already explained, control plots had the highest relative cover followed by meso- and mega-herbivore plots, while cattle plots had the lowest cover. The non-cattle grazing plots had intermediate vegetation cov- er, probably because wildlife were rarer than cattle in the experimental plots. This general trend of the spider community to cluster along control, non-cattle grazing and cattle grazing zones in an MDS analysis (al- though true for only the herb layer fauna) agrees with earlier studies indicating that hab- itat complexity influences the distribution of spiders of the herb layer. For example, work by Halaj et al. (2000) reported that structural habitat complexity had a profound effect on canopy spiders and other arthropods. Rypstra (1983) and Wise (1993) concluded that spider populations are limited by the availability of unique structural features in the habitat rather than by the abundance of prey. Exclosure treatments allowed us to detect changes in plant cover, and showed them to be significant in plots with cattle grazing. Plant cover appears to significantly affect spi- der diversity. Overall, activity by wildlife (mega- and meso-herbivores) had less (non- significant) effect on plant cover and spider diversity compared to that of cattle. The spi- der fauna of the black cotton soil savanna hab- itat is sufficiently rich to be useful for biolog- ical monitoring work in the sense of Kremen et al. (1994), who stated that: “the importance of monitoring is to come up with indicators that respond to anthropogenic disturbances early enough before changes manifest them- selves in the more complex food webs and food chains and even affect the long living organisms,” ACKNOWLEDGMENTS We thank Dr. Nick Georgiadis, Ken and John Wreford-Smith, Fredrick Erei, Kerry Outram and especially the late George Small. We also thank the entire Mpala Research Cen- ter community for contributing to our field work, especially our research assistants: P. Lenguya, E Erei, Hussein M., Abulkadir M. and R. Mwakondi; Dr. Ansie Dippenaar- Schoeman (ARC-Pretoria), Dr. Tony Russell- Smith (Formerly of Natural Resources Insti- tute, University of Greenwich, UK), and Dr. Wanja Kinuthia (National Museums of Kenya, Nairobi) for their moral support; and Sue Abraham (Rhodes University Graphics Ser- vice) for preparing some of the figures. The exclosure plots were built and maintained with grants from James Smithson Fund of the Smithsonian Institution (to Alan Smith), the National Geographic Society (4691-91), the National Science Foundation (#BSR 97-07477 and BSR-03- 16402), and the African Elephant program of the U.S. Fish and Wildlife Service (98210-0-G563)_(to TP. Young). Other fund- ing for this research came from Columbus Zoological Park Association, Inc., Lincoln Park Zoo Africa/Asia Fund and Royal Muse- um for Central Africa, Tervuren, Belgium (to C.M. Warui) and Rhodes University (to Prof M. Villet). I am grateful to all of them for their financial support. Last but not least to the National Museums of Kenya for their offer of research facilities. This research was car- ried out under the auspices of the Mpala Re- search Centre and The Office of the President of the Republic of Kenya (Ref. OP/13/001/8C 20). LITERATURE CITED Abensperg-Traun, G.T., G.T. Smith, G.W. Arnold & D.E. Stevenson. 1996. The effects of habitat fragmentation and livestock-grazing on animal communities in remnants of gimlet Eucalyptus salubris woodland in the Western Australian wheatbelt. 1. Arthropods. Journal of Applied Ecology 33:1281-1301. Abrous-Kherbouche, O., R. Jocque & J.P. Maelfait. 1997. Les effets du paturage intensif sur I’araneofaune dans la region de tala-Guilef (Parc National du Djurdjura, Algerie). Bulletin et An- nales de la Societe royale Beige d’Entomologie 133:71-90 Ahn, P.M. & L.C. Geiger. 1987. Soils of Laikipia District. National Agricultural Research Labora- tories, Nairobi, Kenya, 119 pp. Blandin, P. & M.-L. Celerier. 1981. Les Araignees des savanes de Lamto (Cote- dTvoire). Organi- sation des peuplements, bilans energetiques, place dans I'ecosysteme. Publications du Labor- atoire Zoologie de I’Ecole normale Superieure 21:1-586. Caro, T.M. & G.O Doherty. 1999. On the use of surrogate species in conservation biology. Con- servation Biology 13:805-814. Churchill, T.B. 1998. Spiders as ecological indica- tors in the Australian tropics: family distribution patterns along rainfall and grazing gradients. Pp. 325 — 330. In Proceedings of the 17'*" European 278 THE JOURNAL OF ARACHNOLOGY Colloquium of Arachnology, Edinburgh. (RA. Selden, ed.). 14-18 July 1997. Clarke, B. & R. Gorley. 2001. PRIMER 5 for win- dows version 5.24. Plymouth, PRIMER-E Ltd. [Software]. Clarke, B. & R.M. Warwick. 1994. Change in ma- rine communities; an approach to statistical anal- ysis and interpretation. Plymouth, Plymouth Ma- rine Laboratory, 133 pp. Coddington, J.A., L.H. Young & EA. Coyle. 1996. Estimating spiders’ species richness in a South- ern Appalachian cove hardwood forest. Journal of Arachnology 24:1 1 1-128. Curtis, D.J., E.J. Curtis & D.B.A. Thompson. 1990. On the effects of trampling on montane spiders and other arthropods. Bulletin de la Societe Eu- ropeenne d’Arachnologie 1:103-109. Dippenaar-Schoeman, A.S. & A. Leroy. 2003. A check list of the spiders of the Kruger National Park, South Africa (Arachnida: Araneae). Koe- doe 46:91-100. Dippenaar-Schoeman, A.S., A. Leroy, M. de Jager & A. Van den Berg. 1999. Spider diversity of the Karoo National Park, South Africa (Arachnida). Koedoe 42:31-42. Dippenaar-Schoeman, A.S., A.M. van den Berg & A. van den Berg. 1989. Species composition and relative seasonal abundance of spiders from the field and tree layers of the Roodeplaat Dam Na- ture Reserve. Koedoe 32:25-38. Dublin, H.T. 1995. Vegetation dynamics in the Ser- engeti-Mara ecosystem: the role of elephants, fire and other factors. Pp. 71 — 90. In Serengeti II: Dynamics, Management, and Conservation of an Ecosystem. (A.R.E. Sinclair & R Arcese, eds.). Chicago: University of Chicago Press. Fabricius, C. 1997. The impact of land use on bio- diversity in xeric succulent thicket. South Africa. Ph.D. Thesis, University of Cape Town. Fabricius, C., A.R. Palmer, & M. Burger 2002. Landscape diversity in a conservation area and commercial and communal rangeland in xeric succulent thicket. South Africa. Landscape Ecol- ogy 17:531-537. Foord, S.H., A.S. Dippenaar-Schoeman & M. van der Merwe. 2002. A check list of the spider fau- na of the western Soutpansberg, South Africa (Arachnida: Araneae). Koedoe 45:35-43. Gibson, C.W.D., V.K. Brown, L. Losito & G.C. McGavin. 1982. The response of invertebrate as- semblies to grazing. Ecography 15:166-176. Gibson, C.W.D., C. Hambler & V.K. Brown. 1992. Changes in spider (Araneae) assemblages in re- lation to succession and grazing management. Journal of Applied Ecology 29:132-142. Green, J. 1999. Sampling method and time deter- mines composition of spider collections. Journal of Arachnology 27:176-182. Greenslade, P.J.M. 1964. Pitfall trapping as a meth- od for studying populations of Carabidae. Jour- nal of Animal Ecology 33:301-310. Gunnarsson, B. 1988. Spruce-living spiders and for- j, est decline; the importance of needle loss. Bio- logical Conservation 43:309-319. Halaj, J., D.W. Ross & A.R. Moldenke. 2000. Im- i portance of habitat structure to the arthropod food-web in Douglas-fir canopies. Oikos 90: h 139-152. I Keesing, F. 2000. Cryptic consumers and the ecol- ogy of an African savanna. BioScience 50:205- 215. Kremen, C., R.K. Colwell., TL. Erwin., D.D. Mur- phy., R.E Noss & M.A. Sanjayan. 1994. Terres- trial arthropod assemblages. Their use in conser- vation planning. Conservation Biology 7:796- 808. [LWF] Laikipia Wildlife Forum. 1996. A total count of herbivores in Laikipia District, prelim- inary results. Laikipia (Kenya), Laikipia Wildlife Forum, unpublished report. Lotz, L.N., M.T Seaman & D.J. Kok. 1991. Sur- | face-active spiders (Araneae) of a site in a semi- arid central Africa. Navorsinge van die Nasionale Museum Bloemfontein 7:529-540. Mbugua, S.W. 1986. Monitoring livestock and wildlife in Kenya. Pp. 413 — 430. In Range De- velopment and Research in Kenya. (R.M. Han- sen, B.M. Woie & R.D. Child, eds.). Morrilton, Arizona: Winrock International Institute for Ag- riculture Development. New, TR. 1999. By-catch ethics and pitfall traps. Journal of Insect Conservation 3:1-3. Rambo, J.L. & S.H. Faeth. 1998. Effects of verte- brate grazing on plant insect community struc- ture. Conservation Biology 13:1047-1054. Rivers-Moore, N.A. & M.J. Samways. 1996. Game and cattle trampling, and impacts of human dwellings on arthropods at a game park bound- ary. Biodiversity and Conservation 5:1545- 1556. Russell-Smith, A. 1981. Seasonal activity and di- versity of ground-living spiders in two African savanna habitats. Bulletin of British Arachnol- ogical Society 5:145-154. Russell-Smith, A. 1999. The spiders of Mkomazi Game Reserve, Pp 197 — 222. In Mkomazi: the Ecology, Biodiversity and Conservation of a Tanzanian Savanna. (Coe, M., N. McWilliam, G. Stone & M. Parker, eds.). Royal Geographic So- ciety, London. Russell-Smith, A., J.M. Ritchie & N.M. Collins. 1987. The surface active spider fauna of arid bushland in Kora Reserve, Kenya. Bulletin of British Arachnological Society 7:171-174. Rypstra, A.L. 1983. The importance of food and space in limiting web-spider densities: a test us- ing exclosures. Oecologia 59:312-316. Rypstra, A.L., P. E. Carter, R.A. Balfour & S.D. WARUI ET AL.-EFFECTS OF MAMMAL GRAZING ON SPIDERS IN KENYA 279 Marshall. 1999. Architectural features of agri- cultural habitats and their impact on the spider inhabitants. Journal of Arachnology 27:371-377. Scharff, N. & C.E. Griswold. 1996. Biodiversity and biogeography of the spider fauna of the East- ern Arc Mountains, Tanzania. Preliminary report, 10 pp. Seymour, C. 1998. Different grazing intensities in arid rangelands: effects on invertebrates on a communal farm in the succulent Karoo, South Africa. MSc. Thesis, University of Cape Town. Seymour, C.L., & W.R.J. Dean. 1999. Effects of heavy grazing on invertebrate assemblages in the Succulent Karoo, South Africa. Journal of Arid Environments 43:267-286. Skarpe, C. 1991. Impact of grazing in savanna eco- systems. Ambio 20:351-356. Spence, J.R. & J.K. Niemela. 1994. Sampling ca- rabid assemblages with pitfall traps: the madness and the method. The Canadian Entomologist 126:881-894. StatSoft, Inc. 1999. STATISTICA for Windows. Computer program manual. Oklahoma, Tulsa. Taiti, S.W. 1992. The vegetation of Laikipia district, Kenya. Laikipia-Mt. Kenya Papers, B-2. Laikipia Research Programme, Universities of Nairobi and Bern, pp 21. Uetz, G.W. 1991. Habitat structure and spider for- aging. Pp 325 — 348. In Habitat Structure. The Physical Arrangement of Objects in Space. (Bell, S.S., E.D. McCoy & H.R. Mushinsky, eds.). Chapman & Hall, London. Uetz, G.W. & J.D. Unzicker. 1976. Pitfall trapping in ecological studies of wandering spiders. Jour- nal of Arachnology 3:101-111. Van den Berg, A. & A.S. Dippenaar-Schoeman. 1991. Ground-living spiders from an area where the harvester termite Hodotermes mossambicus occurs in South Africa. Phytophylactica 23:247- 253. Warui, C.M., M.H. Villet & TP. Young. 2004. Spi- ders (Araneae) from black cotton soils of a high- land savanna in Laikipia, central Kenya. Journal of Afrotropical Zoology 1 : 1 3-24. Western, D. 1989. Conservation without parks: wildlife in the rural landscape. Pp 158 — 165. In Conservation for the twenty-first century. (D. Western & M.C. Pearl, eds.). Oxford: Oxford University Press. Whitmore, C., R. Slotow, TE. Crouch & A.S. Dip- penaar-Schoeman. 2001. Checklist of spiders (Araneae) from a savanna ecosystem. Northern Province, South Africa: including a new family record. Durban Museum Novitates 26:10-19. Wise, D.H. 1993. Spiders in Ecological Webs. Cam- bridge, Cambridge University Press, pp 328. Woldu, Z. & M.A. Saleem. 2000. Grazing induced biodiversity in the highland ecozone of E. Africa. Agriculture, Ecosystems and Environment 79: 45-52. Young, TP, N. Patridge & A. Macrae. 1995. Long term glades in Acacia bushland and their edge effects in Laikipia, Kenya. Ecological Applica- tions 5:98-108. Young, T.P., B.D. Okello, D. Kinyua & TM. Palm- er. 1998. KLEE: A long-term multi-species her- bivore exclusion experiment in Laikipia, Kenya. African Journal of Range and Forage Science 14: 94-102. Manuscript received 5 April 2005, revised 1 Sep- tember 2005. 2005. The Journal of Arachnology 33:280-289 SPIDER (ARANEAE) COMMUNITIES OF SCREE SLOPES IN THE CZECH REPUBLIC Vlastimil RMicka: Institute of Entomology, Czech Academy of Sciences, Branisovska 31, CZ-370 05 Ceske Budejovice, Czech Republic Leos Klimes: Institute of Botany, Czech Academy of Sciences, Dukelska 135, CZ- 379 01 Tfeboh, Czech Republic ABSTRACT. We assessed the effects of environmental factors on spider communities in screes (sloping mass of coarse rock fragments) of the Czech Republic, based on catches from 325 pitfall traps, exposed for 177-670 days, from 1984-2000. Bootstrap resampling was applied to test for fuzziness of the partitions in cluster analysis of the samples. Two distinct spider communities were identified. The first one was confined to sites where ice is formed and persists until late summer or over the whole year. This community consists of numerous relict spiders, such as Bathyphantes simillimus buchari Ruzicka 1988, Diplocentria bidentata (Emerton 1882) and Lepthyphantes tripartitus Miller & Svaton 1978, possibly persisting in these cold screes from the early postglacial period. The other community included all other sites, irrespective of their environmental characteristics. Monte Carlo simulations were used to test the significance of en- vironmental factors and their interactions on the studied communities. Ice formation near the traps and position of the traps within individual screes were the most significant factors, followed by the depth of the traps within the scree, diameter of stones forming the scree, and altitude. A marginally significant effect was found for organic content in the scree matter, whereas presence of trees and phytogeographical districts appeared non-significant. Our analyses support the view that spiders inhabiting cold screes in Central Europe belong to a unique relict community of species requiring cold and stable microclimate. Keywords; Scree slopes, environmental factors, ice formation, CCA, Monte Carlo simulations At middle elevations scree slopes (or talus, an accumulation of coarse rock debris that rests against the base of an inland cliff; Allaby & Allaby 2003) represent eommon terrain forms in larger parts of Europe. Most stone aecumulations result from frost weathering of primarily compact rocks. These serees are widespread especially in the subarctic and in mountains in mid-latitudes, where a perigla- cial climate (i.e., climate of areas adjacent to a glacier or ice sheet) prevailing in the recent geological past promoted their development. Screes situated at middle elevations have at- tracted the attention of ecologists only recent- ly, perhaps due to logistic constraints (difficult access, a permanent danger of falling stones, landslides, etc.). Relatively small cavities sit- uated deeply in the scree are usually covered by unstable layers of stones, making non-de- structive studies rather difficult or almost im- possible. Further, the low densities of most in- vertebrates colonizing the inner parts of screes make short-term studies inefficient. In spite of these difficulties, current ecological research showed that some of the screes host peculiar and surprisingly species-rich fauna of inver- tebrates. Recent ecological studies demon- strated that the screes represent island-like ecosystems, supporting species which do not occur in the surrounding areas (Moseler & Molenda 1999; Kubat 2000). Repeated field observations supported by microclimatical measurements have shown that some screes function as large coolers, accumulating cold air that persists in some parts of the screes over the whole season. The thermic regime of the screes is often extremely conservative, largely independent of temperature fluctua- tions above-ground, both within and between years. At the bottom of some screes ice is formed, sometimes persisting there over the whole year, even if temperatures above- ground are higher by 30 °C or even more (Gude et al. 2003). Therefore, the stability of the temperature regime with extremely cold temperatures throughout the year and perma- nently high humidity enables persistence of a unique community of invertebrates, including 280 RUZICKA AND KLIMES-SPIDERS OF SCREE SLOPES 281 several representatives of arctic fauna which retreated from other habitats in Central Europe more than 10,000 years ago. The research focused on the fauna of in- vertebrates in screes of the Czech Republic was considerably intensified after modified pitfall traps were developed (Ruzicka 1982, 1988b). These traps were exposed by the se- nior author for several months up to two years in most important scree localities in the coun- try over the course of the last two decades. Numerous surprising findings have been re- ported, based on this research, including twelve species of spiders, mites and diplopods new to the Czech Republic (Ruzicka 1988b, 1994, 2000, 2002; Ruzicka et al. 1989; Ruz- icka & Aetus 1998; Ruzicka & Hajer 1996; Zacharda 1993) and five species/subspecies new to science (Ruzicka 1988a; Zacharda 2000a, b, c). The results of the studies carried out on nearly 66 localities revealed also a con- siderable heterogeneity of spider communities inhabiting the screes. While some localities host numerous relict species (see Discussion), other sites are relatively poor in biogeograph- ically and ecologically interesting spiders. Also, a considerable variation was observed within individual localities, with great differ- ences between individual parts of the screes. These observations resulted in several ecolog- ical questions that we address in this paper: 1) What are the environmental factors respon- sible for the observed variation in species composition of spider assemblages in the screes? 2) Is the occurrence of relict species in screes correlated with some environmental factors? 3) Is there a sharp boundary between spider communities of cold and warm sites within individual screes? To answer these questions we compiled all available records of spiders from screes of the Czech Republic, obtained by pitfall traps ex- posed for a longer time, and added available information on environmental factors, either measured or estimated in other ways at sites where the spiders were collected. This result- ed in a relatively large and complete dataset which we used in the analysis. METHODS Study area and localities. — The Czech Republic is situated in the temperate zone of Europe between 48° 33' and 51° 03' N, and 12° 05' and 18° 51' E. Major parts of the Czech Republic belong to the Proterozoic and Palaeozoic Bohemian Massif, only the east- ernmost part is pervaded by the Tertiary mountain system of the western Carpathians. The Bohemian Massif was long ago trans- formed by erosion into a levelled terrain, which was, during the Alpine formation of mountains, disrupted by faults, fractures in rock strata; elevated crustal blocks formed mountain regions (particularly border moun- tains), and the volcanic region of Ceske Stfe- dohon Mts. in the north of the Czech Republic was formed. The system of deeply cut river valleys was formed during the Tertiary and, especially, during the Quaternary (Lozek 1988). Due to its geology and geomorpholo- gy, the territory of the Czech Republic is rich in various boulder accumulations (Ruzicka 1993). Material was collected from 66 localities distributed all over the Czech Republic (Fig. 1). Elevations of the localities ranged from 270-1550 m a.s.L The investigated screes are formed by andesite, basalt, conglomerate, limestone, phonolite, quartzite, sandstone, granite and other kinds of rock. The height of scree fields from the foot to the top varied between 10 and 250 m, slope angles ranged between 20° and 40°. Sampling. — The animals were usually trapped in modified pitfall traps made of rigid plastic (Ruzicka 1982, 1988b). The traps con- sisted of a board (20 x 25 cm), which forms an artificial horizontal surface (note that a flat horizontal soil surface is not present in scree slopes) and a can inserted in the centre of the board. Traditional pitfall traps (simple cans) were also used. The cans contained a mixture of 7% formaldehyde and 10% glycerol with a few drops of a surfactant. The traps were placed among the stones. Field research was conducted from 1984-2000. In total, 325 traps were installed, most of them (85%) for more than 300 days. The catch, especially from deeper scree lay- ers, is often poor in species. To obtain more representative samples, we combined catches from traps placed at the same position along the scree slope in individual localities. This resulted in 128 samples. Environmental characteristics. — In total, eight environmental characteristics were reg- istered: elevation (m a.s.L); scree type (1 = bare scree slopes, 2 = scree slopes partly 282 THE JOURNAL OF ARACHNOLOGY Figure 1. — Location of the studied screes on a grid map of the Czech Republic. The circles represent one or several localities used in the analysis. overgrown by solitary trees, 3 = scree forests, corresponding to a gradient from bare to forest screes); position of pitfall traps along the tem- perature gradient in scree field (1 = lower margin, 2 = middle part, and 3 = upper mar- gin of the scree); ice formation near the trap ( 1 = no ice formation, 2 = temporal ice inside scree, melting in summer, 3 = permanent ice forming permafrost-like conditions); typical size of stones in the scree (diameter ranging from 0.1-10 m); depth below the surface in which the trap was installed (ranged from 0- 5.0 m); substrate around the trap (1 = bare stones, 2 = soil, 3 = detritus, 4 = mosses, ranges from sterile to organic substrate); phy- togeographical district (1 = thermo-, 2 = meso-, 3 = oreophyticum, characterized by vegetation on a broader scale, Slavik 1984). In addition, three covariables were used: type of trap (1 = a pot, which was a low ef- hciency trap; 2 = a pot sunken into a board, which was considered a high efficiency trap); number of traps at a site (ranging from 1-1 1); number of days during which the trap was ex- posed. Data analysis. — In the numerical analyses we always used the whole data set, i.e., all localities and all species. The input data were log-transformed prior to analysis. Detrended Canonical Correspondence Analysis (DCA) was used to estimate species turnover along the main direction of variability. After that CCA (Canonical Correspondence Analysis) implemented into the CANOCO program (ter Braak & Smilauer 1998) was performed to test the effects of individual variables and their interactions on species composition. Monte Carlo simulations with 10,000 permu- tations were calculated to assess the signifi- cance of individual environmental factors and their interactions. In these analyses the factors which were not used as explanatory variables were defined as covariables, to remove their effects on the results and to obtain a net effect of individual environmental variables. Using this approach we could perform tests that are counterparts to ANOVA but for multivariate data. Spider communities were classified by a cluster analysis based on Ward’s method, us- ing Euclidean distances for quantitative data (Jongman et al. 1995, p. 178), calculated for log-transformed catches, using the Syn-tax program by Podani (2001). The sharpness of the resulting classification was tested using a bootstrap resampling, in which stability of partition at a given level was tested by resam- pling the original data, according to Pillar RUZICKA AND KLIMES-SPIDERS OF SCREE SLOPES 283 (1999). The outcome of this testing indicates whether groups in the partition reappear more often in resampling data than expected on a random basis. Nomenclature follows the catalogue of spi- ders of the Czech Republic (Buchar & Ruz- icka 2002). Species characteristics were also taken from this source. Voucher specimens are deposited in the collection of V. Ruzicka. RESULTS Site characteristics. — The strongest signif- icant correlation between pairs of the eight en- vironmental characteristics was found be- tween elevation and phytogeographical district (r “ 0.70). This reflects the crucial role of elevation in the distribution of plant com- munities at a broader scale. The position of the traps was strongly negatively correlated with the incidence of ice at the traps (r = -0.49), implying that traps situated at the lower margin of the screes were often sur- rounded by permanent ice whereas traps placed at the upper margin were free of ice. As expected, bare scree slopes were usually built of large boulders whereas forested screes developed on gravel screes (r = 0.40). Further, scree sites with large amounts of organic mat- ter and covered by mosses usually developed in warm regions (r = —0.32), at lower ele- vations (r = —0.34) and at the lower margin of the screes (r = —0.35). Ice formation was negatively correlated with elevation (r ~ —0.20). Even if elevations of the sites (270-1550 m) spanned the range of elevations in the Czech Republic almost completely, surprisingly the screes with ice formation were found at rather low elevations; the screes with permanent ice filling were sit- uated at 350-650 m a.s.l., the screes with tem- poral ice filling at 270-700 m a.s.l. The ice was found more often in sterile screes without organic matter and deeper in the scree than in screes with organic matter and near traps sit- uated closely to the scree surface (r — 0.26 and r — —0.21, respectively). All these cor- relations were significant at P < 0.05. Species composition. — In total, 1047 spi- ders were captured, belonging to 176 species of 22 families. Based on our knowledge on ecological demands of all spider species in the Czech Republic (Buchar & Ruzicka 2002), the following sets of species can be identified among the captured spiders: 1. Species occurring exclusively in bare scree slopes (and in adjacent underground spaces) and in scree forests: Acantholycosa norvegica (Thorell 1872), Bathyphantes sim- iliimus buchari Ruzicka 1988, Clubiona al- picola Kulczyhski 1882, Comaroma simoni Bertkau 1889, Diplocentria bidentata (Emer- ton 1882), Kratochviliella bicapitata Miller 1938, Lepthyphantes notabilis Kulczyhski 1887, Lepthyphantes improbulus Simon 1929, Lepthyphantes zimmermanni Bertkau 1890, Liocranum rutilans (Thorell 1875), Meta men- ardi (Latreille 1804), Micrargus apertus (O. R-Cambridge 1871), Neon levis (Simon 1871) , Pholcomma gibbum (Westring 1851), Porrhomma myops Simon 1884, Porrhomma rosenhaueri (L. Koch 1872), Rugathodes bel- licosus (Simon 1873), Saaristoa firma (O. R- Cambridge 1905), Trogloneta granulum Simon 1922, Wubanoides uralensis (Rakhorukov 1981). 2. Species of scree slopes, occurring also in other habitats (in brackets): Lepthyphantes le- prosus (Ohlert 1865), Liocranum rupicola (Walckenaer 1830), Nesticus cellulanus (Clerck 1757), Pholcus opilionoides (Schrank 1781), Sitticus pubescens (Fabricius 1775) (synanthropic), Ceratinella major Kulczyhski 1894, Megalepthyphantes collinus (L. Koch 1872) , Tegenaria silvestris L. Koch 1872 (for- ests), Lepthyphantes tripartitus Miller & Sva- ton 1978, Theonoe minutissima (O. R. -Cam- bridge 1879) (peat bogs), Agraecina striata (Kulczyhski 1882) (lowland forests), Wal- ckenaeria capita (Westring 1861) (rock steppes), Cryphoeca silvicola (C.L. Koch 1834) (spruce forests), Porrhomma egeria Si- mon 1884 (caves and subalpine belt). The cluster analysis of samples revealed two distinct groups, indicating that two clearly separated spider communities can be identi- fied in the screes (Fig. 2). The bootstrap re- sampling showed a partitioning at this level. For lower levels, i.e. when considering parti- tioning to a higher number of clusters, we ob- tained non-significant results. Accordingly, all clusters except for those labelled in Fig. 2 A and B should be interpreted as fuzzy, not dis- tinctly separated from each other. All ten sam- ples from sites at which permanent ice was observed and most samples from sites at which temporal ice formation was registered were included in cluster A. The other cluster, B, included all other localities. 284 THE JOURNAL OF ARACHNOLOGY 14 12 10 Zi 4 2 ICE DB BSB LT Figure 2. — Spider communities in Czech screes, as revealed by a cluster analysis of samples. The A and B clusters were supported by bootstrap resampling. ICE = ice formation (2 = temporal ice inside scree, melting in summer; 3 = permanent ice forming permafrost). Presence of three indicators of ice incidence (DB = Diplocentha bidentata, BSB — Bathyphantes simillimus buchari, LT = Lepthyphantes tripartitus) is indicated by stars. The longest gradient in DCA was 6.4, in- dicating that the response model suitable for the analysis is unimodal. Therefore, we used CCA for direct and CA for indirect gradient analysis. First four axes of the CCA explained 8.9 % of total variance and a test of all ca- nonical axes was strongly significant (P = 0.001). The CCA ordination of samples (Fig. 3) clearly separated localities belonging to the two clusters of localities. The overlap between envelopes encompassing localities belonging to the two clusters was relatively large, how- ever, localities belonging to cluster A showed clearly higher scores of the first ordination axis in comparison with localities belonging to cluster B. While direct gradient analyses, such as CCA, are searching for the pattern caused by environmental variables, its coun- terpart, correspondence analysis (CA), reflects similarity in species composition and abun- dance of individual samples. The CA ordina- tion (not shown) revealed a pattern very sim- ilar to that of CCA, and the amount of variation explained by the first ordination axis was similar (3.7 and 2.7 %, respectively). This indicates that the environmental factors used in the analysis belong to those that are the most important for the pattern of similarity among samples. The CCA analysis used to test the effect of individual variables showed that ice formation and position of the traps on the scree slopes had the strongest effects on species composi- tion and abundance of spiders. The effect of the two factors was strongly significant, as documented by the ordination diagram. The depth at which the traps were placed also showed a strong effect on species composi- tion, similarly to elevation and stone diameter. The last factor showing a significant effect was the substrate, however, its effect was only marginally significant. Several interactions also played a significant role (Table 1). Even if their effect was usually less strong than that of the above mentioned variables, they were still significant, suggesting that species com- position and abundance of spiders in screes is affected by numerous factors, often interact- ing with each other in a complicated way. Therefore, it is not surprising that spider com- munities of screes are not sharply delimited and, except for the two largest groups, they show a fuzzy pattern with little distinctness. Spiders and environmental factors* — The RUZICKA AND KLIMES-SPIDERS OF SCREE SLOPES 285 Figure 3. — CCA ordination of samples. Arrows indicate the effect of environmental variables. Full circles = samples belonging to A cluster in Fig. 2, crosses = samples belonging to B cluster. ELE: elevation; ICE: ice formation near the trap; DEPTH: depth below the surface, in which the trap was installed; POS: position of pitfall trap along the temperature gradient in scree field; DIAM: typical size of stones; SUBSTR: substrate around the trap; PHYTO: phytogeographical district; SCREETYP: from bare scree to scree forest. relatively high number of samples scattered over the whole Czech Republic enabled a more detailed analysis of the effect of envi- ronmental factors on individual species. Analyses of the relationship between envi- ronmental variables and the number of cap- tured spiders of individual species (linear re- gression) showed that species found mainly at higher elevations included Clubiona alpicola (only above 700 m a.s.L), Wubanoides ura- lensis (only above 930 m), Bathyphantes sim- illimus, and Walckenaeria capita. Several spe- cies were caught mainly at lower elevations: Lepthyphantes improbulus (up to 400 m a.s.L), Pholcus opilionoides (up to 750 m), Liocranum rupicola (up to 800 m) and Phol- comma gibbum. Metellina merianae occurred primarily in scree forests; Acantholycosa norvegica, Lep- thyphantes notabilis, Clubiona alpicola were recorded exclusively on bare, open scree slopes. 286 THE JOURNAL OF ARACHNOLOGY Table 1 . — Results of the CCA analyses applied to log abundances of spiders caught in pitfall traps, r: species-environment correlation on the first axis, var: percentage of species variability explained by the first ordination axis. F: the F-ratio statistics for the test on the trace. P: corresponding probability value obtained by the Monte Carlo permutation test. The variables not used as explanatory variables in individual analyses, were used as covariables. Type of the trap, number of traps at a site and number of days for which the trap was exposed were used as covariables in all analyses. ELF: elevation; ICE: ice formation near the trap; DEPTH: depth below the surface, in which the trap was installed; POSITION: position of pitfall trap along the temperature gradient in scree field; DIAM: typical size of stones; SUBSTR: substrate around the trap; PHYTO; phytogeographical district; SCREETYP: from bare scree to scree forest. Explanatory variables r var F P ELE 0.812 1.5 1.67 0.0017 ICE 0.828 1.7 1.84 0.0001 DEPTH 0.833 1.4 1.54 0.0003 POSITION 0.803 1.7 1.87 0.0001 DIAM 0.809 1.4 1.59 0.0089 SUBSTR 0.8 1.2 1.3 0.0497 PHYTO 0.732 0.9 1.03 0.3920 SCREETYP 0.767 0.9 0.95 0.5951 ELE X DIAM 0.778 1.3 1.47 0.0215 DIAM X ICE 0.843 1.4 1.54 0.0185 SUBSTR X ICE 0.847 1.3 1.42 0.0192 ELE X DEPTH 0.809 1.3 1.38 0.0323 SUBSTR X DEPTH 0.81 1.2 1.3 0.0396 PHYTO X SUBSTR 0.774 1.1 1.23 0.1060 SCREETYPE X SUBSTR 0.783 1 1.07 0.3015 Lepthyphantes notahilis and Pholcus opi- lionoides colonized mainly upper scree mar- gins; Lepthyphantes tripartitus and Diplocen- tria bidentata occurred mainly at lower margins. The dependence on ice formation is sharp in some spiders: Rugathodes bellicosus, Lep- thyphtantes notabilis, Tegenaria silvestris, Nesticus cellulanus, Pholcomma gibbum, Meta menardi, Acantholycosa norvegica, Lio- cranum rupicola, Pholcus opilionoides avoid- ed sites with ice formation, whereas Lepthy- phantes tripartitus, Diplocentria bidentata and Bathyphantes simillimus buchari at lower elevations were confined to these sites. Diplocentria bidentata and Lepthyphantes tripartitus occurred together at sites where ice is formed. Along the gradient of substrate type, L. tripartitus preferred more detritus-rich sites, whereas D. bidentata colonised more mossy habitats. Diplocentria bidentata tended to occur at the surface, whereas L. tripartitus occurred in deeper layers. These trends were Linivocally supported by the separate sieving of moss and detritus on Klic Mt. on 12th Oc- tober 1999: 28 specimens of L. tripartitus and 6 specimens of D. bidentata were collected by detritus sieving, whereas 66 specimens of D. bidentata and 1 specimen of L. tripartitus were collected by moss sieving. Orb weavers Metellina merianae and Meta menardi preferred spaces among larger stones; in contrast, Lepthyphantes notabilis, Pholcom- ma gibbum, Acantholycosa norvegica were more abundant at sites with smaller stones. The species occurring mainly at the scree surface included Acantholycosa norvegica, Diplocentria bidentata, whereas Porrhomma myops, Rugathodes bellicosus, Meta menardi, Nesticus cellulanus, Wubanoides uralensis were found mainly in the depth of the screes. Finally, species found mainly at the surface of bare stones included Lepthyphantes nota- bilis, Clubiona alpicola, Rugathodes bellico- sus, Meta menardi, Theonoe minutissima, and Wubanoides uralensis. DISCUSSION Balch (1900) was possibly the first who documented ice formation in scree slopes at middle (but not at higher) elevations. We showed that ice is regularly formed in screes also from 270-700 m. According to Gude et al. (2003) lower parts of scree slopes are in- tensively cooled during short periods of win- ter frost. Cold air penetrates inside the screes RUZICKA AND KLIMES-SPIDERS OF SCREE SLOPES 287 only during periods with no (or limited) snow cover. Long-term data from three south Bo- hemian meteorological stations support this hypothesis. They indicate a negative relation- ship between the number of frost days without snow cover and altitude: there are 23 frost days without snow per year in Ceske Bude- jovice at 389 m, 19 such days at Kasperske Hory (737 m) and only 10 frost days without snow per year on Churahov at 1,118 m a.s.L Mountain scree slopes do not markedly cool in winter, because frost air cannot penetrate the scree slopes through the snow cover. Spatial heterogeneity of invertebrates in scree slopes was studied by Molenda 1989; Ruzicka et al. 1995; V. Ruzicka 1990, 1996, 2002; J. Ruzicka 1996. The position of a site along the scree slope was designated as the main factor influencing species distribution by Brabec (1973) in his pioneer study. We found that the effects of position of the trap on the scree slope and ice formation are strongly sig- nificant. Ice formation is a principal factor and ice is usually formed on the lower margin of scree slopes. However, concave slope forms can be formed by various slope denudation processes also in the middle part of a scree slope (Demek et al. 1975; Ruzicka 1999c). In such cases, ice can be formed also in the mid- dle part of a scree slope. Elevation patterns in spiders and mites of screes also have been documented by Ruzicka & Zacharda (1994) who focused on scree hab- itats in our highest mountains in the Krkonose National Park, and by V. Ruzicka (1996), who studied spiders in screes at low elevations of the Podyji National Park. Spatial distribution of spiders in screes was studied by Ruzicka 1999a, 2002 and Ruzicka et al. 1995. Temperature is a key factor re- sponsible for the presence or absence of spe- cies in individual parts along the slope. The dependence of L. tripartitus on detritus ex- plains the fact, that this species colonizes the whole profile of the lower margin of the screes, from the surface to the depth of about one meter, whereas D. bidentata, which is re- stricted to moss cushions, cannot colonize deeper layers. Land surfaces at higher latitudes in the northern hemisphere support a range of forest, scrub, tundra and peatland communities at the present day that may collectively be called the “coldland complex”. Physiognomically and floristically similar communities also occur at higher elevations of mountains further south (Tallis 1991). Current disjunct distribution of some spider species is a result of their with- drawal from Central Europe caused by chang- ing climatic conditions in the Pleistocene. Twenty-seven spider species of the Czech arachnofauna exhibit boreo-montane type of geographical distribution (Ruzicka in prep.). They occur in higher latitudes and have dis- junct, island-like populations in Central Eu- rope. The present findings indicate that some scree spiders in Central Europe could be re- garded as relicts of former climatic periods (‘‘glacial relicts”). Four of these species occur exclusively in scree slopes (Buchar & Ruzicka 2002). Having a distribution center in Siberia/ North Asia, Wubanoides uralensis and Acan- tholycosa norvegica occur only in several lo- calities in Central Europe (Schikora 2004; Marusik et al. 2003), independent of ice for- mation. Bathyphantes similUmus shows about the same general distributional pattern. In con- trast, Diplocentria bidentata has a Holarctic distribution. In Scotland, northern England and Wales it occurs locally with a low abun- dance in highlands (Harvey et al. 2002). In Central Europe it is known only from the peat bogs in Harz, Germany, situated at the highest elevations (Wiehle 1965); in the Czech Re- public on hilltops in the Krkonose Mountains (2 specimens) and on lower margin of frozen scree slopes (243 specimens). The occurrence of the latter two species is closely tied to ice formation in scree slopes. The same is true for the Central European mountain species Lep- thyphantes tripartitus. The occurrence of the three species at lower elevations is closely tied to the present periglacial temperature regime in frozen scree slopes, and the presence of these species indicates the palaeorefugial character of these habitats (Zacharda et al. in press), i.e. island-like habitats inhabited by populations of formerly more widespread spe- cies (Nekola 1999). Deep layers of screes represent shallow subterranean spaces, in which gradual adap- tation to the stable environment of deep sub- terranean spaces takes place (Ruzicka 1999b). Species, which preferentially colonise deep scree layers, exhibit leg elongation, depigmen- tation, body diminishing, and eye reduction (Ruzicka 1988a, 1990, 1998; Schikora 2004). 288 THE JOURNAL OF ARACHNOLOGY We found some of these species on several localities (R. bellicosus, B. s. buchari); on the other hand, the occurrence of Porrhomma myops and Comaroma simoni is known from one locality only. The reason for their rarity (a special combination of environmental fac- tors vs. our inability to penetrate more deep in scree?) remains unknown. ACKNOWLEDGMENTS This study was supported by the Grant Agency of the Czech Republic — project No. IAA6007401 and by the Institute of Entomol- ogy— project No. Z50070508 (VR) and proj- ect AV0Z6005908 of the Czech Academy of Sciences (LK). Thanks to Dr. Miloslav Za- charda for help with collecting spiders, Dr. Svatava Kfivancova for providing us with me- teorological data, Pierre Paquin and Soren Toft for valuable comments on a previous ver- sion of this paper. LITERATURE CITED Allaby, A. & M. Allaby. 2003. Oxford Dictionary of Earth Sciences. Oxford University Press, Ox- ford, 608 pp. Balch, E.S. 1900. Glacieres or freezing caverns. Al- len, Lane & Scott, Philadelphia, 377 pp. Brabec, E. 1973. The ecology of stony debris in the Ceske Stredohon Mt. — a second approximation. Thesis, Charles Univerity, Praha, 95 pp. (in Czech). Buchar, J. & V. Ruzicka. 2002. Catalogue of spiders of the Czech Republic. Peres, Praha, 351 pp. Demek, J., J. Pasek, & J. Rybar. 1975. The princi- ples of erosion-denudation slope processes. Stu- dia geographica 51:195-204 (in Czech). Gude, M., S. Dietrich, R. Mausbacher, C. Hauck, R. Molenda, V. Ruzicka & M. Zacharda. 2003. Probable occurrence of sporadic permafrost in non-alpine scree slopes in central Europe. Pp. 331-336. In Proceedings of the 8th International Conference on Permafrost. (M. Phillips, S. M. Springman & L. U. Arenson, eds.). Lisse (NL). Harvey, P.R., D.R. Nellist & M. G. Telfer. 2002. Provisional atlas of British spiders (Arachnida, Araneae). Biological Records Centre, Hunting- don, 406 pp. Jongman, R.H., C.J.E ter Braak, & O.F.R. van Ton- geren. 1995. Data analysis in community and landscape ecology. Wageningen, Pudoc, 299 pp. Kubat, K. (ed.). 2000. Stony Debris Ecosystems. Acta Universitatis Purkynianae, Biologica 4, 204 pp. Lozek, V. 1988. Inorganic nature in relation to flora and vegetation. Pp. 31-35, 156-157. In Kvetena Ceske socialisticke republiky [The flora of the Czech Socialist Republic]. (S, Hejny & B. Sla- vik, eds.). Academia, Praha, (in Czech, English summary). Marusik, Y.M., G.N. Azarkina & S. Koponen. 2003. A survey of East Palaearctic Lycosidae (Aranei). II. Genus Acantholycosa E Dahl, 1908 and re- lated new genera. Arthropoda Selecta 12(2): 101- 148. Molenda, R. 1989. Kafer in kaltlufterzeugenden Blockhladen — dkologische Untersuchungen an einem stark bewetterten Spaltenokosystem. 4 Rundschreiben der Arbeitsgemeinschaft Rhein- ischer Koleopterologen. Pp. 103-111. Moseler, B.M. & R. Molenda. (eds.). 1999. Le- bensraum Blockhlade. Decheniana-Beihefte 37. 170 pp. Nekola, J.C. 1999. Paleorefugia and neorefugia: the influence of colonization history on community pattern and process. Ecology 80:2459-2473. Pillar, V.D. 1999. How sharp are classifications? Ecology 80:2508-2516. Podani, L. 2001. Syn-Tax 2000. User’s Manual. Scientia, Budapest. Ruzicka, J. 1996. The beetles (Insecta: Coleoptera) in rock debris of the Plesivec hill (northern Bo- hemia, Ceske stredohon Protected Landscape Area). Klapalekiana 32:229-235 (in Czech, En- glish summary). Ruzicka, V. 1982. Modifications to improve the ef- ficiency of pitfall traps. Newsletter of the British Arachnological Society 34:2-4. Ruzicka, V. 1988a. Problems of Bathyphantes eu- menis and its occurrence in Czechoslovakia (Ar- aneae, Linyphiidae). Vestnrk Ceskoslovenske spolecnosti zoologicke 52:149-155. Ruzicka, V. 1988b. The longtimely exposed rock debris pitfalls. Vestnik Ceskoslovenske spolec- nosti zoologicke 52:238-240. Ruzicka, V. 1990. Spider (Araneae) communities of rock debris on a typical hillside in the Ceske Stredohon Mts. (North Bohemia). Acta Ento- mologica Bohemoslovaca 86[ 1 989] :4 19-431. Ruzicka, V. 1993. Structure and ecology of inver- tebrate communities of stony debris in Czech Re- public. Research project. Bulletin de la Societe neuchateloise des Sciences naturelles 116(1): 209-214. Ruzicka, V. 1994. Spiders in south Bohemian stony debris. Acta Universitatis Purkynianae, Biologica 1[1993]:9-18. Ruzicka, V. 1996, Species composition and site dis- tribution of spiders (Araneae) in a gneiss massif in the Dyje river valley. Revue suisse de Zool- ogie, voL hors serie:56 1—569. Ruzicka, V. 1998. The subterranean forms of Lep- thyphantes improbulus, Theonoe minutissima and Theridion bellicosum (Araneae: Linyphiidae, Theridiidae). Pp. 101-105. In Proceedings of the 17th European Colloquium of Arachnology. (P. RUZICKA AND KLIMES-SPIDERS OF SCREE SLOPES 289 A. Selden, ed.)- British Arachnological Society, Burnham Beeches, Bucks. Ruzicka, V. 1999a. Spiders (Araneae) of scree slopes in the Moravian Karst. Acta Musei Mo- raviae, Scientiae biologicae (Brno) 84:205-210. Ruzicka, V. 1999b. The first steps in subterranean evolution of spiders (Araneae) in Central Europe. Journal of Natural History 33:255-265. Ruzicka, V. 1999c. The freezing scree slopes and their arachnofauna. Pp. 141-147. In Lebensraum Blockhlade. (B. M. Moseler & R. Molenda, eds.). Decheniana-Beihefte 37. Ruzicka, V. 2000. Porrhomma myops and Porrhom- ma egeria in the Czech Republic. Acta Univer- sitatis Purkynianae, Biologica 4:149-150. Ruzicka, V. 2002. Spatial distribution of spiders (Araneae) on scree slopes in Kfivoklatsko and Moravsky Kras Protected Landscape Areas. Acta Societatis Zoologicae Bohemicae 66:321-328. Ruzicka, V. & P. Antus. 1998. Spiders in stony de- bris in central Bohemia. Acta Universitatis Pur- kynianae, Biologica 2:19-29. Ruzicka, V., J. Bohac, O. Syrovatka & L. Klimes. 1989. Invertebrates from rock debris in north Bo- hemia (Araneae, Opiliones, Coleoptera, Diptera). Sbomik Severoceskeho Muzea — Pnrodni Vedy, Liberec 17:25-36 (in Czech, English summary). Ruzicka, V. & J. Hajer. 1996. Spiders (Araneae) of stony debris in North Bohemia. Arachnologische Mitteilungen 12:46-56. Ruzicka, V., J. Hajer & M. Zacharda. 1995. Arach- nid population patterns in underground cavities of a stony debris field (Araneae, Opiliones, Pseu- doscorpionidea, Acari: Prostigmata, Rhagidi- idae). Pedobiologia 39:42-51. Ruzicka, V. & M. Zacharda. 1994. Arthropods of stony debris in the Krkonose Mountains, Czech Republic. Arctic and Alpine Research 26:332- 338. Schikora, H.-B. 2004. Wubanoides uralensis (Pak- horakov 1981)-Geographic variation, mating be- haviour, postembryonic development and de- scription of a new subspecies (Araneae, Linyphiidae). Pp. 327-341. In Diversitat und Biologie von Webspinnen, Skorpionen und an- deren Spinnentieren. (K. Thaler ed.). Denisia 12. Slavik, B. 1984. Grundlegende Phytochorotypen der Tschechischen Sozialistischen Republik. Preslia 56:241-265. Tallis, J.H. 1991. Plant Community History. Chap- man and Hall, London. ter Braak, C.J.F. & P Smilauer. 1998. CANOCO reference manual and users guide to “Canoco for Windows” software for canonical community or- dination, version 4. Microcomputer Power, Itha- ca. Wiehle, H. 1965. Die Spinnenfauna des Harzes. Na- tur und Museum 95(4): 133-142. Zacharda, M. 1993. Glacial relict Rhagidiidae (Ac- ari: Prostigmata) from superficial underground enclosures in the Krkonose Mountains, Czecho- slovakia, Journal of Natural History 27:47-61. Zacharda, M. 2000a. New and little-known species of Rhagidiidae from talus ecosystems, in the Czech Republic and Austria. Journal of Zoology, London 251:105-118. Zacharda, M. 2000b. New species of the rhagidiid genus Foveacheles (Acari: Prostigmata: Eupo- doidea) with a lyrifissure-like structure of the chelicerae. Journal of Natural History 34:247- 265. Zacharda, M. 2000c. New species of the genus Troglocheles (Acari: Prostigmata: Rhagidiidae) from Oetztaler Alps, Tyrol, with a key to adult species of the genus. Journal of Natural History 34:463-478. Zacharda, M., M. Gude, S. Krause, C. Hauck, R. Molenda & V. Ruzicka. In press. The Relict Mite Rhagidia gelida (Acari, Rhagidiidae) as Biolog- ical Cryoindicator of Periglacial Microclimate in European Highland Screes. Arctic, Antarctic and Alpine Research. Manuscript received 20 July 2005, revised 15 Sep- tember 2005. 2005. The Journal of Arachnology 33:290-299 FAUNISTIC SIMILARITY AND HISTORIC BIOGEOGRAPHY OF THE HARVESTMEN OF SOUTHERN AND SOUTHEASTERN ATLANTIC RAIN FOREST OF BRAZIL Ricardo Pinto-da-Rocha and Marcio Bernardino da Silva: Departamento de Zoologia, Institute de Biociencias, Universidade de Sao Paulo, Caixa Postal 11461, 05422-970, Sao Paulo, SP, Brazil. E-mail: ricrocha@usp.br Cibele Bragagnolo: Museu de Zoologia, Universidade de Sao Paulo, Brazil ABSTRACT. Harvestmen show a high degree of endemism in the Atlantic Rain Forest (eastern coast of Brazil). This biome shows the highest diversity of harvestmen inhabiting Brazil; 2/3 of the species are found in this area. Most of the species are distributed in a few thousand square kilometers, almost always within one mountain range. The similarities of 26 localities were studied, including sites from the Brazilian savanna, using data from recent collections (more than 8,000 specimens) and published data. A cluster analysis using Sprensens Coefficient indicated a high degree of endemism of species of harvestmen (sim- ilarity indexes below 0.5). It resulted in six main clusters related to the large mountain ranges and near sites. A high variation in richness was observed; 4-64 species per locality. The distribution of 84 species of four recently reviewed subfamilies of Gonyleptidae (Goniosomatinae, Caelopyginae, Progonylepto- idellinae and Sodreaninae) was studied. Eleven areas of endemism, with 3—14 endemic species each, were proposed. A primary Brooks Parsimony Analysis showed a possible first vicariant event splitting the fauna of two northern areas from the rest, and a second event splitting the fauna of southern areas (until 24°35"S) from those areas related to certain mountain ranges in the central Atlantic Rain Forest. The vicariant events were related to the uplifting of the Serra do Mar and the Serra da Mantiqueira, and the appearance of large rivers and climatic changes. Keywords: Atlantic Rain Forest, biodiversity. Brooks Parsimony Analysis, harvestmen, Neotropics. The Atlantic Rain Forest is located in the largest part of the Brazilian coastal region be- tween 6-30° S, also occupying the central to southern interior part of the country. This bi- ome comprises two types of vegetation for- mation: the Coastal Atlantic Rain Forest, close to the coast line, with elevations from sea lev- el to approximately 1,000 m a. s. 1., and with a hot, warm climate lacking a dry season; and the Atlantic Semi-deciduous Forest, which ex- tends across the plateau in the interior of the country (usually above 600 m elevation), that can have a severe dry season, normally be- tween April and September (Oliveira-Filho & Fontes 2000). The Atlantic Rain Forest was almost completely continuous in 1500, the year of the discovery of Brazil by Europeans, but is currently totally fragmented and re- duced to less than 7.6% of the original area (Morellato & Haddad 2000). This occun*ed because colonization was mainly on the coast and most state capitals are in this biome. We should stress that anthropogenic pressure is still strong on the remaining fragments. The few areas without or with a low anthropogenic pressure are in governmental reserves or in steep regions. Diversity in the Atlantic Rain Forest seems to be higher than in most parts of the Ama- zonian Rain Forest, and endemism is remark- able; 50% on an average and as high as 95% in some groups of amphibians according to Morellato & Haddad (2000). However, such statements are mainly based on data for plants and vertebrates, the invertebrates remaining poorly studied. An examination of the records of Laniatores harvestmen of the Atlantic Rain Forest (see the catalog of Kury 2003) and Eupnoi (Tourinho-Davis & Kury 2003; Tour- inho-Davis 2004) revealed that the group rep- resents an exclusive fauna, with the highest level of endemism (97.5%) in this biome. The opilionids are hygrophilous, have low vagility and are primarily nocturnal and cryp- 290 PINTO-DA-ROCHA ET AL.— BIOGEOGRAPHY OF HARVESTMEN 291 Colatina (20) Santa Teresa (32) Santa Leopoldina (18) PelrPf^olis (47) P.N.S^rra dos 6rgaos (64) E.E E.:Paraiso (19) lIL Figure L-— Richness (.between brackets) of Opiliones (Laniatores and Eupnoi) recorded in 26 localities in south and southern Brazil. tic. Apparently the Laniatores possess low ca- pability of dispersion. These aspects suggest the group is a good model for biogeographic studies. The only two studies dealing with op- ilionids which analyzed biogeography based on cladistic analysis were Briggs & Ubick (1989) and Ubick & Briggs (1989) for two genera of Laniatores, Phalangodidae, endemic to coastal California. Few studies deal with historic biogeographic aspects of the Atlantic Rain Forest. Some stud- ies related this biome as sister area with the Andean-Amazonian region (Amorim & Fires 1996) or as sister area with southern region (from Sao Paulo to Rio Grande do Sul) in case of groups with more southern occurrence in South America (Morrone et al. 1994; Perez- Losada et al. 2004). A few biogeographic hy- potheses based on phylogenetic reconstruc- tions were proposed for the Atlantic Rain Forest, as the comprehensive studies of Costa (1995) for fishes and Amorim & Pires (1996) for dipterans and monkeys. The main goals of this article are to demonstrate the great diver- sity of opilionids in the Atlantic Rain Forest including their high endemism, and to present a biogeographic hypothesis for the region. METHODS Similarity and richness. — The Sprensen index was applied to the analysis of similarity among the records of occurrence of 363 named species of Opiliones (Laniatores and Eupnoi) in 26 sites in south and southeastern Brazil. Morphospecies were not included be- cause they were not standardized among all sites. The most intensively sampled sites were chosen, using the following criteria: more than 200 specimens collected; or stability or little increasing of richness with recent collecting. The analyses were performed with the MVSP 3.1 software (Kovach Computing Services 1999). The records were obtained from the lit- erature (Laniatores from Kury 2003; Eupnoi from original descriptions and Tourinho-Davis 2004) in addition to museum records of the Museu de Zoologia da Universidade de Sao Paulo (MZSP), Institute Butantan (IBSP) and Museu Nacional do Rio de Janeiro. These col- lections include old material and 8,879 spec- 292 THE JOURNAL OF ARACHNOLOGY imens recently collected (2000-2004) for the project Biodiversity of Arachnida and Myri- apoda of the State of Sao Paulo (IBSP, MZSP). The observed richness of each local- ity (Fig. 1) was calculated including mor- phospecies, records from literature and mate- rial from museums. Biogeographic analyses. — Four subfami- lies of Gonyleptidae, for which we have da- distic hypotheses at species level (Gonioso- matinae, Caelopyginae, Progonyleptoidellinae and Sodreaninae), were used for biogeograph- ic analyses. The areas of endemism were cho- sen by overlapping areas of distribution of at least three endemic species. The areas of en- demism basically follow Pinto-da-Rocha (2002) with some modifications: the compo- nent Santa Catarina (SC) was split from Pa- rana (PR); the southern region of Sao Paulo (SSP), located in the Vale do Ribeira was split from the Serra do Mar of Sao Paulo; and the Serra dos Orgaos (Org) was split from the Ser- ra do Espinhago (SEsp). Other abbreviations are: Bahia (BA), Espirito Santo (ES), low- lands of the northern part of the Sao Paulo coast and the southern part of Rio de Janeiro state (LSRJ), Serra da Bocaina (Boc), Serra da Mantiqueira (Mnt), Serra do Espinhago (SEsp), Serra do Mar de Sao Paulo (SMSP), and Serra dos Orgaos (Org). The primary Brooks Parsimony Analysis (BPA) was performed to infer relationships among areas. In this analysis each terminal of the species’ cladograms was replaced by the species’ areas of distribution; terminals and nodes were transformed into a binary matrix (Brooks et al. 2001). The function of primary BPA is to determine whether there is a general pattern among areas (Brooks et al. 2001). Widespread taxa were considered informative and their area considered as monophyletic (Assumption 0). The matrix of the area was constructed with the patterns of distribution of 84 species (Table 1) of two species clado- grams: one for the subfamily Goniosomatinae; and another for the monophyletic group com- posed of the subfamilies (Sodreaninae (Progonyleptoidellinae, Caelopyginae)). The hypotheses of relationship among subfamilies of Gonyleptidae and species of Caelopyginae are in Pinto-da-Rocha (2002). The revisions and hypotheses for the Goniosomatinae (M.B. da Silva and P Gnaspini), Sodreaninae and Progonyleptoidellinae (both R. Pinto-da-Ro- cha) are in preparation for publication. Taxa cladograms (for species names see Table 1): Caelopyginae - (((7, 6)(2(4(3, 1)))) ((19(11 (10(12(13(18)) (14(15, 16, 17))))) (5(23(21 (20, 22))) ((8, 9) (27(24, 25, 26)))))))); Gon- iosomatinae = ((((62, 63) ((50, 54) (64, 68))) ((67(51, 55)) (73(66(52, 65))))) (((70(53(59, 75))) (60(57(49(56(61, 58)))))) (82(84((79(80 (74, 78))) (77(83((71, 72) (81(69, 76)))))))))); Progonyleptoidellinae = ((45, 46) ((48(47(41 (40,42)))) (36(43, 44)(35(37(38(34, 39))))))); Sodreaninae - (30(28(29(33(32, 31))))). The parsimony analysis of the biogeographic ma- trix was conducted with the PAUP 4.0 (Swof- ford 2002), using Branch-and-Bound algo- rithm with the commands holdlOOOO, mult* 1000 and hold/ 1000. RESULTS Richness. — Richness varied from 4-64 species per locality in south and southeastern Brazil (Fig. 1). The areas of low diversity are in cerrado forests (Brazilian savanna) with 4- 7 species (Piracicaba and Pirassununga) and in the Atlantic Semi-deciduous Forest with 8- 12 species (Foz do Iguagu, Porto Cabral, Japi and Atibaia). Localities in the Costal Atlantic Rain Forest are richer with 12-64 species. However, it must be stressed that some areas on the coast, such as Ilha Anchieta and Par- aiso, were undersampled (40-50 h of noctur- nal sampling) and there are no records either from literature or from museums. Thus, these estimates should be taken with care. Localities considered as well-sampled such as Cantarei- ra, Morro Grande, Boraceia, Paranapiacaba, Itatiaia and Serra dos Orgaos, present 27-64 species. Therefore, the fauna of harvestmen of the Coastal Atlantic Rain Forest is richer than the Atlantic Semi-deciduous Forest and the cerrado. Faunistic similarity. — Analyses showed clusters among localities of the same moun- tain range (Fig. 2). From the 363 species in- cluded in the matrix, 213 (58.7%) occurred in just one locality. Among the 150 species re- corded in more than one locality, only 93 were in two. Therefore, the groups possess only a few species in common, generating very low indices of similarities, thus indicating the high level of endemism of harvestmen species. The fauna of the Atlantic Rain Forest of the State of Sao Paulo forms a distinct group from other regions and also from the interior region PINTO-DA^ROCHA ET AL.— BIOGEOGRAPHY OF HARVESTMEN 293 Porto Cabrai P.N. Iguagu Santa Teresa Santa Leopoldina Colatina Piracicaba Pirassununga Piraquara A.EJT. Marumbi P.N.S. Orgios Petropolis E.E.E. Paraiso P.E. Campos do Jordio P. N. itatiaia P.N.S. Bocaina P.E. Ilha Anchieta P.E. ilhabela Atibaia S. Japi P.E.S. Cantareira P.E.T. AltoRibeira P.E. Jureia-ltatins Miracatu R. Morro Grande E.B. Boraeeia R.B.A.S. Paranapiacaba Sorensen's Coefficient Figure 2.-™-Cluster analysis (S0rensen index) showing the similarity among harvestmen faunas (Lan- iatores and Eupnoi) of 26 areas in south and southern Brazil, Abbreviations in Table L * = areas not included in biogeographic analyses; ** = Cerrado (savanna). of Sao Paulo (Cantareira, Japi and Atibaia). This group is divided into a cluster from the southern coast of Sao Paulo (Miracatu, Jureia, Morro Grande and PETAR), and another in the center of the Sao Paulo Coastal Rain For- est (Boraeeia and Paranapiacaba). A high similarity was found between the Sen ra do Japi (about 70%) and Atibaia, and can be explained by their proximity (ca. 30 km). The same occurs with the Paranapiacaba and Bora- ceia localities; besides their proximity, they be- long to the same geomorphological formation. The fauna of the northern coast of the state of Sao Paulo is distinct from other regions of the State of Sao Paulo and even its two localities showed a low similarity (index below 0.30). It is possible to identify a great cluster com- posed of localities of Serra do Mar of Rio de Janeiro, Serra da Maetiqueira and Serra da Bocaina. Within this cluster, the localities are even distinctly different from each other, the Serra do Mar of Rio de Janeiro, including the Serra dos Orgaos, formed a distinct subgroup from other localities, including Serra da Mae- tiqueira (Campos do Jordao and Itatiaia) and Serra da Bocaina. Other groups can be iden- tified: localities of the cerrado (Piracicaba and Pirassununga); Serra do Mar of Parana (Pira- quara and Marumbi); Semi-deciduous Forest (P.N. Iguagu and Porto Cabral); and a group formed by three localities in Espirito Santo. Historical Biogeographic Analysis.— Among the 84 species included in the analy- sis, 66 occur in one area, 13 were recorded in 2-“3 areas, four in 4-8 areas and two have no precisely known locality (Table 1). The stud- ied groups were not recorded in the cerrado and the more interior areas of the Atlantic Semi-deciduous Forest, except for the Met and the SEsp. Eleven areas of endemism were recognized as having at least three endemic species belonging to the subfamilies Caelo- pyginae, Goeiosomatinae, Progonyleptoidel- linae and Sodreaeinae (Fig. 3). Primary BPA 294 THE JOURNAL OF ARACHNOLOGY Figure 3. — Strict consensus area cladogram of harvestmen subfamilies Caelopyginae, Goniosomatinae, Progonyleptoidellinae and Sodreaninae (Gonyleptidae) based on three equally parsimonious trees (L = 248; Cl = 0.66; RI = 0.57). Abbreviations of names in Table 1; * = Paraiba do Sul River; ** = Ribeira do Iguape River. analyses resulted in three equally parsimoni- ous cladograms (L = 248; Cl = 0.66; RI = 0.57) and the strict consensus (L = 254; Cl = 0.64; RI “ 0.54) is shown in Fig. 3. The three cladograms varied in positions of BA and SEsp, which formed a group in two clad- ograms, SEsp being more related to other ar- eas than to BA in the third cladogram. The placement of Org and Mnt also varied, with the Org sister of Boc+LSRJ being in two cladograms and Mnt sister of the Boc + LSRJ in one. The consensus cladogram shows a se- quence of vicariant events from North to South. However, it should be noted that the basal position of Bahia and Serra do Espin- haqo in the cladogram could be due to the few endemic species recorded in these areas be- sides the insufficient information to relate them to the southern areas, such as Serra do Espinhago. The vicariant events that followed separated Espirito Santo from the southern ar- eas and the central areas of Sao Paulo, and Rio de Janeiro from the southernmost areas. The Serra do Mar of the State of Sao Paulo was split from the continuous remaining part just south of Sao Paulo, as this area showed affinities with southern Brazil from which it is separated by the Ribeira do Iguape River. DISCUSSION A Brazil harbors about 900 species of har- vestmen (see Kury 2003 and Hallans catalog PINTO-DA-ROCHA ET AL.— BIOGEOGRAPHY OF HARVESTMEN 295 at http://entowww,tamu.edu/research/collection/ hallan/OpilRpt2Txt). The Coastal Atlantic Rain Forest possesses most of this diversity (about 600 described species), which makes this area the most diverse in the world for this taxonomic group. Among the 16 subfamilies of Gonyleptidae, the predominant group in the Atlantic Rain Forest, nine are exclusively found in this vegetation formation, whereas two occur mainly in this region (Tricomma- tinae and Hernandariinae). These 1 1 subfam- ilies include 223 described species. Other di- versified groups found in this region are Pachylinae, Gonyleptinae and Sclerosomati- dae, among others. The study of similarity patterns (Fig. 2) among the well-sampled localities (Fig. 1) showed an almost total coincidence with the areas of endemism herein proposed. The very low similarity between localities and groups of localities show how isolated these faunas are. These results indicate the high influence of geomorphoiogy and geographical isolation in the pattern of harvestmen species distribu- tion. The clusters show, in general, that local- ities in the same mountain range are more similar to each other than to those in other mountain ranges. There are two main biogeographic studies in South America that consider south and southeastern Brazil as belonging to more than one biogeographic component. Costa (1995) presented an area cladogram for three groups of Cyprinodontiformes, 23 other groups of fishes, and one genus of frog. Amorim & Pires (1996) presented a general area cladogram based on several groups of neotropical dipter- ans (Ditomyiidae, Sciaridae and Scatopsidae) and monkeys (Callitrichidae). Both hypothe- ses considered the Atlantic Rain Forest as having 4-6 components. However, the vicar- iant events postulated by these authors con- sidered different areas of endemism for ter- restrial and freshwater animals. Biogeography of the freshwater fauna seems to be related to paleodrainages that flowed to the interior or the Atlantic Ocean (Lundberg et al. 1998). On the other hand, the terrestrial fauna occurs on both sides of the mountain range. Costa (1995) suggested that the coastal areas form a component that is sister to a biogeographic component (his area “f”) that encompasses a large interior region around the La Plata and Sao Francisco Basins, and includes our area BA. In his study, the components of the Serra da Mantiqueira and the Serra do Espinhago do not possess any taxa. According to Costa (1995), the coastal components share a unique ancestral area in which the first vicariance event split areas “i+h” from “g” (similar to our areas SMSP, LSRJ, Boc, Org, and ES), followed by a second divergence between his areas “i” (with no taxon in our study), and “h” (our SC, PR e SSP). Amorim & Pires (1996) related SE Amazonia with other Bra- zilian regions. According to them, the areas comprising the center and northeast sections of Brazil (areas MGBA and NEBr in their fig. 26) form a component which is sister to south- east Bahia (our BA). This whole component is sister to a clade comprised of north Rio de Janeiro (our ES), Sao Paulo-Rio de Janeiro (our Org, Mnt, Boc, LSRJ, SSP, PR and SC), and southern Brazil and the northeast of Ar- gentina (areas for which there are no opilion- ids related to this study). The main vicariant events are related to mountain uplift and the appearance of valleys. The origin of the Serra do Mar and the Serra da Mantiqueira was during the Paleocene (Pe- tri & Eulfaro 1988), or early in the Upper Cre- taceous, as a result of tectonic activity (Al- meida & Carneiro 1998). Although the great orographic ascension occurred between the Pliocene and the Pleistocene, we should stress the origin as being recent. Valleys seem to represent important geographical barriers, such as the valley of the Paraiba do Sul River, whose origin was during the Oligocene-Mio- cene (Petri & Fulfaro 1988), and isolated the Serra da Mantiqueira in the west from the Ser- ra do Mar, Serra da Bocaina and the Serra dos Orgaos in the east. In addition, the same val- ley isolated the northern areas (Espirito Santo, Serra do Espinha^o and Bahia) from the re- maining southern ones (Fig. 3). It is interesting to note that the Atlantic side of most coastal mountains receives a large amount of rain (up to 4,000 mm a year). On the other hand, the interior side is in a rain shadow, and a valley such as Paraiba do Sul, receives one-third as much rain as the adjacent mountain range (Behling & Lichte 1997). An- other remarkable fact was the generation of new environments during glaciations in the Pleistocene, such as grasslands and short gal- lery forests, in areas previously covered by rain forests, where currently there are semi- 296 THE JOURNAL OF ARACHNOLOGY Table 1. — Distribution of species of harvestmen subfamilies Caelopyginae, Goniosomatinae, Progony- leptoidellinae and Sodreaninae (Gonyleptidae) in eleven areas of endemism in south and southeastern Brazil. SC = Santa Catarina; PR = Parana; SSP = Sul de Sao Paulo; SMSP = Serra do Mar de Sao Paulo; Mnt = Serra da Mantiqueira; Boc = Serra da Bocaina; LSRJ = north coast of Sao Paulo and south of Rio de Janeiro; Org = Serra dos Orgaos; ES = Espmto Santo; SEsp = Serra do Espinhago; BA = southern coastal Bahia. Species/area SC PR SSP SMSP Mnt Boc LSRJ Org ES SEsp BA Caelopyginae 1 . Ampheres fuscopunctatus X X 2. A. leucopheus X X X X X X X X 3. A. luteus X 4. A. tocantinus 7 7 ? 7 ? 7 7 7 7 ? 7 5, Arthrodes xanthopygus X 6. Caelopygus elegans X 1. C. melanocephalus 8. Garatiba bisignata 9. G. bocaina X X X 10. Metarthordes albotaeniatus X 11. M. bimaculatus 12. M. hamatus X X X 13. M. laetabundus 14. M. leucopygus 15. M. longipes X X X X X 16. M. nigrigranulatus 17. M. pulcherrimus 18. M. xango 19. Metampheres albimargina- X X X X X tus X 20. Pristocnemis albimaculatus X 21. F. farinosus X X X X 22. P. perlatus X X 23. P. pustulatus 24. Thereza albiornata X X X X X X X X 25. T. amabilis X 26, T. poranga 27. T. speciosa X X X Sodreaninae 28. Gertia hatschbachi 29. Sodreana sodreana X X X X 30. Stygnobates barbiellinii 31. Zortalia bicalcarata X X 32. Z. inscripta 33. Z. leprevosti X X X X Progonyleptoidellinae 34. Cadeadoius niger X 35. Gonyleptoides acanthoscelis 36. G. curvifemur 37. G. marumbiensis X X X 38. Heliella singularis X 39. Iguapeia melanocephala X X 40. Iporangaia pustulosa 41. Leptocnema sulphurea 42. Mitopernoides variabilis 43. Moreiranula mamillata 44. M. moreirae X X X X X PINTO=DA-ROCHA ET AL.— BIOGEOGRAPHY OF HARVESTMEN 297 Table 1. — Continued. Species/area SC PR SSP SMSP Mnt Boc LSRJ Org ES SEsp BA 45. Progonyleptoidellus fuscop- ictus 46. P. striatus 47. Gen, sp.n. 1 48. Gen. sp.n. 2 Goniosomatinae 49. Acutisoma banhadoae X X X X 50. A. discolor SLA. hamatum 52. A. indistinctum 53. A. inerme X X X X 54. A. inscriptum 55. A. longipes 56. A. moUe X X X 57, A. proximum X X X 58. A. thalassinum X 59. A. sp.n. 1 X 60. A. sp.n. 2 61. A. sp.n. 3 62. A. sp.n. 4 63. A. sp.n. 5 X X X X 64. A. sp.n. 6 65. A. sp.n. 7 X X 66. A. sp.n. 8 67. A. sp.n. 9 X X 68. Goniosoma albiscriptum 69. G. calcar X X 70. G. Catarina 11. G. dentipes X X 72. G. ensifer 13. G. mode stum X X 74. G. roridum X 75. G. spelaeum 76. G. unicolor X X X X X 11. G. vatrax 78. G. venustum X X 19. G. varium X X 80. G. sp.n. 1 81. G. sp.n. 2 82. Gen n spn 83. Lyogoniosoma macracan- X X X thum X 84. Xulapona cara X Species/area SC PR SSP SMSP Mnt Boc LSRJ Org ES SEsp BA Total species 6 13 10 15 11 10 16 23 7 4 3 Endemic species 5 7 3 7 6 4 9 14 5 3 3 deciduous forests as in the interior of the State plant diverse and , more open eeviroemeets of Sao Paulo (Behling & Lichte 1997) or low- could have decreased the diversity of opilion- lands as is the case in the State of Santa Ca- ids in those sites. The tree floras of semi-de- tariea (Behlieg & Negrelle 2001; BeMieg 2002). The replacemeet of rain forest by less ciduous forests are less diversified than coast^ al forests, so they have been considered a 298 THE JOURNAL OF ARACHNOLOGY subgroup of the former (Oliveira-Filho & Fontes 2000). However, this difference is not as remarkable as it is in harvestmen. Never- theless, we should stress that only two areas of the semi-deciduous forests were well sam- pled (Serra do Japi and Atibaia). This char- acteristic could lead to a misunderstanding of the relationships between coastal and interior areas. The high diversity of opilionids in the Coastal Atlantic Rain Forest, an area of higher diversity than any other country in the world, can be explained by the high number of geo- graphical barriers on the Brazilian coast that isolated populations creating new species, and also by many events of forest fragmentation, hence leading to population divergence, due to climatic changes during the Pliocene-Pleis- tocene. The unique opilionid faunas represented in each of the 11 areas of endemism call atten- tion to the necessity of preserving those en- vironments, The Atlantic Rain Forest is a hot- spot, and the decimation of the Brazilian Atlantic Forest is one of the most alarming conservation problems in the world (Terborgh 1992). This biome possesses a great number of protected areas along the coast in the south- southeastern part of Brazil. In fact, most col- lecting was done in reserves. However, the op- ilionid faunas of three areas of endemism (ES, SEsp and BA) are poorly or not represented in terms of governmental reserves (see Con- servation International do Brazil 2000 or the online atlas at http://www.sosmatatlantica. org.br/?secao= atlas), and their remaining hab- itats are suffering high anthropic pressure (Morellato & Haddad 2000), and deserve bet- ter attention in future planning of new pro- tected areas in order to maintain the diversity of the group. ACKOWLEDGMENTS We are grateful to Adriano Kury and Paulo Inacio who first showed RPR the importance of opilonids in biogeographic studies, to An- tonio D. Brescovit and the team of the BIO- TA-Arachnida project; and to Fernando Marques for critical review of the manuscript, Raimon Clark and Marcos Hara helped with language. Grants from FAPESP # 99/05446- 8 (RPR) and 03/02673-0 (MBS) and CNPq 133994/2003-1 (CB). LITERATURE CITED Almeida, EF. & C.D.R. Carneiro, 1998. Origem e evolugao da Serra do Mar. Revista Brasileira de Geociencias 28(2): 135-150. Amorim, D.S. & M.R.S. Pires. 1996. Neotropical biogeography and a method for maximum bio- diversity estimation. P. 326. In C.E.M. Bicudo & N. A. Menenzes (eds.) Biodiversity in Brazil. A first approach. CNPq. Behling, H. 2002. South and southeast Brazilian grasslands during Late Quaternary times: a syn- thesis. Palaeogeography, Palaeoclimatology, Pa- laeoecology 177:19-27 Behling, H. & M. Lichte. 1997. Evidence of dry and cold climatic conditions at glacial times in tropical southeastern Brazil. Quaternary Re- search 48:348-358 Behling, H. & R.B. Negrelle. 2001. Tropical rain forest and climate dynamics of the Atlantic low- land, southern Brazil, during the late Quaternary. Quaternary Research 56:383-389. Briggs, TS. & D. Ubick. 1989. The harvestmen family Phalangodidae. 2. The new genus Micro- cina (Opiliones: Laniatores). Journal of Arach- nology 17:207-220. Brooks, D.R., M.G.P. van Veller & D.A. MacLen- nan. 2001. How to do BPA, really. Journal of Biogeography 28:345-358. Conservation International do Brasil. 2000. Avalia- gao e agoes prioritarias para a conserva^ao da biodiversidade da Mata Atlantica e campos su- linos. Ministerio do Meio Ambiente, Brasilia, 40 p. (available at http://www.conservation.org.br/ publicacoes/files/Sumario.pdf) Costa, W.J.E.M. 1995. Pearl killifishes, the Cyno- lebiatinae: systematics and biogeography of the neotropical annual fish subfamily (Cyprinodon- tiformes, Rivulidae). Neptune City: TEH Pub- lications, 128 p. Kovach Computing Services. 1999. MVSP 3.1. Multi-variate statistical package. Computer pro- gram. Wales. Kury, A.B. 2003. Annotated catalogue of the Lan- iatores of the New World (Arachnida, Opiliones). Revista Iberica de Aracnologfa, Zaragoza, vol. especial monografico 1:1-337. Lundberg, J.G., L.G. Marshall, J. Guerrero, B. Hor- ton, M.C.S. Malabarba & E Wesselingh. 1998. The stage for Neotropical fish diversification: A history of tropical South American rivers. In: L.R. Malabarba, R.E. Reis, R.R Vari, Z.M.S. Lu- cena & C.A.S. Lucena (eds.). Phylogeny and Classification of Neotropical Fishes. EDI- PUCRS. 603 p. Morellato, L.PC. & C.F.B. Haddad. 2000. Introduc- tion: the Brazilian Atlantic Forest. Biotropica 32(4b):786-792. Morrone, J., S.R. Junent & J.V. Crisci. 1994. South PINTO^DA~ROCHA ET AL.— BIOGEOGRAPHY OF HARVESTMEN 299 American beetles. National Geogriphic Research Explores 10(1): 104-1 15. Oliveira-Filho, A. & M.A.L. Fontes. 2000. Patterns of floristic differentiation among Atlantic forests in southeastern Brazil, and the influence of cli- mate. Biotropica 32:903-913. Perez-Losada, M., G. Bond-Buckup, C. Jara ,& K.A. Crandall. 2004. Molecular Systematics and Biogeography of southern South American fresh- water “crabs” Aegla (Decapoda: Anomura: Ae- glidae) using multiple heuristic tree search ap- proaches. Systematic Biology: 53(5):767-780. Petri, S. & V.J. Fulfaro. 1988. Geologia do Brasil (Faeerozoico). EDUSP. 631 p. Pinto-da-Rocha, R. 2002. Systematic review and cladistic analysis of the Brazilian subfamily Cae- lopyginae. Arquivos de Zoologia 36(4):357-364. Swofford, D. L. 2002. PAUP*. Phylogenetic Anal- ysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Mas- sachusetts. Terborgh, J. 1992. Maintenance of the diversity in tropical forests. Biotropica 24:283-292. Tourinho-Davis, A.L.M. 2004. A new genus of Ga- grellinae from Brazil, with a comparative study of some of the southernmost tropical and sub- tropical South American species (Eupnoi, Scle- rosomatidae, Gagrellinae). Revista Iberica de Ar- acnologia 9:157-177. Tourinho-Davis, A.L.M. & A.B. Kury. 2003. A re- view of Jussara, with descriptions of six new species (Arachnida, Opiliones, Sclerosomatidae) from Brazil. Tropical Zoology 16(2):209-275. Ubick, D. & T.S. Briggs. 1989. The harvestmen family Phalangodidae. 1. The new genus Cali- cina, with notes on Sitalcina (Opiliones: Lania- tores). Proceedings of the California Academy of Sciences 46(4):95-136. Manuscript received 10 December 2004, revised 5 August 2005. 2005. The Journal of Arachnology 33:300-305 A SURVEY OF SPIDERS (ARANEAE) WITH HOLARCTIC DISTRIBUTION Yuri M. Marusik: Institute for Biological Problems of the North, RAS, Portovaya Str. 18, Magadan 685000, Russia. E-mail: yurmar@maiLru Seppo Koponen: Zoological Museum, University of Turku, FI-20014 Turku, Finland. ABSTRACT. Of the 13,800 species distributed in the Holarctic Region only 395 are known both from Eurasia and North America. Of these only 105 species are distributed throughout the whole Holarctic (circum-Holarctic species). In addition, 28 species have an almost complete Holarctic distribution, occur- ring from Europe to northwestern North America (subcircum-Holarctic species). Species with a circum- Holarctic distribution were found in 13 families. The highest numbers of circum-Holarctic species were in the families Linyphiidae (37), Theridiidae (14), Araneidae (13) and Gnaphosidae (11). The percentage of the circum-Holarctic species among the Holarctic spiders is highest in Philodromidae (2.4%), Araneidae (2.2%), Theridiidae (2.0%) and Tetragnathidae (1.9%). These families encompass mainly herb-bush-tree dwellers. Somewhat unexpectedly it was found that most circum-Holarctic species occupy the boreo- nemoral zone (41%), or may even have a polyzonal range (23%). Twenty-nine species (28%) of the circum-Holarctic spiders have a northern distribution; most of them occurring both in arctic and boreal zones. Keywords: Holarctic region, circum-Holarctic species, distribution types, zonal distribution, spiders The Holarctic region, an area covering the Northern Hemisphere approximately north of 25° N, is the largest zoogeographical realm of the Earth. Around 13,800 species of spiders are listed in Platnick’s (2004) catalogue as in- habitants of this realm. Without a doubt, the Holarctic is the best studied region in all groups of living organisms. Most biogeographers divide the Holarctic region into two subunits, Palaearctic and Ne- arctic, lying in the Old and New World re- spectively. Among the species of spiders known in the Holarctic, only 395 species (or around 3%) are known from both Palaearctic and Nearctic regions. Most of them are listed in PlatnickN (2004) and other catalogues as Holarctic or Cosmopolitan species. Considering different meanings of the word Holarctic, we wish to stress that in this paper under the term Holarctic species (or distribu- tion, range) or circum-Holarctic species (or distribution, range) we mean species occur- ring (distributed) throughout the whole or at least most of Eurasia and North America. Many authors consider distribution of species as Holarctic if they are known from two con- tinents, although a species may be known only from one locality in one continent (e.g.. Plat- nick 2004). Holarctic species possibly intro- duced by man, long ago or more recently, have been treated here like the others, “nat- urally Holarctic” species. The longitudinal width of the range of the circum-Holarctic species restricted to boreal or hypoarctic zones is slightly wider than that of species occurring in the nemoral zone (Figs. 1-2), although the real length (in kilo- meters) is longer in the nemoral zone. The ne- moral zone starts in the Palaearctic at the Ca- nary Islands (15°W) and continues to Kamchatka (160°E) (total length of the zone is about 180°); in the Nearctic this zone stretches from about 150°W (Alaska) to about 60°W (Nova Scotia) (length = 90°). Altogeth- er the nemoral zone covers about 270°. The boreal and hypoarctic zones start at about 10°E (Fennoscandia) and continue almost without break to about 40° W (Greenland), and altogether comprise 310°. Species having po- lyzonal ranges or those that are synantropic have the widest ranges and can occur almost throughout the whole Holarctic. The goal of this paper is to list all species of spiders which have a wide Holarctic range (either circum- or subcircum-Holarctic). Such a list can be a useful source for many fields 300 MARUSIK & KOPONEN— HOLARCTIC SPIDERS 301 Figure -1. Distributional zones in the Hoiarctic region. 2. The width of nemoral and boreal- hypoarctic zones in the Hoiarctic region. of arachnology, like population genetics (var- iability across the wide range), ecology (com- parative ecology and ethology of widespread species), taxonomy (morphological variation across the range), and physiology (study of cold resistance or thermal preferences in dif- ferent parts of the wide range). Often com- parative study, either ecological or morpho- logical, on distant populations of widespread species reveals important differences which can lead to separation of new taxa. METHODS The major source of potential Hoiarctic species is the catalogue of Platnick (2004), from which the species mentioned as Hoiarc- tic were chosen. These species were studied using personal knowledge and recent species lists (e.g., Doedale et aL 1997; Marusik et ak 2000; Buckle et al. 2001). Many other publi- cations have also been used, the most impor- tant are, in alphabetic order: Doedale & Red- ner (1990), Levi & Randolph (1975), Logunov (1996), Logunov & Marusik (2001), Marusik (1994), Marusik et al. (1992, 2002a, 2002b), Mikhailov (1997), Rybalov et al. (2002), Saaristo & Eskov (1996), Song et al. (1999) and Yoshida (2003). The following main distribution types (ab- breviations in brackets) have been distin- guished (cf. Appendix 1): Arctic = tundra zone (ar); Boreal = taiga or coniferous forest belt (bo); Hypoarctic = arctic + northern tai- ga + mountain tundra in boreal zone (hy); Ne- moral = zone south of boreal: mixed or de- ciduous forest, steppe, desert (ne); Polyzonal = wide range within above types (po); Mon- tane = mountains in nemoral zone (mo); Cos- mopolitan (cos), see also Fig. 1. RESULTS AND DISCUSSION Of the more than 13,800 species recorded in the Hoiarctic Realm only 395 are known in both Eurasia and North America, and only 105 of them are distributed throughout the en- tire Hoiarctic, i.e. they are circum-Holarctic. In addition, 28 species have an almost Hoi- arctic distribution, occurring from Europe to northwestern North America, i.e. they are sub- circum-Holarctic. This means that less than 1% of all species in the Hoiarctic region are circum- or subcircum-Holarctic (Table 1). Thus, the number of truly Hoiarctic species of spiders is much lower than usually estimated (cf. Platnick 2004). Of the 65 species listed as Hoiarctic in Pro- szyhski & Star^ga (1971) at least 16 are not really Hoiarctic. On the other hand, the num- ber of species within the Hoiarctic has sub- sequently increased considerably due to active research in Siberia and the Nearctic (cf, Ma- rasik et al. 2000). Marusik listed most of the present Hoiarctic species ten years ago (Ma- 302 THE JOURNAL OF ARACHNOLOGY Table 1. — Number of species found both in Nearctic and Palaearctic regions (1), number of species with circum- (2) and subcircum- (3) Holarctic distribution, percentage of circum-Holarctics of all species found in the Holarctic Realm (4), number of species found within the Holarctic Realm (5) and worldwide (6). Basic data from Platnick (2004). “Others” include Dysderidae, Hahniidae, Liocranidae, Miturgidae, Nesticidae, Oecobiidae, Oonopidae, Scytodidae, Sicariidae, Sparassidae, Theridiosomatidae, Uloboridae, and Zorapsidae. Family 1 2 3 4 5 6 Agelenidae 7 1 — 0.3 295 490 Araneidae 23 13 — 2.2 599 2824 Clubionidae 10 3 — 1.2 259 529 Dictynidae 9 3 3 0.7 446 555 Gnaphosidae 27 11 — 0.9 1162 1955 Linyphiidae 160 38 13 1.3 3003 4247 Lycosidae 25 5 2 0.5 1041 2262 Philodromidae 15 7 2 2.4 286 512 Pholcidae 7 1 — 0.5 187 836 Salticidae 21 3 — 0.2 1382 4975 Tetragnathidae 7 3 — 1.9 160 1026 Theridiidae 45 14 5 2.0 690 2209 Thomisidae 15 3 1 0.5 609 2028 Amaurobiidae 3 — 1 — 444 626 Titanoecidae 1 — 1 — 34 46 Others 21 — — — 3189 13302 Total 395 105 28 0.8 13786 38432 rusik 1994); however, nine species have now been added and eleven omitted. Species occurring both in the New and Old World parts of the Holarctic belong to 28 spi- der families (Table 1); however, species with circum-Holarctic distribution are known in 13 families only. Two additional families each have one subcircum-Holarctic species. The following families have most Holarctic species: Linyphiidae (38), Theridiidae (14), Araneidae (13), Gnaphosidae (11), Philod- romidae (7) and Lycosidae (5). The percent- age of circum-Holarctic species among all the species found in the Holarctic Region, is high- est in Philodromidae (2.4%), Araneidae (2.2%), Theridiidae (2.0%) and Tetragnathidae (1.9%) (Table 1). These families encompass mainly herb-bush-tree dwellers. Among gen- era, the most rich in circum-Holarctic species are Micaria (5), Thanatus (5) and Theridion (6). The last-mentioned genus seems to be paraphyletic, and its six species with circum- Holarctic range belong to three different groups. We expected that, like in many other groups of living organisms (see Banks 1981), most of the species with a circum-Holarctic distribu- tion would be restricted to the northern (boreal and/or arctic) zones. The main reasons for this expectation were the smaller area and post- glacial history of the boreal, and especially the arctic zones, compared to the nemoral zone. However, it was found that most of circum- Holarctic species occur in the boreo-nemoral zone (41% or 43 species), and many even have a polyzonal range (23% or 24 species). Among circum-Holarctic species, only 28% (or 29 species) have a northern distribution (arctic, hypoarctic and/or boreal range). Most of them occur both in tundra and boreal zones, and three species are known from the boreal zone only. Among the 28 subcircum-Holarctic species, as many as 16 (57%) have this kind of northern (arctic, hypoarctic and/or boreal) distribution pattern. The proportion of cosmopolitans among circum-Holarctic species was highest in Ther- idiidae (4 species). Many species with a cos- mopolitan range are absent from Siberia, like Ostearius melanopygius (O.R-Cambridge 1879), Tenuiphantes tenuis (Blackwall 1852), Diplocephalus cristatus (Blackwall 1833), and two Oecobius species. ACKNOWLEDGMENTS This work was partly supported by the Academy of Finland (grants 207667 and MARUSIK & KOPONEN— HOLARCTIC SPIDERS 303 211596) and by the Russian Foundation for Basic Research (grant 04-04-48727). LITERATURE CITED Buckle, DJ., D. Carroll, R.L. Crawford & V.D. Roth. 2001. Linyphiidae and Pimoidae of Amer- ica north of Mexico: checklist, synonymy, and literature. Fabreries, Suppl. 10:89-191. Danks, H.V. 1981. Arctic Arthropods. Entomolog- ical Society of Canada. Ottawa. 608 pp. Dondale, C.D. & J.H. Redner. 1990. The Wolf Spi- ders, Nurseryweb Spiders, and Lynx Spiders of Canada and Alaska (Araneae: Lycosidae, Pisaur- idae, and Oxyopidae). The Insects and Arachnids of Canada 17. 383 pp. Dondale, C.D., J.H. Redner & Yu.M. Marusik. 1997. Spiders (Araneae) of the Yukon. Pp.73- 113. In Insects of the Yukon (H.V. Danks & J.A. Downes, eds.). Biological Survey of Canada Monograph series 2. Ottawa. Levi, H.W. & D.E. Randolph. 1975. A key and check list of American spiders of the family Theridiidae north of Mexico (Araneae). Journal of Arachnology 3:1-51. Logunov, D.V. 1996. A critical review of the spider gQn&rsL Apollophanes O. P.-Cambridge, 1898 and Thanatus C. L. Koch, 1837 in North Asia (Ara- neae, Philodromidae). Revue Arachnologique 11: 133-202. Logunov, D.V. & Yu.M. Marusik. 2001. Catalogue of the jumping spiders of northern Asia (Arach- nida, Araneae, Salticidae). KMK Scientific Press Ltd. Moscow. 299 pp. Marusik, YM. 1994. A check-list of spiders with trans-Palaearctic distribution. Bollettino dell’Accademia Gioenia di Scienze Naturali 26(345):273-279. Marusik, Y.M., K.Y. Eskov & J.P. Kim. 1992. A check list of spiders (Aranei) of Northeast Asia. Korean Arachnology 8:129-158. Marusik, YM., D.V. Logunov & S. Koponen. 2000. Spiders of Tuva, South Siberia. IBPN FEB Rus- sian Academy of Sciences Magadan. 252 pp. Marusik, Y.M., L.B. Rybalov, S. Koponen & A.V. Tanasevitch. 2002a. Spiders (Aranei) of Middle Siberia, an updated check-list with a special ref- erence to the Mirnoye Field Station. Arthropoda Selecta 10(4):323-350. Marusik, YM., S. Koponen, N.N. Vinokurov & S.N. Nogovitsyna. 2002b. Spiders (Aranei) from northernmost forest-tundra of northeastern Yak- utia (70°35'N, 134°34'E) with description of three new species. Arthropoda Selecta 10(4): 351-370. Mikhailov, K.G. 1997. Catalogue of the spiders of the territories of the former Soviet Union (Arach- nida, Aranei). Zoological Museum, Moscow State University, Moscow. 416 pp. Platnick, N.I. 2004. The World Spider Catalog, Ver- sion 4.5, online at 2004 American Museum of Natural History, http://research.amnh.org/ento- mology/spiders/catalog/index.html Proszyhski, J. & W. Star^ga. 1971. Paj?iki- Aranei. Katalog Fauny Polski 33:1-382. Rybalov, L.B., YM. Marusik, A.V Tanasevitch, LG. Vorobyeva, S. Koponen. 2002. Spiders (Ar- anei) of the Yenisei River middle flow and en- virons of “Mirnoye” field station. Study of bi- ological diversity along Yenisei Ecological transect. Animal World. Moscow, IEEP:70-95. [In Russian] Saaristo, M.I. & K.Y Eskov. 1996. Taxonomy and zoogeography of the hypoarctic erigonine spider genus Semljicola (Araneida, Linyphiidae). Acta Zoologica Fennica 201:47-69. Song, D.X., M.S. Zhu & J. Chen. 1999. The Spiders of China. Hebei Science and Technology Pub- lishing House. Shijiazhuang. 640 pp. Yoshida, H. 2003. The spider family Theridiidae (Arachnida: Araneae) from Japan. Arachnologi- cal Society of Japan. Osaka. 224 pp. Manuscript received 22 December 2004, revised 1 Sepetember 2005. Appendix 1. — Species with a circum- and subcircum-Holarctic range. The distribution types used are: arctic (ar), boreal (bo), hypoarctic (hy), nemoral (ne), polyzonal (po), montane (mo) and cosmopolitan (cos), see also Methods. Species with a subcircum-Holarctic dis- tribution have been marked with an asterisk (*). Agelenidae 1/0 (circum-Holarctic/subcircum- Holarctic) Tegenaria domestica (Clerck 1757)-Cos. Amaurobiidae 0/1 * Arctobius agelenoides (Emerton 1919) — ar-bo, unknown in eastern half of the Nearctic. Araneidae 13/0 Aculepeira carbonarioides (Keyserling 1892) — hy; A. packardi (Thorell 1875) — hy-ne; Araneus dia- dematus Clerck 1757 — bo-ne; A. marmoreus Clerck 1757 — bo-ne; A. nordmanni (Thorell 1870) — bo-ne; A. saevus (L. Koch 1872)-bo-ne; Araniella displicata (Hentz 1847) — bo-ne; A. proxima (Kulczyhski 1885) — bo; Cercidia prom- inens (Westring 1851) — po; Cyclosa conica (Pal- las 1771) — bo-ne, although absent in NE Siberia (east of Lena river); Hypsosinga pygmaea (Sun- devall 1831) — po; Larinioides cornutus (Clerck 1757) — po; L. patagiatus (Clerck 1757) — po. Clubionidae 3/0 Clubiona kulczynskii Lessert 1905 — bo-ne; C. pal- 304 THE JOURNAL OF ARACHNOLOGY Udula (Clerck 1757) — bo-ne; C. trivialis C.L. Koch 1874 — bo-ne. Dictynidae 3/3 * Arctella lapponica (Holm 1945) — ar-bo, in Ne- arctic known only from the North-West; * Dic- tyna alaskae Chamberlin & Ivie 1947 — bo, in Nearctic is known from Alaska only; D. arun- dinacea (Linnaeus 1758) — po; D. major Menge 1869 — po; Emblyna annulipes (Blackwall 1846) — bo-ne; * Hackmania prominuia (Tullgren 1948)^ — ar-bo, in Nearctic known only from the North-West. Gnaphosidae 11/0 Gnaphosa microps Holm 1939 — hy-bo; G. musco- rum (L. Koch 1866) — po, absent in NE Siberia and Far East; G. orites Chamberlin 1922 — hy; Haplodrassus signifer (C.L. Koch 1839) — po; Micaria aenea Thorell 1871 — bo-ne; M. alpina L. Koch 1872 — hy-bo; M. pulicaria (Sundevall 1831) — bo-ne; M. rossica Thorell 1875-po; M. tripunctata Holm 1978 — hy-bo; Trachyzelotes jaxartensis (Kroneberg 1875) — steppe-semides- ert; Zelotes puritanus Chamberlin 1922 — dis- junctive polyzonal range, restricted to warm and xeric habitats from tundra zone to steppes and mountains. Linyphiidae 38/13 Agyneta olivacea (Emerton 1882) — hy-ne; * Allo- mengea scopigera (Grube 1861) — bo-ne, in Nearctic restricted to the western half; Aphileta misera (O.P.-Cambridge 1882) — bo-ne; Bathy- phantes gracilis (Blackwall 1841) — bo-ne; Ca- rorita limnaea (Crosby & Bishop 1927) — bo; Centromerus sylvaticus (Blackwall 1841) — bo- ne; Cnephalocotes obscurus (Blackwall 1834) — bo-ne; Collinsia holmgreni (Thorell 1872) — hy- bo; Diplocentria bidentata (Emerton 1882) — bo-ne; Diplocentria rectangulata (Emerton 1915) — bo; Dismodicus bifrons (Blackwall 1841) — bo-ne; Erigone arctica (White 1852) — ar-bo-mo, represented by series of subspecies, none of which has Holarctic range; E. atra Blackwall 1833 — po; E. psychrophila Thorell 1872 — hy; E. tirolensis L. Koch 1872— hy (ar- mo); Estrandia grandeva (Keyserling 1886) — hy-ne; Helophora insignis (Blackwall 1841) — bo-ne; Hilaira herniosa (Thorell 1875) — hy-bo-mo; * H. nubigena Hull 1911— bo, in Nearctic known from Alaska only; * Horcotes strandi (Sytchevskaya 1935) — hy-bo, in Nearctic known from Yukon Territory only; * Hybauch- enidium ferrumequinum (Grube 1861)— hy, in Nearctic known from Yukon Territory only; * Hypselistes jacksoni (O.P.-Cambridge 1902) — hy-ne, in Nearctic known only from Alaska to Utah; Improphantes complicatus (Emerton 1882) — bo-mo; * Islandiana falsifica (Keyserling 1886) — hy, in Nearctic known from NW part (Alaska to Northwestern Territories, south to British Columbia); Kaestneria pullata (O.P.- Cambridge 1863) — bo-ne; ''Lepthyphantes” le- prosus (Ohlert 1867) — po, north of 55°N exclu- sively synantropic; Macrargus multesimus (O.P.-Cambridge 1875) — hy-ne; * Maso sundev- alli (Westring 1851) — bo-ne, in Nearctic known in Alaska only; * Mecynargus monticola (Holm 1943)— hy-bo, in Nearctic known only from Western Canada only; M. paetulus (O.P.-Cam- bridge 1875) — bo-ne; M. sphagnicola (Holm 1939) — hy-bo, in Nearctic it occurs in Yukon and NW Territories; Megalepthyphantes nebulosus (Sundevall 1830) — po, north of 55°N exclusively synantropic; * Metopobactrus prominulus (O. R- Cambridge 1872) — bo-ne, in Nearctic known east of Saskatchewan; Microlinyphia impigra (O.P.-Cambridge 1871) — bo-ne; M. pusilla (Sun- devall 1830) — po; Microneta viaria (Blackwall 1841) — po; Neriene clathrata (Sundevall 1830) — bo-ne; N. radiata (Walckenaer 1841) — bo-ne; Oreonetides vaginatus (Thorell 1872) — hy-bo-mo; Pelecopsis mengei (Simon 1884) — bo-ne; Pocadicnemis pumila (Blackwall 1841) — bo-ne; * Poeciloneta variegata (Blackwall 1841) — hy-bo-mo, in Nearctic restricted to the West; * Semljicola lapponicus (Holm 1939) — hy, in Nearctic known from Alaska only; Sisicus apertus (Holm 1939) — bo-ne; Thyreostenius par- asiticus (Westring 1851) — bo-ne; * Tibioplus diversus (C.L. Koch 1879) — bo, in Nearctic known from Alaska and Yukon Territory; Tiso aestivus (L. Koch 1872) — hy-bo, in Nearctic known from Yukon Territory and Greenland; * Walckenaeria capita (Westring 1861) — bo-ne, in Nearctic known from Ontario only; W. cuspidata Blackwall 1833 — bo-ne; W. karpinskii (O.P.- Cambridge 1837) — ar-bo-mo; W. lepida (Kul- czyhski 1885) — bo-ne. Lycosidae 5/2 Alopecosa aculeata (Clerck 1757) — po; Pardosa hyperborea (Thorell 1872) — hy-bo, absent be- tween Lena River and Alaska; * P. lapponica (Thorell 1872) — hy-bo-mo, in Nearctic unknown west of the Hudson Bay; * P. palustris (Linnaeus 1758) — bo-ne, in Nearctic known from Alaska, Yukon Territory and northern British Columbia; Pirata piraticus (Clerck 1757) — po; ''Tricca’' al- pigena (Dolleschall 1852) — hy-bo-mo; Trochosa terricola Thorell 1856 — bo-ne, in Siberia found only in areas free of permafrost, and rather rare in South Siberia. Philodromidae 7/2 Philodromus cespitum (Walckenaer 1802) — po; P. rufus Walckenaer 1826 — bo-ne; Thanatus arcti- MARUSIK & KOPONEN— HOLARCTIC SPIDERS 305 cus Thorell 1872 — ar-bo-mo (=?hy), in Siberia it has polyzonal range, in Scandinavia and Ne- arctic it is restricted to northern taiga and tundra; * T. coloradensis Keyserling 1880 — bo-mo, this species has disjunctions between Alps and Sibe- ria (Marusik et aL 2000); T. formicinus (Clerck 1757) — bo-ne; T. striatus C.L. Koch 1845 — po; * r. vulgaris Simon 1870 — ne (?), in Siberia it has disjunction between Xinjiang and Far East; Tibellus maritimus (Menge 1875) — bo-ne; T. ob- longus (Walckenaer 1802) — bo-ne. Pholcidae 1/0 Pholcus phalangioides (Fuesslin 1775)—- Cos. In northern Palaearctic it is exclusively synantropic species, and most probably absent in South Si- beria. Salticidae 3/0 Chalcoscirtus alpicola (L.Koch 1876) — disjunctive hy-bo-mo range, in Eurasia known from Alps and Siberia east of Ural; Salticus scenicus (Clerck 1757)— bo-ne; Sitticus ranieri Peckham & Peckham 1909 — hy-bo. Tetragnathidae 3/0 Pachygnatha clercki Sundevall 1823 — po; Tetrag- natha dearmata Thorell 1873 — -bo-ne; T. extensa (Linnaeus 1758) — po. Theridiidae 14/5 Achaearanea tepidariorum (C.L. Koch 1841) — Cos, in northern Holarctic it is a synantropic spe- cies; Crustulina sticta (O.P.-Cambridge 1861) — bo-ne; Enoplognatha cartels (Fickert 1876)— bo-ne; Euryopis saukea Levi 1951 — po (?), until 1972 it was known from the Nearctic and Poland only. Later it was found in many places in Ne- arctic, Europe and Asia; * Neottiura bimaculata (Linnaeus 1757) — bo-ne, in Nearctic known from British Columbia and Washington State; * Robertus lividus (Blackwall 1836) — bo-ne, in Nearctic known from Alaska only; * R. lyrifer Holm 1939 — hy-mo, in Nearctic know from Alaska only; Rugathodes aurantius (Emerton 1915) — bo; Steatoda albomaculata (De Geer 1778) — po; * S. bipunctata (Linnaeus 1758) — bo-ne, in Nearctic known from Ontario to New- foundland; S. grossa (C.L. Koch 1838) — Cos, in northern Eurasia it is an exclusively synantropic species; S. triangulosa (Walckenaer 1802) — Cos; * Theridion impressum L. Koch 1881 — po, in Nearctic known from Alaska to western North- west Territories and southward to northern Al- berta; T. montanum Emerton 1882 — bo-ne; T. ohlerti (Thorell 1870) — hy-bo-mo; T. petraeum L. Koch 1872 — bo-ne; T. pictum (Walckenaer 1802) — bo-ne; T. varians (Hahn 1833) — bo-ne; Theridula gonygaster (Simon 1873) — Cos, in Pa- laearctic disjunctive distribution: in Asia known from Caucasus, Guangxi and Sichuan and Japan; in Nearctic known from Arizona and Florida. Thomisidae 3/1 Misumena vatia (Clerck 1757) — po; * Ozyptila arc- tica Kulczyhski 1908 — hy-mo, in Nearctic known from Alaska to western Northwest Terri- tories, south to northern British Columbia; Xys- ticus luctuosus (Blackwall 1836) — bo-ne; X. ob- scurus Collet 1877 — bo-mo. Titanoecidae 0/1 * Titanoeca nivalis Simon 1874— bo-mo, in Nearc- tic restricted to the western half. 2005. The Journal of Arachnology 33:306-312 FAUNA AND ZOOGEOGRAPHY OF SPIDERS (ARANEAE) IN BULGARIA Christo Deltshev: Institute of Zoology, Bulgarian Academy of Sciences, 1 Tsar Osvoboditel Bid., 1000 Sofia, Bulgaria. E-mail: cdeltshev@zoology.bas.bg ABSTRACT. Bulgaria is home to 975 species of spiders in 41 families. This number was established after a critical review of the existing literature and taxonomic review of the available collections. The spiders are distributed in all districts of Bulgaria, occurring in lowlands, forests, mountains, caves and urban territories. According to their current distribution the established 975 species can be split into 27 zoogeographical categories, grouped into five major chorotypes (Cosmopolitan, Holarctic, European, Med- iterranean, Endemics). The largest number of species belongs to the widely distributed species in the Holarctic, but the most characteristic are the endemics. Their established number (76 species) is high and reflects the local character of the fauna. This phenomenon can be attributed to the relative isolation of the mountains compared with the lowlands in the context of paleo-enviromental changes since the Pliocene. Keywords: Europe, diversity, distribution, chorotypes The first information on the spiders fauna of Bulgaria came from the end of 19^'’ century (Pavesi 1876). Systematic investigation start- ed in the beginning of 20^^ century by P. Dren- sky (1913, 1921, 1929, 1931, 1936a, b, 1937, 1938, 1939, 1940, 1942, 1943). Drensky (1936a) published the only catalogue of the spiders on the Balkan Peninsula in which 624 species from Bulgaria were reported. More re- cent publications are a result of intensive fau- nistic research after 1967 (Deltshev 1967, 1972a, b, 1973, 1974, 1977a, b, 1978, 1980, 1983a, b, c, 1984, 1985, 1987a, b, 1988, 1990, 1992, 1993, 1996, 1997a, b, 1998, 2003; Deltshev & Blagoev 1995, 1997, 2001; Hels- dingen et al. 1977 2001]; Blagoev & Deltshev 1989; Blagoev et al. 2002; Dimitrov 1993, 1994, 1996, 1997, 1999, 2003; Dimitrov & Lazarov 1999, 2002; Thaler et al. 1994; La- zarov 1998, 2003, 2004; Lazarov et al. 2001; Tzonev & Lazarov 2001). The accumulation of new data makes possible a critical taxo- nomic and faueistic review, together with a zoogeographic analysis. METHODS The material treated herein can be divided into two major parts: the first comprises a crit- ical incorporation of all available literature re- cords concerning the distribution of spiders in Bulgaria; the second concerns the original col- lections obtained from 1965-2002 during a field survey covering most of the districts in Bulgaria, kept in the collections of Institute of Zoology, Bulgarian Academy of Sciences. RESULTS AND DISCUSSION The spider fauna is represented in Bulgaria by 975 species, included in 41 families and 285 genera. The number of species is high compared with the number of spiders recorded from other European countries with similar territories (Tables 1,2). The number of fami- lies is also high compared with the data for the world: 110 (Platnick 2005; Austria 40, Germany 39, Switzerland 39 (Blick et al. 2002). Best represented are the families Lin- yphiidae (226 species or 23.2%), Gnaphosidae (98 species or 10%), Salticidae (91 species or 9.3%), Lycosidae (80 species or 8.2%) and Theridiidae (74 species or 7.5%). The genera with the highest number of species are: Cen~ tromerus (16 species or 73%), Walckenaeria (14 species or 6.4%), Tenuiphantes (11 spe- cies or 5%) and Diplocephalus (9 species or 4.1%) (Table 2). This richness, however, de- pends not only on the size of the regions, but also on the degree of exploration by araneol- ogists. According to their current distribution Bul- garian spiders can be divided into 27 zoogeo- graphical chorotypes, grouped into 5 zoogeo- graphical complexes (I = Cosmopolitan, II = Holarctic, III = European, IV — Mediterra- nean, V = Endemic) (Fig. 1). The data con- cerning general distribution of spiders are tak- 306 DELTSHEV— THE SPIDERS OF BULGARIA 307 Table 1. — Comparison of area and spider species richness of some European countries. Country Area (km^) Spider species Sources Austria 83,858 961 Blick et al. (2002) Bulgaria 110,993 975 Blagoev et al. (2002) Czech Republic 77,280 830 Buchar & Ruzicka (2002) Greece 128,900 810 Bosmans (pers. comm.) Hungary 92,340 725 Samu & Szinetar (1999) Macedonia 25,713 558 Blagoev (2002) Portugal 91,500 660 Cardoso (1999) Serbia 102,000 618 Deltshev et al. (2003) Slovenia 20,120 529 Kuntner & Sereg (2002) en from Michailov (1997), Marasik et al. (2000), Platnick (2004) and Vigna Taglianti et al. (1999) (Fig. 1). Cosmopolitan species complex ( COS + SCO, 20, 2%): Includes especially widespread species associated with lowlands, woodlands and high elevation zones of mountains. Complex of species widely distributed in the Holarctic Region (HOL + OLW + PAT + PAL + WPA + ECA + EEC + SEC + EEE P WPA): Is best represented and comprises 561 (57.5%) species widespread in Bulgaria (Fig. 1). Palearctic species (sensu lato) are dominant (36.1%), followed by Holarctic (10.5%), European Central Asiatic (7%) and West Palearctic (3%). The remaining choro- types (EEC, SEC & EEE) are represented by single species. The complex includes espe- cially widespread species associated with low- lands, woodlands and high elevation zones of mountains. Most of the species are well rep- resented in the mountains. Characteristic mountain species are represented by the lin- yphiids Bolyphantes alticeps (Sundevall 1833), B. luteolus (Blackwall 1833), Fronti- nellina frutetorum (C. L. Koch 1834), Gona- tium rubens (Blackwall 1833), Pityohyphantes phrygianus (C. L. Koch 1836), Tenuiphantes alacris (Blackwall 1853), T. tenebricola (Wid- er 1834). High mountain species are the lin- Table 2. — The spider fauna of Bulgaria listed by families, depicting numbers of genera and species. Families Genera Species Families Genera Species Atypidae 1 2 Oxyopidae 1 3 Nemesiidae 2 4 Zoropsidae 1 2 Filistatidae 2 2 Zoridae 1 6 Scytodidae 1 1 Agelenidae 6 32 Leptonetidae 1 2 Cybaeidae 2 3 Pholcidae 4 7 Hahnidae 4 9 Segestridae 1 3 Dictynidae 8 15 Dysderidae 4 29 Amaurobiidae 5 22 Oonopidae 4 4 Titanoecidae 2 7 Mimetidae 2 4 Miturgidae 1 12 Eresidae 1 2 Anyphenidae 1 2 Oecobiidae 1 1 Liocranidae 7 13 Uloboridae 2 4 Clubionidae 1 26 Nesticidae 1 3 Corinnidae 3 5 Theridiidae 17 74 Zodariidae 1 11 Theridiosomatidae 1 1 Gnaphosidae 19 96 Linyphiidae 94 226 Sparassidae 2 3 Tetragnathidae 4 17 Philodromidae 4 32 Araneidae 16 56 Thomisidae 13 60 Lycosidae 11 80 Salticidae 31 91 Pisauridae 2 3 Total 285 975 308 THE JOURNAL OF ARACHNOLOGY 400 350 300 250 200 150 100 50 0 1 ■ -B-j-M-t-mL-m I — , ■ 1 . , - 0°“^ ^^v ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ Chorotypes Figure 1. — Zoogeographical types in the spider fauna of Bulgaria, showing the number of species represented in each. Abbreviations: COS = cosmopolitan; SCO = subcosmopolitan; HOL = Holarctic; OLW = Old World; PAT = Palearctic-Paleotropic; PAL = Palearctic; WPA = west-Palearctic; EC A = European-central Asian; EEC = east European-central Asian; SEC = south European-central Asian; EEE = east European-east Mediterranean; MCA = Mediterranean-central Asian; BCA = Balkan-central Asian; EUR = European; MEE = middle-east European; MSEE = middle-southeast European; SEU = south European; EEU = east European; SEE = southeast European; BCAU = Balkan-Caucasian; BAN = Balkan Anatolian; MED = Mediterranean; EME = east Mediteranean; NME = north Mediterranean; NEM = northeast Mediterranean; BALK = Balkan endemics; BULG = Bulgarian endemics. yphiids Entelecara media (Kulczynski 1887) and Mecynargus paetulus (O.R-Cambridge 1875), which are not established in the forest belt. Some xenotopic species (Thaler 1988) are widely distributed in the mountains and reach the highest summits as aeronauts. To this group belong the linyphiids Dicymbium nigrum (Blackwall 1834), Diplostyla concolor (Wider 1834), Meioneta rurestris (C.L. Koch 1836), Oedothorax agrestis (Blackwall 1853), O. apicatus (Blackwall 1850), O. fuscus (Blackwall 1834) which inhabit the mountain zone in dense populations (Deltshev 1990, 1995). European species complex (EUR + MEE + MSEE -h EEU+ SEE): Comprises 191 (20%) species, widespread in Europe and Bulgaria (Fig. 1). European species (sensu lato) are dominant (14%), followed by East European species (3%), and Middle Southeast European species (1.5%). The remaining chorotypes (MEE & SEE) are represented by single spe^ cies. The complex comprises widespread spe- cies which inhabit both lowland and moun- tains. Interesting is the group of European mountain species, best represented in the for- est, subalpine and alpine belts. Characteristic mountain species are the linyphiids, Araeon- cus anguineus Deltshev 1987, Bolyphantes kolosvaryi (Caporiacco 1936), Cinetata gra- data (Simon 1881), Diplocephalus foraminb fer (O.R-Cambridge 1875), Improphantes im- probulus (Simon 1929), Maso gallicus Simon 1894, Mughiphantes pulcher (Kulczynski 1881), Oreonetides glacialis (C.L. Koch 1872), Tiso vagans (Blackwall 1834). Other linyphiid species such as Palliduphantes istri- anus (Kulczynski 1914), Centromerus capu- cinus (Simon 1884), C. cavernarum (L. Koch 1872), Porrhomma lativelum Tretzel 1956 and P. microps (Roewer 1931), are characteristic of caves. Mediterranean species complex (MCA — MED + EME -h NME + NEM P SEUP BCA P BAN P BCAU): Includes 127 species (13%) that occur in the Mediterranean area or a part of it. The complex forms only 13% of the total spider fauna of Bulgaria, but the real percentage is probably higher, because a large part of the endemics have a Mediterranean or- DELTSHEV— THE SPIDERS OF BULGARIA 309 igin. Most of the species in the complex are widely distributed in the Mediterranean re- gion. Very interesting are the mountain-Med- iterranean species [Aculepeira talishia (Za- vadsky 1902), Pardosa incerta 1905], which may be regarded as ancient elements in the high mountains. Endemic species complex (BALK -h BULG): Includes 76 species (10%) estab- lished in Bulgaria (35 species) and other ter- ritories of the Balkan Peninsula (41 species). The established number is high and reflects the local character of the fauna. The question about the status and distribution of endemic spiders found in Bulgaria is complicated. Some of them are found only in restricted ar- eas, while others show wider distributions, sometimes even over the whole peninsula. According to their origin, the endemics form two groups. Some of the species can be regarded as probable remnants of ancient Mediterranean mountain fauna (paleoendem- ics), and others came from the northern parts of Europe during the glacials and evolved un- der isolation on mountains during the inter- glacials (neoendemics). The curious is the dis- tribution of the genus Antrohyphantes Dumitrescu 1971 (Linyphiidae), found only in the high elevation zone and in caves. It is re- lated to the genus Fageiella Kratochvfl 1934 (Linyphiidae), an endemic from the caves of the western part of the Balkan Peninsula (Bos- nia, Montenegro). Their allopatric distribution indicates that they had already separated be- fore the establishment of the Vardar tectonic zones (Deltshev 1996). This suggests that these two genera are paleoendemics. Concerning the formation of cave fauna, Deeleman-Reinhold (1976) wrote that “many European cave spiders are probably relics of populations of moist Tertiary forests”. Due to the lack of knowledge, it is difficult to deter- mine with certainty which of the cave spider endemics of Bulgaria are Tertiary and which are Quaternary elements. Nevertheless, the blind species in the family linyphiid, Cen- tromerus bulgarianus (Drensky 1931), Trog- lohyphantes drenskii Deltshev 1973 and Trog~ lohyphates bureschianus Deltshev 1975, all species with primitive three branched para- cymbia, also can be regarded as probable pa- leoendemics (Deltshev 1996). The linyphiid spiders Araeoncus clivifrons Deltshev 1987, Diplocephalus altimontanus Deltshev 1984, Drepanotylus pirinicus Deltshev 1992, Erigone L pirini Deltshev 1983, Incestophantes annulatus (Kulczynski 1882), Mughiphantes lithoclasicolus Deltshev 1983, Metopobactrus orbelicus Deltshev 1985, known only from the high alpine parts of the Pirin and Rila Mountains are high al- pine elements?. Here, also can be placed Thenuiphantes drenskyi Helsdingen 1977, oc- curring in the high elevation belts of Pirin, Rila, Central Stara Planina and Vitosha moun- tains. These species are regarded as derivative of their respective North or Middle European species {Diplocephalus picinus (Blackwall 1841), Drepanotylus borealis Holm 1945, Er- igone longipalpis (Sundevall 1830), Metopo- bactrus prominulus (O.P.-Cambridge 1872), due to the disjunction of ranges during the glacial and interglacial (Deltshev 1996; Deltshev & Blagoev 1997). The largest frac- tion of endemics was encountered mainly in caves, coastal sites, woodlands and high alti- tude zones. The presence of the 975 spider species shows that Bulgaria is a territory of consid- erable species richness. This conclusion is supported also by the existence of 76 endemic species. In a zoogeographical respect, the widely distributed spiders in the Holarctic re- gion are dominant. However, the most char- acteristic faunal elements are the endemics. Their number is high, and their faunistic com- position reflects the local character of the fau- na. According to their origin the endemics be- long to two principal faunistic complexes: Mediterranean and European. This phenome- non can be explained by the relative isolation of the mountains compared with the lowlands, in the context of palaeo-environmental chang- es that have occurred since the Pliocene. ACKNOWLEDGMENTS I am especially indebted to Dr. K. Thaler, Dr. J. Dunlop and E. Stojcoska for access to the materials in the collections of University of Innsbruck, Museum of Natural History, Berlin and Macedonian Museum of Natural History, and to my colleagues Gergin Bla- goev, Stoyan Lazarov, Z. Hubenov and S. Abadjiev for discussion and helpful assis- tance. LITERATURE CITED Blagoev, G. 2002. Check list of Macedonian spiders (Araneae). Acta Zoologica Bulgarica. 54(3):9- 34. 310 THE JOURNAL OF ARACHNOLOGY Blagoev, G. & C. Deltshev. 1989. Biotopical dis- tribution of wolf-spiders (Araneae, Lycosidae) in the Zemen Gorge, Southwestern Bulgaria. Ecol- ogy, Bulgarian Academy of Sciences 22:73-80. Blagoev, G., Deltshev, C. & S. Lazarov. 2002. The Spiders (Araneae) of Bulgaria. Institute of Zo- ology, Bulgarian Academy of Sciences. Online at http://cl.bas.bg/bulgarianspiders/ Blick, T, A. Hanggi & K. Thaler. (2002). Checklist of the arachnids of Germany, Switzerland, Aus- tria, Belgium and the Netherlands (Arachnida: Araneae, Opiliones, Pseudoscorpiones, Scorpi- ones, Palpigradi). Version 2002 June 1. Online at http://www. AraGes.de/checklist_e.html Buchar, J. & V. Ruzicka. 2002. Catalogue of Spi- ders of the Czech Republic. Edited by P. Merrett. Praha: Peres. 351 pp. Cardoso, P. 1999. Portuguese spiders checklist, http: //www/geocites.com/RainForest/Vines/ 1 5 1 97/ checklist.html Deeleman-Reinhold, C.L. 1976. Distribution pat- terns in European cave spiders. Bulletin of the South African Speleological Association: 25-35. Deltshev, C. 1967. On the studies of spiders (Ara- neae) in the Vitosha Mountain. Bulletin de Insti- tut de Zoologie et Musee, Sofia 24:51-56. Deltshev, C. 1972a. A contribution to the study of spiders (Araneae) from the caves in Bulgaria. II. Genus Lepthyphcmtes in Bulgarian caves. Bulle- tin de Institut de Zoologie et Musee, Sofia 36: 137-147. Deltshev, C. 1972b. A new genus of Bulgarian cave spiders (Protoleptoneta hiilgarica, n. g., n. sp., Leptonetidae). International Journal of Speleol- ogy 4:275-283 Deltshev, C. 1973. A new Troglohyphantes from Bulgarian caves (Araneae, Linyphiidae). Inter- national Journal of Speleology 5:103-109. Deltshev, C. 1974. A new Centromerus from Bul- garian caves (Araneae, Linyphiidae). Internation- al Journal of Speleology 6:81-86. Deltshev, C. 1977a. A new Protoleptoneta from caves (Araneae, Leptonetidae). Acta Zoologica Bulgarica 1:3-8. Deltshev, C. 1977b. Genus Nesticus (Nesticidae, Araneae) from Bulgarian caves. In V. Panos (ed.), Proceedings of 6th International Congress of Speleology, Olomouc 5:73-78. Deltshev, C. 1978. A new Histopona (Araneae, Agelenidae) from Bulgarian caves. Acta Zoolo- gica Bulgarica 10:57-59. Deltshev, C. 1980. A contribution to the taxonom- ical study of pallidus group of genus Lepthy- phantes Menge (Araneae, Linyphiidae) in Bul- garia. Acta Zoologica Bulgarica 16:44-56. Deltshev, C. 1983a. A contribution to the taxonom- ical study of sylvaticus group of genus Centrom- erus F. Dahl (Araneae, Linyphiidae) in Bulgaria. Acta Zoologica Bulgarica 21:53-58. Deltshev, C. 1983b. Notes on the spiders of genus Erigone Audouin (Araneae, Erigonidae) in Bul- garia. Acta Zoologica Bulgarica 22:71-75. Deltshev, C. 1983c. A contribution to the taxonom- ical and faunistical study of genus Lepthyphantes Menge (Araneae, Linyphiidae) from Pirin Moun- tain. Acta Zoologica Bulgarica 23:25-32. Deltshev, C. 1984. A new Diplocephalus species from Bulgarian mountains (Arachnida, Araneae, Erigonidae). Reichenbachia 22:91-93. Deltshev, C. 1985. A contribution to the study of the family Erigonidae (Araneae) from Pirin Mountain, Bulgaria, with a description of a new species (Metopobactrus orbelicus sp. n.). Bulle- tin of the British Arachnological Society 6:359- 366. Deltshev, C. 1987a. A critical review of genus Zo- darion Walckenaer (Araneae, Zodariidae) in Bul- garia. Acta Zoologica Bulgarica 33:19-25. Deltshev, C. 1987b. A critical review of genus Ar- aeoncus Simon in Bulgaria, with description of a new species (Araeoncus clivifrons sp. n.) (Arachnida, Araneae, Erigonidae). Reichenba- chia 25:97-102. Deltshev, C. 1988. The genus Fageiella Kratochvil and Antrohyphantes Dumitresco (Araneae, Lin- yphiidae, Lepthyphanteae) in the caves of Balkan Peninsula. TUB-Documentation 38:293-302. Deltshev, C. 1990. A critical review of genus Coe- lotes Blackwall in Bulgaria with description of a new species {Coelotes drenskii sp. n.) (Araneae, Agelenidae). Acta Zoologica Bulgarica 40:29- 43. Deltshev, C. 1992. A critical review of family Ther- idiidae (Araneae) in Bulgaria. Acta Zoologica Bulgarica 43:13-22. Deltshev, C. 1993. The genus Tegenaria Latreille in Bulgaria: A critical review with description of two sibling species (Arachnida, Araneae: Age- lenidae). Berichte des Naturwissenschaftlich- Medizinischen Vereins in Innsbruck 80:167-174. Deltshev, C. 1995. Spiders (Araneae) from the high altitude zone of Rila Mountain (Bulgaria). Beri- chte Natwissenschaftlich-Medizinischen Verein Innsbruck 82:217-225. Deltshev, C. 1996. The origin, formation and zo- ogeogaphy of endemic spiders of Bulgaria (Ar- aneae). Revue Suisse de Zoologie hors seriel41- 151. Deltshev, C. 1997a. A new Cybaeus from the Mountains of Balkan Peninsula. Reichenbachia 32:1-4. Deltshev, C. 1997b. Crypoecina deelemanae g. n., sp. n., a remarkable spider from the mountains of Montenegro (Yugoslavia) (Arachnida, Ara- neae, Hahnidae). Revue Suisse de Zoologie 104: 485-489. Deltshev, C. 1998. Spiders from the high altitude zone of Central Stara Planina Mountain (Bulgar- DELTSHEV^THE SPIDERS OF BULGARIA 311 ia) (Araneae). Berichte des Naturwissenschaf- tlich-Medizinischee Vereins in Innsbruck 85: 213-221 Deltshev, C. 2003. A critical review of the spider species (Araneae) described by P. Drensky in the period 1915-1942 from the Balkan. Berichte des Naturwissenschaftlich-Medizinischen Vereins in Innsbruck 90:135-150. Deltshev, C. & G. Blagoev. 1995. A critical review of family Lycosidae (Araneae) in Bulgaria. Re-- vue Arachnologique 10:171-198. Deltshev, C. & G. Blagoev. 1997. The spiders of Pirin Mountain (Bulgaria). Taxonomic, faunistic and zoogeographical analysis (Araneae). Beri- chte des Naturwissenschaftlich-Medizinischen Vereins in Innsbruck 84:269-286. Deltshev, C. & G. Blagoev. 2001. A critical check list of Bulgarian spiders (Araneae). Bulletin of the British Arachnoiogical Society 12:110-138. Deltshev, C., B.P.M. Cercic & G. Blagoev. 2003. The Spiders of Serbia. Institute of Zoology, Fac- ulty of Biology, University of Belgrade. 833 pp. Dimitrov, D. 1993. Nev/ and rare spiders (Araneae) to the fauna of Bulgaria. Second National Sci- entific Entomological Conference Sofia. Pp. 73- 75. Dimitrov, D. 1994. A record of Achaearanea ta- bulata from the Balkan Peninsula (Araneae: Theridiidae). Arachnologische Mitteilungen 8: 77-79. Dimitrov, D. 1996. Coelotes deltshevi sp. n., a new spider species from Bulgaria (Arachnida, Ara- neae, Zodariidae). Berichte des Naturwissen- schaftlich-Medizinischen Vereins in Innsbruck 83:159-161. Dimitrov, D. 1997. Description of Harpactea strandjica, sp. n. from Bulgaria (Araneae, Dys- deridae). Bulletin of the British Arachnoiogical Society 10:322. Dimitrov, D. 1999. The spider fauna of the Strand- zha Mountain (South-East Bulgaria). L Faunistic data and taxonomic remarks (Arachnida: Ara- neae). Acta Zoologica Bulgarica 52:15-26. Dimitrov, D. 2003. Erigonoplus spinifemuralis sp, n. (Araneae: Linyphiidae)— a new species from Bulgaria. Acta Zoologica Bulgarica 55:33-35. Dimitrov, D. & S. Lazarov. 1999. Two new species of Harpactea from Bulgaria (Araneae: Dysderi- dae), Berichte des Naturwissenschaftlich-Medi- zinischen Vereins in Innsbruck 86:127-129. Dimitrov, D. & S. Lazarov. 2002. A contribution to the study of the spiders (Araneae) in Chepun Mountain and Dragoman Swampland (NW Bul- garia). Acta Zoologica Bulgarica 54(2):47-53. Drensky, P. 1913. Uber die Spinnenfauna Bulgar- ians. Spisanie Bulgarska Akadademia na Nau- kata 2:1-146. Drensky, P. 1921. Contribution a 1 'etude des araig- nees de la Macedoine orientale et de Pirine plan- ina. Spisanie Bulgarska Akadademia na Naukata 23:1-50. Drensky, P. 1929, Spinnen (Araneae) aus Mittel und Sud-West Mazedonien. Spisanie Bulgarska Aka- dademia na Naukata 39:1-76, Drensky, P. 1931, Hohlen-Spinnen aus Bulgarien. Spisanie Bulgarska Akadademia Na Naukata 44: 1-50. Drensky, P. 1 936a. Katalog der echten Spinnen (Ar- aneae) der Balkanhalbinsel. Sbornik Bulgarskata Akademia na Naukata 32:1-223. Drensky, P. 1936b. Studien fiber die Bulgarischen Spinnenfauna ihre okologischen und biogeogra- phischen Besoederheiten. Travaux Societe Bul- gare des Sciences Naturelles 17:71-1 15. Drensky, P. 1937. Die Spinnenfauna Bulgariens. 1. Unterordnung Mygalomorphae, Family Ctenizi- dae und Atypidae. Izvestia Tsarski Prirodonauch- ni Instituti 10:259-280. Drensky, P. 1938. Die Spinnenfauna Bulgariens. 11. Unterordnung Arachnomorphae, I Gruppe Te- trastica, Familien: Filistatidae, Dysderidae und Oonopidae. Izvestia Tsarski Prirodonauchni In- stituti 11:81-1 13. Drensky, P. 1939. Die Spinnenfauna Bulgariens. III. Unterordnung Arachnomorphae, II Gruppe Trionichia, Familien: Urocteidae, Uloboridae, Si- caridae, Pholcidae, Eresidae. Izvestia Tsarski Pri- rodonauchni Instituti 12:231-252. Drensky, P. 1940. Die Spinnenfauna Bulgariens. IV Unterordnung Arachnomorphae, II Gruppe Trionichia, Familien: Zodariidae, Dictynidae und Amaurobiidae. Izvestia Tsarski Prirodonauchni Instituti 13:169-194. Drensky, R 1942. Die Spinnenfauna Bulgariens. V. Unterordnung Arachnomorphae, II Gruppe Trionichia, Familie: Agelenidae. Izvestia Tsarski Prirodonauchni Instituti 15:33-60. Drensky, P. 1943. Die Spinnenfauna Bulgariens. VL Unterordnung Arachnomorphae, II Gruppe Trionichia, Familie Euetrioidae. Izvestia Tsarski Prirodonauchni Instituti 16:219-254. Helsdingen, van, P.I., K. Thaler & C. Deltshev. 1977 The tenuis group of Lepthyphantes Menge (Araneae, Linyphiidae). Tijdschrift voor Ento- mologie 120:1-54. Helsdingen, van, P.I., K. Thaler & C. Deltshev. 2001. The European species of Bolyphantes with an attempt of a phylogenetic analysis (Araneae, Linyphiidae). Memorie della Societa Entomolo- gica Italaliana 80:3-35. Kuntner, M. & I. Sereg. 2002. Additions to the spi- der fauna of Slovenia, with a comparison of spi- der species richness among European countries. Bulletin of the British Arachnoiogical Society 12:185-195. Lazarov, S. 1998. A contribution to the study of the spiders (Araneae) in Sushtinska Sredna Gora 312 THE JOURNAL OF ARACHNOLOGY Mountain, Bulgaria. Historia Naturalis Bulgarica 9:27-34. Lazarov, S. 2003. A review of the family Dysder- idae (Araneae) in Bulgaria: faunistic and zoo- geographical analysis. Pp. 259-265. In European Arachnology 2002 (E Samu & C. Szinetar, eds) Plant Protection Institute and Berzsenyi College, Budapest Lazarov, S. 2004. A contribution to the study of spiders (Araneae) in Macedonia. Acta Zoologica Bulgarica 56:155-166. Lazarov, S., C. Deltshev & G. Blagoev. 2001. Spi- ders (Araneae) of Sushtinska Sredna Gora Moun- tain, Bulgaria. Acta Zoologica Bulgarica 53:3- 28. Marusik, Y.M., D.V. Logunov & S. Koponen. 2000. Spiders of Tuva, South Siberia. Magadan. IBPN FEB RAS, 2000. 252 pp. Michailov, K.G. 1997. Catalogue of the spiders of the territories of the former Soviet Union (Arach- nida, Aranei). Moscow, Zoological Museum of the Moscow State University. 416 pp. Pavesi P. 1876. Gli Arachnid! Turchi. Atti della So- cieta Italiana di Scienze Natural! e del Museo Civico di Storia Naturale 19:1-27. Platnick, N. 2004. The World Spider Catalog, Ver- sion 4.5 American Museum of Natural History, online at http://research.amnh.org/entomology/ spiders/catalogSl -87/index. html. Samu, F. & C, Szinetar. 1999. Bibliographic check list of the Hungarian spider fauna. Bulletin of the British Arachnological Society 11:161-184. Thaler, K. 1988. Arealformen in der nivalen Spin- nenfauna der Ostalpen (Arachnida, Aranei). Zoologischerl Anzeiger 220:233-244. Thaler, K., P. Van Helsdingen & C. Deltshev. 1994. Vikariante Verbreitung im Artenkomplex von Lepthyphantes annulatus in Europa und ihre Deutung (Araneae, Linyphiidae). Zoologischer Anzeiger 232:111-127. Tzonev, G. & S. Lazarov. 2001. A contribution to the study of spiders (Araneae) in Osogovo Mountain, south-west Bulgaria. Acta Zoologica Bulgarica 53:67-78. Vigna Taglianti, A., PA. Audisio, M. Biondi, M.A, Bologna, G.M. Carpaneto, A. De Biase, S. Fat- torini, E. Piattella, R. Sindaco, A. Venchi & M. Zapparoli. 1999. A proposal for a chorotype clas- sification of the Near East fauna, in the frame- work of the Western Palearctic region. Biogeo- graphia 20:31-59. Manuscript received 10 June 2005, revised 15 Sep- tember 2005. 2005= The Journal of Arachnology 33:313-322 GEOGRAPHICAL CONTEXT OF SPECIATION IN A RADIATION OF HAWAIIAN TETRAGNATHA SPIDERS (ARANEAE, TETRAGNATHIDAE) Rosemary G, Gillespie: Division of Insect Biology, University of California Berkeley, 137 Molford Hall, Berkeley, California 94720-3114, U=S=A= ABSTRACT. Adaptive radiation involves the diversification of species each adapted to exploit different ecological roles. I have studied a radiation of spiders in the genus Tetragnatha (Tetrageathidae) in the Hawaiian Islands to elucidate processes involved in such diversification. The temporal framework of the Hawaiian Islands allows examination of the changing pattern of adaptive radiation over time, as lineages have generally progressed down the island chain from older to younger islands. Species of Tetragnatha in the spiny-leg clade on any one island are typically most closely related to others on the same island, and the same set of ecological forms (ecomorphs) has evolved repeatedly on different islands. These results indicate that adaptive radiation frequently involves ecological divergence between sister taxa to allow multiple close relatives to co-occur in the same habitat. The current study examines the geographical context within which these species arose. I focus on a clade of 5 species that occur on the volcano of East Maui; at any given site 3 species can co-occur, one of each of 3 different ecomorphs. Mitochondrial DNA sequences from populations of these 5 species from throughout their distribution (Maui, Lanai and Molokai) were used to infer the geographic history of the species on East Maui and to determine whether diversification likely occurred in situ, or alternatively whether diversification occurred in allopatry on different volcanoes. Although ecological differentiation between taxa is evident, allopatry is clearly im- plicated in the initial divergence of taxa. Further study is required to understand the nature of the interplay between allopatry and ecological divergence in species formation. Keywords: Adaptive radiation, biogeography, allopatry, parapatry, evolution One of the most hotly debated aspects of the speciatioe process is its geographical con- text, and the nature and importance of isola- tion in the initial divergence of taxa (Coyne & Orr 2004). As a result of the infiueetial work of Mayr (1963), ideas of speciation were dominated for many years by the importance of allopatry in initiating divergence (Coyne 1994; Howard & Berlocher 1998). However, theoretical studies have demonstrated that sympatric speciation can occur and can cause species to form much more rapidly than by allopatric speciation (Turelli et al. 2001; Gav- rilets 2003). The importance of sympatric spe- ciation in nature, however, remains question- able (Coyne & Orr 2004). Recent studies on species and speciation have started to recog- nize the validity of some of the predominant ideas of the earlier part of this century, in- cluding the role of both divergent natural se- lection (Schluter 2001) and hybridization (Seehausen 2004) in generating new species. In particular, ecological speciation, in which reproductive isolation evolves as a conse- quence of divergent natural selection on traits between contrasting environments, is now rec- ognized as an important mechanism of spe- ciation (Schluter 2001). However, the geo- graphic context of speciation in situations of adaptive radiation, where multiple close rela- tives occur in sympatry, is still the subject of considerable debate (Glor et al. 2004). The current study focuses on the Hawaiian Islands, the most isolated archipelago in the world and well known for some of the most extraordinary illustrations of adaptive radia- tion (Simon 1987; Wagner & Funk 1995). The Hawaiian island chain is a hotspot archipela- go, arranged in chronological series, the youn- gest island being Hawaii, the oldest Kauai (Carson & Clague 1995; Price & Clague 2002). The biogeographic pattern that pre- dominates in most Hawaiian taxa, both spe- cies and populations, is a step-like progression down the island chain from the oldest to the youngest islands (Wagner & Funk 1995), of- ten with repeated bouts of diversification within islands (Roderick & Gillespie 1998). 313 314 THE JOURNAL OF ARACHNOLOGY 5.1 myrs KAUA'I 3.7 m 1. 6 myrs % OAHU MOLOKA’KMAUI nui -J ^ — " — ■ 9 myrs^ LANA’I^ 1.3 myrs 1 Maui\\ .3 myrs MAUI 0.8 fjfiyrs ^ Maui KAHO’OLAWE Figure 1. — Map of the Hawaiian Islands. Names in bold type indicate islands that were the focus of the current study. Ages of the different volcanoes are given (myrs = million years). Accordingly, the islands are considered a “natural laboratory” as they allow study of patterns of species formation on islands of dif- ferent age (Gillespie 2004, 2005). Here, I fo- cus on an adaptive radiation in the spider ge- nus Tetragnatha (Tetragnathidae) in the Hawaiian Islands to examine the geographic context involved in the initial divergence of taxa. The genus Tetragnatha is strikingly diverse in the Hawaiian Islands, with multiple species occurring in sympatry throughout the islands. Until 1991, only 8 species had been described from the islands. Over the last few years I have described an additional 29 species of Ha- waiian Tetragnatha (Gillespie 1991, 1992a, 1994, 2002, 2003) and am currently describ- ing approximately 15 more species. This spe- cies radiation encompasses forms representing a huge spectrum of colors, shapes, sizes, eco- logieal affinities and behaviors. Many species are web builders, with their shapes modified to allow concealment within specific micro- habitats (Blackledge & Gillespie 2004). Some species have modifications of the jaws, appar- ently to allow specialization on specific prey types (Gillespie 2005). However, several groups have abandoned the characteristic web-building behavior of the genus (Gillespie 1991, 1992b). For example, one entire clade, or lineage, of 16 species (the “spiny leg” clade), has “lost” web building behavior, with the concomitant development of long spines along the legs and adoption of a vagile, cur- sorial, predatory strategy. Recent studies have shown that represen- tatives of the spiny leg clade occur as four distinct ecomorphs associated with specific habitat types: "'green spiny"' on leaves; "‘ma- roon spiny” on moss, "large brown spiny” on tree bark, and "small brown spiny” on twigs (Gillespie et al. 1994; Gillespie et al. 1997). Similar sets of ecomorphs occur in most native habitats and phylogenetic analy- ses have shown that ecomorphs have arisen repeatedly and independently (Gillespie 2004). In particular, the most ubiquitous eco- morph, green spiny, has evolved (or been lost) at least once on each of the older islands, Kau- ai (T. kauaiensis), Oahu (T. tantalus, T po~ lychromata), and Maui Nui, the once con- nected volcanoes of Molokai, Lanai, and Maui {T. brevignatha, T. macracantha, T. waika- moi). Likewise, the maroon spiny has evolved independently on Oahu {T. perreirai) and Maui Nui {T. kamakou); both species are GILLESPIE^SPECIATION IN HAWAIIAN TETRAGNATHA 315 closely related to species of the green spiny ecomorph. Also, one of the small brown spiny ecomorphs {T. restricta) has evolved indepee= dently oe Maui. The island of Hawaii, pre= sumably because it is still very young, con- tains mostly populations of the same species that occur oe Maui (Gillespie 1991). The distribution of ecomorphs across hab- itats is significantly different from random (Gillespie 2004): there is a remarkably similar representation of ecomorphs in different hab- itats. Not all habitats have all ecomorphs, but there is never more than one representative of a given ecomorph at a site. The finding that similar ecomorphs never co-occur is most striking on East Maui. Here, a representative of each ecomorph is found at almost every site on the volcano, yet the species composition of the array of four different ecomorphs changes quite markedly between different locations (Gillespie 2005). Different species of the same ecomorph have very clear cut parapatric dis- tributions. Moreover, different ecomorphs that co-occur are frequently sister species, sug- gesting the possibility that ecological differ- ences may have arisen in situ. However, pop- ulations of these species also occur on other volcanoes, suggesting a potential role for ak lopatry in the initial divergence of taxa. Here, I focus oe a group of sympatric spe- cies to determine how divergence may have occurred within the geographic context of the Maui Nui island complex. Specifically, I ex- amined 5 species that are found oe East Maui: T. kamakou (maroon spiny), T. restricta (small brown spiny), and T. waikamoi, T, ma- cracantha, and T. brevignatha (green spiny). (The large brown ecomorph is represented at all sites by T. quasimodo, but this species falls outside the clade of 5 species which form the focus of the current study). Tetragnatha ka- makou and T. restricta co-occur with each other and with one of the green spiny eco- morphs (T. waikamoi, T. macracantha, or T. brevignatha) at different locations oe the vol- cano. The question addressed here is whether species in the Maui Nui clade formed through diversification within the single volcano of East Maui or alternatively whether divergence occurred in allopatry, prior to their current distribution. METHODS Study Sites and Organisms.' — The study was focused oe the more recent part of the Hawaiian archipelago, Maui Nui and Hawaii (Fig. 1). Maui Nui is a composite of 4 separate islands, Maui, Molokai, Lanai, and Kahoola- we. Until 300,000-400,000 years ago, these islands were all connected, much like the is- land of Hawaii is today (Carson & Clague 1995) . Glacially mediated fluctuations in sea level have alternately flooded and exposed the land connecting islands of the complex of is- lands. Except for Kahoolawe, all islands of Maui Nui have been sufficiently high to main- tain native forest. Each island has a single high volcano except for Maui itself, which has two. These volcanoes range in age from Mo- lokai (1.8 MY), through Lanai and West Maui (1.3MY) to East Maui (0.8MY). The island of Hawaii is the largest in the archipelago and the youngest. It consists of 5 volcanoes, the oldest being Kohala (0.43MY), then Hualalai (0.40MY), Mauna Kea (0.38MY), Mauna Loa (0.20MY), and Kilauea (O.IOMY). The five focal species for the study were T. brevignatha, T. macracantha, T. kamakou, T. restricta, and T. waikamoi. Specimens collect- ed from different sites are shown in Table 1. As outgroups, I used two populations of T. quasimodo. East Maui, Waikamoi, Carruthers, 6100ft, 26 June 1994; and Hawaii, Puu Ma- kaala, 11 July 1994. Phylogenetic Hypotheses. — An approxi- mately 730 base pair piece of Cytochrome ox- idase subunit I (COI) was amplified using primers ECO- 1628 (ATAATGTAATTGT- TACTGCTCATGC) and HCO-2396 (ATTGT- AGCTGAGGTAAAATAAGCTCG) (Palumbi 1996) . Genbank accession numbers are given in Table 1. Historical hypotheses of phyloge- netic relationships were reconstructed using three methods: (i) Maximum Parsimony as the optimality criterion in the program PAUP* version 4.0b 10 (Swofford 2000). Heuristic searches were performed by step-wise addi- tion of taxa, with TBR branch swapping and 1000 step-wise random taxon addition repli- cates. Characters were weighted (transver- si'Ons: transitions) 2:1. Of the total characters: 543 characters were constant, 143 variable characters were parsimony-informative, and 44 characters were parsimony-uninformative, (ii) Maximum Likelihood as the optimality criterion. MODELTEST v, 3.04 (Posada and Crandall 1998), which makes use of log like- lihood scores to establish the model of DNA evolution that best fits the data, was first 316 THE JOURNAL OF ARACHNOLOGY Figure 2. — Phylogenetic hypothesis based on DNA sequence of partial mitochondrial cytochrome oxi- dase I. Phylogeny reconstruction used Maximum Parsimony (tree length 523, Consistency index = 0.58, Retention index = 0.74), Maximum Likelihood (-Ln likelihood = 3376.47142), and Bayesian inference of likelihood, with each analyses producing a similar topology. Support for each node was assessed through bootstrap values > 50% for Maximum Parsimony (below node) and for Maximum Likelihood (above each node, left), and Posterior Probabilities for the Bayesian analysis (above each node, right). GILLESPIE-SPECIATION IN HAWAIIAN TETRAGNATHA 317 brevignatha macracantha brevignatha kamakou restrkta kamakou waikamoi Figure 3. — Ancestral state reconstruction to infer transformations in ecomorph category and island (or volcano within island). Ecomorph category and island affiliation were mapped on to the molecular phy- logeny using the accelerated transformation (ACCTRAN) optimization. used to determine parameter values. The HKY85+G+I model was selected: negative log likelihood, 3381.6406, and Akaike Infor- mation Criterion (AIC), 6775.2812. This is a general time-reversible model of DNA substi- tution with a gamma distribution for the rate of substitution at a site and a shape parameter gamma distribution of 1.4870. The proportion of invariable sites (I) was estimated as 0.612. The base frequencies were estimated as: A, 0.2586; C, 0.1650; G, 0.2058; T, 0.3705 with a ti/tv ratio of 2.9285. (iii) Bayesian Inference of Likelihood with posterior probability of phytogenies approximated by sampling trees from the posterior probability distribution. The program MrBayes (Huelsenbeck 2000) uses Markov chain Monte Carlo (MCMC) to sample phylogenies according to their poste- rior probabilities, with the marginal probabil- ity of trees calculated from the trees visited during the course of the MCMC analysis. The proportion of the time any single tree is found in this sample is an approximation of the pos- terior probability of the tree. The estimates from MODELTEST were used as priors for the Bayesian inference of phylogeny using MrBayes 3.0. MacClade 4 (Maddison and Maddison 2000) was used to overlay eco- morph category on the molecular phylogeny and determine the most parsimonious scenario for ecomorph evolution. Both accelerated and delayed transformation optimization options were applied. RESULTS Phylogenetic estimation using Maximum Parsimony, Maximum Likelihood and Bayes- ian Inference gave a similar topology (Fig. 2). Support for nodes, which was provided by bootstrap support for Maximum Parsimony 318 THE JOURNAL OF ARACHNOLOGY and Maximum Likelihood and Posterior Prob= abilities for the Bayesian Inference of Likeli- hood, was generally high. Geographical loca- tions were mapped against the tree topology to determine probable ancestral geographic af- hnities for each species/ population. The re- sults gave strong support for the following clades: “brevignatha Maui”; “brevignatha Hawaii”; “all macracantha Maui + Lanai”; “kamakou Maui”; “kamakou Molokai”; “all restricta Maui + Hawaii”; and “all waikamoi Maui”. There was also strong support for the clade “restricta + only kamakou Molokai”; this renders T. kamakou paraphyletic with re- spect to r. restricta. Most importantly, how- ever, it suggests that, although T. restricta (Maui + Hawaii) co-occurs with T. kamakou on Maui, and T. kamakou is the sister species of T. restricta, the population of T. kamakou which is sister to T. restrica is not on Maui, but on Molokai, another volcano in the Maui Nui group of islands. There was weak support for a clade of “brevignatha Maui” + “macracantha Maui + Lanai”, which, if upheld, would render T. bre- vignatha paraphyletic relative to T. macracan- tha. The character reconstructions to examine the evolution of ecomorphs are shown in Fig. 3, using ACCTRAN, which minimizes paral- lel evolution. The topology indicates that a minimum of 3 character transformations are required, one with the divergence of T. ka- makou from T. brevignatha + T. macracan- tha, the second with the divergence of T. res- tricta from T. kamakou Molokai. Also, based on the reconstruction, E. Maui appears to be the ancestral geographic locality, although this may be because E. Maui, with by far the larg- est land mass in the island complex, has the largest number of species and thus is the is- land most likely to have haplotypes repre- sented across the tree. DISCUSSION Throughout the native forest on the volcano of Haleakala, East Maui, four species of spiny leg Tetragnatha can co-occur, and when they do, only one of each of the primary eco- morphs is found in any given location; green spiny {T. waikamoi, T. brevignatha, or T. ma- cracantha), maroon spiny {T. kamakou), small brown spiny {T. restricta or T. kikokiko), and large brown spiny {T. quasimodo). In this pa- per I focused on five species of Hawaiian Te- tragnatha that form a Maui Nui clade and are each others’ closest relatives (Gillespie 2004): T. waikamoi, T. brevignatha, T. macracantha, T. kamakou, and T. restricta. Tetragnatha ka- makou, and T. restricta co-occur with T. wai- kamoi, T. brevignatha, or T. macracantha throughout their distribution. The predominant ecomorph and the ecomorph represented by the most number of species, is green spiny, with T. waikamoi being sister to the remaining species in the clade, and T. brevignatha and T. macracantha forming sister species. Previous work has shown that different spe- cies of green spiny never co-occur (Gillespie 2004, 2005), suggesting that ecological diver- gence between taxa might involve some de- gree of isolation. The results presented here do not reject the hypothesis that isolation plays a role in the divergence of sister species of the same ecomorph. For example, the green spiny T. macracantha (from East Maui and Lanai) is monophyletic and sister to the East Maui population of green spiny T. brevigna- tha. However, because T. macracantha has a population on Lanai as well as East Maui, it is possible that divergence between these two taxa occurred in allopatry. The most informative result comes from the clade including T. kamakou and T. restricta, which is sister to the clade comprising T. bre- vignatha and T. macracantha. In particular, the mitochondrial tree shows that East Maui populations of T. kamakou and T. restricta, which occur in sympatry throughout much of the volcano, are each more closely related to populations on other volcanoes: for example, T. restricta is most closely related to T. ka- makou from Molokai and individuals of T. ka- makou on East Maui are most closely related to individuals of T. kamakou on West Maui. One explanation for this pattern could be col- onization of Hawaii by T. restricta, diver- gence in allopatry, and subsequent recoloni- zation of Maui. However, given the tendency of taxa to colonize from older to younger is- lands, and not the reverse (Wagner & Funk 1995), this scenario would be unusual. A sec- ond scenario is suggested by the reconstruc- tion of historical geographical areas (Fig. 3), which indicates that T. restricta diverged from T. kamakou on East Maui, with T. kamakou going on to colonize Molokai. However, it is difficult to suggest a process that might lead GILLESPIE-SPECIATION IN HAWAIIAN TETRAGNATHA D C n o S ^ J Pii S J ! I s d pQ •>' r)^^^i^r^vO'Ot^ooo ^-■o^^'l>^'a^o^^-^-^'t^l>r'0^r-r-!^o^r-r^o^^^r-l^c^o^r^^^!m CN|00r^fNCN|00 00CNCMfNrs|(N(N00CNlCNrN|00(NrN|00rNl(NrN|(N00(NOO oor^ooooooi>r'OOoo(»oooooor'00oooor^oooot^cx)ooooooi^ooaNON oodooSoSdSdodoooSdSoddddddd>< ? QQQQQQPQQQQQQQQQClQQQQQQQQQQ<< O ^ g g ^ ,D -D O (N 00 cs| > 1—1 ON ON o o o o\ On On ON ON d d ON ON ON On On ON *_> ON ON M_j >5 ON On CM CM CM 1— H 1— M r— ( d G F— I 1—1 I— ( F— ( 1—1 00 F-H 00 5-< o D d (0 d o « > O (U d 00 ns O d 0) d (D d Q d d bX) d d 3 U4 d bJ) d d d d 0) d d d ►-S d 1—5 3 d d 1—5 d d d d 1—5 3 1—5 d < 3 s <0 Ph < d d 3 ON cn r-M CM Ov ,-M in VO r- o o ON r- cn in in m cn r- VO m VO m cn VO -H a a i+Q •a «S iS tS tS t+Q vS iS in 1 — 1 F— ( CM o f-H CM o cn in in F— ( cn in in cn m cn cn cn cn VO cn cn cn cn VO VO in VO 00 w w ^ d d O O P •d X d d f2 d d 5 3 ^ g fG d 2 d o d ^ ^ X is P w 3 S ^ o . . o cd (U W ^ ^ -P d d “ c3 ^ d S S w ^ ^ w >3 s ,b0 -4^ d I S I w 5 319 320 THE JOURNAL OF ARACHNOLOGY to such an inferred sequence of events. Final- ly, a third scenario is that the divergence of T. restricta from T. kamakou was initiated by colonization from the older Molokai. For ex- ample, if East Maui were occupied by T. ka- makou and the volcano was subsequently col- onized a second time by T. kamakou from Molokai, disruptive selection could lead to the formation of a new, ecologically differentiated species. This final scenario of an older ances- tor is also supported by the inferred time of divergence between T. restricta and T. ka- makou: The maximal uncorrected pairwise ge- netic divergence between these species is 9.1 %, an amount which, when scaled to a global arthropod rate of mitochondrial sequence di- vergence of 2.3% per million years (Brower 1994), minimally dates the ancestor to 1.98 MYA, the approximate age of Molokai. How- ever, if T. restricta diverged from T. kamakou Molokai before E. Maui was formed 0.8MYA (Fig. 1) as the dating suggests, the geograph- ical context of the divergence remains enig- matic, as T. restricta has not been found on any island older than E. Maui. Further sam- pling of populations and genetic loci may help resolve this issue. The results also indicate that both T. bre- vignatha and T. kamakou are paraphyletic when considering populations on the different volcanoes: T. brevignatha (E. Maui + Hawaii) with respect to T. macracantha (E. Maui + Lanai) and T. kamakou (W. Maui + E. Maui) with respect to T. restricta (E. Maui + Ha- waii). Paraphyly at the species level is not un- common (Funk & Omland 2003), and may even be expected under the scenario of eco- logical speciation. Among island radiations, the phenomenon has been clearly demonstrat- ed in species of beetles (Rees et al, 2001) and lizards (Thorpe et al. 1994) in the Canary Is- lands and also a radiation of anole lizards in the Caribbean (Thorpe & Stenson 2003). In the current study, T. restricta has clearly emerged within T. kamakou: all extant allo- patric populations of T. kamakou are very similar morphologically and do not warrant distinct species status. One limitation of the current study is that it is based entirely on mitochondrial DNA se- quences. Mitochondrial gene trees can fre- quently conflict with species trees, generally as a result of unsorted ancestral variation (lin- eage sorting) or hybridization (Rokas et al. 2003). Among Hawaiian arthropods, studies of both flies {Drosophila) and crickets {Lau- pala) have shown marked differences between trees generated from nuclear DNA versus mi- tochondrial DNA. For example, interspecific hybridization has been a regular occurrence in the history of both Drosophila (DeSalle & Giddings 1986) and Laupala (Shaw 2002; Mendelson & Shaw 2005). However, for the Hawaiian Tetragnatha, there is no evidence that hybridization between species occurs reg- ularly. Considerable work on the phylogenetic relationships among the species considered here based on nuclear loci (allozymes and minisatellites) (Pons & Gillespie 2003, 2004; Gillespie 2004) shows no evidence of unsort- ed historical variation or recent introgression between any of the species or populations within species. Thus, we can assume that the mitochondrial data do accurately represent the phylogenetic history of the species examined in the current study. That Tetragnatha do not hybridize, despite the young age of many spe- cies, is interesting in comparison with Dro- sophila and Laupala: In both the flies and crickets, sexual selection has been implicated as a major force in driving speciation (Kane- shiro 1989; Shaw & Herlihy 2000). By con- trast, sexual selection appears not to play such a key role in Hawaiian Tetragnatha (Roderick & Gillespie 1998); rather, ecological affinities appear to be of greater significance in the ini- tial stages of differentiation. Although still largely conjecture at this point, ecological af- finity may play a key role in reinforcing iso- lation of gene pools in Tetragnatha, a process that perhaps is not as important in Drosophila and Laupala. Together with previous studies on the spiny leg clade (Blackledge & Gillespie 2004; Gil- lespie 2004), the results of the current study suggest that there is a strong ecological com- ponent to species diversification. These results corroborate similar findings that have ap- peared in the literature for other adaptive ra- diations. In particular, extensive within-habitat proliferation, and repeated evolution of simi- lar ecomorphs in different habitats has been found in cichlid fish in the Great African Lakes (Ruber et al. 1999), sticklebacks in Ca- nadian glacial lakes (Schluter & McPhail 1993; Schluter 1998; Schluter 2000), and An- olis lizards in the Caribbean (Losos et al. 1998). In most of these cases, species pairs GILLESPIE--SPECIATION IN HAWAIIAN TETRAGNATHA 321 appear to have had ae allopatric phase in their recent history (e.g., threespiee sticklebacks and Darwin’s ground inches). The only ex- ception in which no allopatric phase is indi- cated is that of the Rhagoletis flies (Feder et ah 1988; Filchak et aL 2000), although even here the possible role of allopatry cannot be ruled out (Coyne & Orr 2004). The radiation of Hawaiian Tetragnatha suggests, as do re- sults from studies of many other adaptive ra- diations, that allopatry, together with ecolog- ical divergence, plays an important role in the formation of species. Future studies on the in- terplay between divergence in allopatry and ecological differentiation will be critical to understanding the mechanism of speciation within an adaptive radiation. ACKNOWLEDGMENTS The work reported here was supported by funds from the Schlinger Foundation, the Na- tional Science Foundation, the University of Hawaii Research Council, and UC Berkeley. Additional support was provided by the Bish- op Museum, Haleakala National Park, the Na- ture Conservancy of Hawaii, the State De- partment of Land and Natural Resources, and the Hawaii Natural Areas Reserve System. For help with collecting I owe particular thanks to M. Amedo, G. Binford, L. Brom- ham, C. Ewing, M. Heddle, A.C. Medeiros, G.S. Oxford, D. Polhernus, M. Rivera, G. Roderick, K. Shaw, and A.M. Tan. LITERATURE CITED Blackiedge, T.A. & R.G. Gillespie. 2004. Conver- gent evolution of behavior in an adaptive radia- tion of Hawaiian web-building spiders. Proceed- ings of the National Academy of Sciences 101: 16228-16233. Brower, A.V.Z. 1994. Rapid morphological radia- tion and convergence among races of the butter- fly Heliconius erato inferred from patterns of mi- tochondrial DNA evolution. Proceeding of the National Academy of Sciences 91:6491-6495. Carson, H.L. & D.A. Clague, 1995. Geology and biogeography of the Hawaiian Islands. Pp. 14- 29. In Hawaiian Biogeography Evolution on a Hot Spot Archipelago. (W.L. Wagner & V.A. Funk, eds.). Smithsonian Institution Press, Wash- ington, D.C. Coyne, J.A. 1994. Ernst Mayr and the origin of spe- cies. Evolution 48:19-30. Coyne, J.A. & H.A. Orr. 2004. Speciation. Sinauer Associates, Sunderland, Massachusetts. Desalle, R. & L.V, Giddings. 1986. Discordance of nuclear and mitochondrial DNA phylogenies in Hawaiian Drosophila, Proceedings of the Na- tional Academy of Sciences 83:6902-6906. Feder, J.L., C.A. Chilcote & G.L. Bush. 1988. Ge- netic differentiation between sympatric host rac- es of the apple maggot fly Rhagoletis pomonella. Nature (London) 336:61-64. Filchak, K.E,, J.B. Roethele & J.L. Feder, 2000. Natural selection and sympatric divergence in the apple maggot, Rhagoletis pomonella. Nature 407:739-742. Funk, D.J. & K.E. Omland. 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mito- chondrial DNA. Annual Review of Ecology, Evolution and Systematics 34:397-423. Gavrilets, S. 2003. Models of speciation: what have we learned in 40 years? Evolution 57:2197-2215. Gillespie, R.G. 1991. Hawaiian spiders of the genus Tetragnatha: 1. Spiny leg clade. Journal of Ar- achnology 19:174-209. Gillespie, R.G. 1992a. Hawaiian spiders of the ge- nus Tetragnatha II. Species from natural areas of windward East Maui. Journal of Arachnology 20:1-17. Gillespie, R.G. 1992b. Impaled prey. Nature 355: 212-213. Gillespie, R.G. 1994, Hawaiian spiders of the genus Tetragnatha: III. Tetragnatha acuta clade. Jour- nal of Arachnology 22:161-168. Gillespie, R.G. 2002. Hawaiian spiders of the genus Tetragnatha: IV: New, small species in the spiny leg clade. Journal of Arachnology 30:159-172. Gillespie, R.G. 2003. Hawaiian spiders of the genus Tetragnatha: V. Elongate web-builders from Oahu. Journal of Arachnology 31:8-19. Gillespie, R.G. 2004. Community assembly through adaptive radiation in Hawaiian spiders. Science 303:356-359. Gillespie, R.G. 2005. The ecology and evolution of Hawaiian spider communities. American Scien- tist 93:122-131. Gillespie, R.G., H.B. Croom & G.L. Hasty. 1997. Phylogenetic relationships and adaptive shifts among major clades of Tetragnatha spiders (Ar- aneae: Tetragnathidae) in Hawaii, Pacific Sci- ence 51:380-394. Gillespie, R.G., H.B. Croom & S.R. Palumbi. 1994. Multiple origins of a spider radiation in Hawaii. Proceedings of the National Academy of Scienc- es 91:2290-2294. Glor, R.E., M. E. Gifford, A. Larson, J. B. Loses, L. Rodriguez-Schettino, A. R. Chamizo-Lara, and T. R. Jackman. 2004, Partial island submer- gence and speciation in an adaptive radiation: a multilocus analysis of the Cuban green anoles. Proceedings of the Royal Society of London B 271:2257-2265. Howard, D.J, & S.H. Berlocher. 1998. Endless 322 THE JOURNAL OF ARACHNOLOGY Forms: Species and Speciation. Oxford Univer- sity Press, Oxford. Huelsenbeck, J.P. 2000. MrBayes: Bayesian infer- ence of phylogeny. University of Rochester. Kaneshiro, K.Y. 1989. The dynamics of sexual se- lection and founder effects in species formation, Pp. 279-296. In Genetics, Speciation and the Founder Principle (L.V. Giddings, K.Y. Kaneshi- ro and W.W. Anderson, eds.). Oxford University Press, Oxford. Losos, J.B., T.R. Jackman, A. Larson, K. de Quei- roz & L. Rodriguez-Schettino. 1998. Contingen- cy and determinism in replicated adaptive radi- ations of island lizards. Science 279:2115-2118. Maddison, D. & W. Maddison. 2000. MacClade. Sinauer, Sunderland, Massachusetts. Mayr, E. 1963. Animal Species and Evolution. Belknap Press of Harvard University Press, Cambridge, Massachusetts, Mendelson, TC,, & K.L. Shaw. 2005. Sexual be- haviour: rapid speciation in an arthropod. Nature 433:375-376. Palumbi, S.R. 1996. Nucleic acids II: The poly- merase chain reaction, Pp. 205-247. In Molecu- lar Systematics (D. Hillis, C, Moritz and B. Ma- ble, eds.). Sinauer, Sunderland, Massachusetts. Pons, J. & R.G. Gillespie. 2003. Common origin of the satellite DNAs of the Hawaiian spiders of the genus Tetragnatha: evolutionary constraints on the length and nucleotide composition of the re- peats. Gene 313:169-177. Pons, J., & R.G. Gillespie. 2004. Evolution of sat- ellite DNAs in a radiation of endemic Hawaiian spiders: Does concerted evolution of highly re- petitive sequences reflect evolutionary history? Journal of Molecular Evolution 59:632-641. Posada, D. & K.A. Crandall. 1998. Modeltest: test- ing the model of DNA substitution. Bioinfor- matics 14:817-818. Price, J.P & D.A. Clague. 2002. How old is the Hawaiian biota? Geology and phylogeny suggest recent divergence. Proceedings of the Royal So- ciety B 269:2429-2435. Rees, D.J., B.C. Emerson, R Oromi & G.M. Hewitt. 2001. Reconciling gene trees with organism his- tory: the mtDNA phylogeography of three Ne- sotes species (Coleoptera: Tenebrionidae) on the western Canary Islands. Journal of Evolutionary Biology 14:139-147. Roderick, G.K. & R.G. Gillespie. 1998. Speciation and phylogeography of Hawaiian terrestrial ar- thropods. Molecular Ecology 7:519-531. Rokas, A., G. Melika, Y. Abe, J.L. Nieves- Aldrey, J.M. Cook & G.N. Stone, 2003. Lifecycle clo- sure, lineage sorting, and hybridization revealed in a phylogenetic analysis of European oak gall- wasps (Hymenoptera: Cynipidae: Cynipini) us- ing mitochondrial sequence data. Molecular Phy- logenetics and Evolution 26:36-45. Ruber, L., E. Verheyen & A. Meyer. 1999. Repli- cated evolution of trophic specializations in an endemic cichlid fish lineage from Lake Tangan- yika. Proceedings of the National Academy Sci- ence 96:10230-10235. Schluter, D. 1998. Ecological causes of speciation. Pp. 114-129. In Endless forms: Species and spe- ciation (D.J. Howard and S.H. Berlocher, eds.). Oxford University Press, Oxford, England. Schluter, D. 2000. Ecological character displace- ment in adaptive radiation. American Naturalist 156:S4-S16. Schluter, D. 2001. Ecology and the origin of spe- cies. Trends in Ecology & Evolution 16:372- 380. Schluter, D. & J.D. Mcphail. 1993. Character dis- placement and replicate adaptive radiation. Trends in Ecology & Evolution 8:197-200. Seehausen, O. 2004. Hybridization and adaptive ra- diation, Trends in Ecology & Evolution 19:198- 207. Shaw, K.L. 2002, Conflict between nuclear and mi- tochondrial DNA phylogenies of a recent species radiation: What mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proceedings of the National Academy of Scienc- es 99:16122-16127. Shaw, K.L., & D. Herlihy. 2000. Acoustic prefer- ence functions and song variability in the Ha- waiian cricket Laupala cerasina. Proceedings of the Royal Society of London B 267:577-584. Simon, C. 1987. Hawaiian evolutionary biology: An introduction. Trends in Ecology and Evolu- tion 2:175-178. Swofford, D.L. 2000. PAUP*. Phylogenetic Anal- ysis Using Parsimony (* and other methods). Sinauer, Sunderland, Massachusetts. Thorpe, R.S,, D.P. McGregor, A.M. Gumming & W. Jordan, 1994. DNA evolution and colonization sequence of island lizards in relation to geolog- ical history: mtDNA RFLP, cytochrome b, cy- tochrome oxidase, 12s rRNA sequence, and nu- clear RAPD analysis. Evolution 48:230-240. Thorpe, R.S., & A.G. Stenson. 2003. Phylogeny, paraphyly and ecological adaptation of the colour and pattern in the Anolis roquet complex on Mar- tinique. Molecular Ecology 12:117-132. Turelli, M., N.H. Barton & J.A. Coyne. 2001. The- ory and speciation. Trends in Ecology and Evo- lution 16:330-343. Wagner, W.L. & V. Funk. 1995. Hawaiian Bioge- ography: Evolution on a Hot Spot Archipelago. Smithsonian Institute Press, Washington, D.C. Manuscript received 19 January 2005, revised 17 June 2005, 2005. The Journal of Arachnology 33:323-333 DIVERSITY OF ARBOREAL SPIDERS IN PRIMARY AND DISTURBED TROPICAL FORESTS Andreas Floren: Dept, of Animal Ecology and Tropical Biology, University Wuerzburg, Biozeetrum Am Hubland, D-97074 Wuerzburg, Germany. E-mail: floren@biozentrum.um-wuerzburg.de Christa Deeleman-Reiehold: Sparrenlaan 8, 4641 GA Ossendrecht, The Netherlands ABSTRACT. This study investigates how arboreal spider communities in SE-Asian primary lowland rain forests change after anthropogenic disturbance. Two types of secondary forests were distinguished: 1) forests adjacent to each other, which finally merged into primary forest and 2) forests that were isolated by at least 10 km from the primary forest. Three forests of different age were investigated from each type and compared with undisturbed primary forest. All disturbed forests had been used some years for agri- culture and were then left between 5 and 50 years to regenerate naturally. Spiders from at least seven trees per forest type were collected using insecticidal knockdown fogging and sorted to species or mor- phospecies level. Spiders represented between 5-10% of all canopy arthropods. A similar number of spiders were collected per square meter from all trees. However, communities in the primary forest differed greatly in their alpha- and beta-diversity and in community structure from those in the disturbed forest types. Diversity was high in the regenerating forests connected to the primary forest and approximated the conditions of the primary forest during the course of forest succession. In contrast, the isolated forests were of low diversity and communities showed little change during forest regeneration. These results indicate the importance of a species-source from which disturbed forests can be recolonized. However, even under optimal conditions this process needed decades before spider communities became similar to those of the prim.ary forest. With no species-source available, spider diversity changed little during 50 years of forest regeneration. In the isolated forest we observed a drastic turnover from forest species towards species characteristic of open vegetation and shrubs. Our results give an indication of how large a loss in diversity can be expected in isolated forest fragments. Keywords^ Fogging, fragmentation, forest isolation, recolonization, species-source The canopy of tropical lowland rain forests forms a highly complex habitat. Here lives the most diverse arthropod fauna of the world, which influences many ecosystem processes and ecosystem services (examples in Linsee- mair et al. 2001; Basset et al. 2003). This as- sessment has been based on the faunistic-eco- logical analysis of taxa such as Coleoptera, Lepidoptera or Formicidae (e.g., Erwin 1983; Morse et al. 1988; Floren et al. 2001, 2002; Brehm et aL 2003; Davidson et al. 2003) while comparatively little work has been done on other groups. However, these latter groups can be rich in species and of great ecological importance, such as Araeeae (Hoefer et al. 1994; Deeleman-Reiehold 2001; Santos et al. 2003) which, next to Formicidae, are the most abundant predators in the trees (Stork 1991; Floren & Linseemair 1997; Wagner 1997). Despite political declarations, tropical forests are recklessly destroyed and reduced to forest fragments which are much simpler in species diversity and habitat complexity. Although this destruction will certainly change many ecosystem properties, the consequences of this transformation have never been adequately in- vestigated. This study aims at providing such knowledge. We investigated the diversity and stracture of arboreal spider communities in SE-Asian primary forests. Furthermore, we analyzed how communities differ in disturbed forest types and how they reorganized follow- ing anthropogenic disturbance. We collected arboreal spiders by insecticidal knockdown fogging. Besides primary forest, we studied 1) three secondary forests of different ages that merged into each other and finally into pri- mary forest and 2) three isolated secondary forests of different ages which were separated by at least 10 km from the primary forest. This study design allowed us to assess the im- portance of species recolonization for the re- 323 324 THE JOURNAL OF ARACHNOLOGY Figure 1. — Map of study sites. SI, SII, Sill were adjacent forests of 5, 15, and 40 years age that merged into primary forest at Kinabalu National Park substation Sorinsim. CRl, CRII, CRIII, were isolated forest plots of 10, 20, and 50 years age that were at least ten kilometers away from the primary forest. P = primary forest plots. organization of spider communities during forest regeneration. After eight years of tax- onomic analysis by the senior author we are now able to present the results of our inves- tigation. METHODS Study sites* — Arboreal arthropods were collected by insecticidal knockdown fogging in a Dipterocarp lowland rain forests of Kin- abalu National Park (500 — 650 meters a.s.l.) in Sabah, Malaysia on Borneo (6° 2.75 'N, 116° 42.2'E) during various field periods from 1992 — 2001 (Table 1). The area has a relative- ly constant climate with a main rainy season from November to February and a shorter one from June to July. The level of precipitation varies between 2000 and 4000 mm. In total, 15 trees of the genus Aporusa (Euphorbi- aceae) were fogged, 6 trees of Xanthophyllum affine (Polygalaceae) and 9 trees of various other genera (for details see Horstmann et ak in press). The disturbed forests were situated at substation Sorinsim in Kinabalu National Park and in the vicinity of the Crocker Range National Park. A map of all forest types is shown in Fig. 1. Details on the study sites are published elsewhere (Floren et al. 2001; Horstmann et al. in press). All secondary for- ests were clear-cut for crop planting, aban- doned and left for natural regeneration. Three forests of 5, 15 and 40 years, each of 5 — 6 ha (abbreviated SI, SII, Sill), which merged into one another and finally into primary forest, were investigated at National Park substation Sorinsim, Foggings were carried out from February — -March 1997. Three isolated forest plots of 10, 20 and 50 years were found within at least 10 km distance from the primary for- est of the Crocker Range National Park (ab- breviated CRI, CRII, CRIII). They were about 4 — 6 hectares in size and surrounded by cul- tivated land (fruit, oil palm, rubber planta- tions, pastures, etc.). Fieldwork was carried out between January and February 2001. All disturbed forests had only a single canopy lay- er which was in no case closed and differed both in tree height and girth at breast height of the study trees. Collecting methods* — A full description of the fogging method is given in Adis et al. (1998). Natural pyrethrum was used as an in- secticide and all arthropods that dropped into the collecting funnels two hours following fogging were used in the analysis. In order to collect arboreal arthropods as completely as possible, 80 — 90% of a crown projection area was covered with collecting funnels installed beneath a tree. In total, 102 foggings were car- ried out, consisting of the first and subsequent foggings (mostly on consecutive days). Fau- nistic analysis is based on all these foggings while only the first foggings were used for community level analysis (Table 1), Seven primary forest trees were re-fogged after three years and two trees after an eight month pe- riod. Spider communities from these samples could not be distinguished from those of the first foggings and were, therefore, considered independent samples. As no tree species grew in all forests, a common tree was fogged in each forest type. However, as tree specific as- sociations of broad-leaved trees are thought to be of minor importance for spiders and were also not indicated by our results, we refer to Floren & Linsenmair (2001) for the general discussion of this aspect. Analysis is based on FLOREN & DEELEMAN-REINHOLD— SPIDER DIVERSITY IN MALAYSIA 325 Table 1. — Forests investigated and focal trees. Individual trees were refogged several times on consec- utive days. SI, SII, SIII = secondary forests connected with primary forests; CRI, CRII, CRIII = isolated secondary forests. Focal tree species Number of foggings Tree . height (m) Girth in breast height (cm) Fog 1 Re-fog Primary forest Aporusa lagenocarpa 27 3 24-30 70.24 ± 18.12 A. subcaudata (Euphorbiaceae) SI (5 yrs.) Meiochia umbeiiata (Sterculiaceae) 8 10 6-8 57.91 ± 9.46 SII (15 yrs.) Vitex pinnata (Verbenaceae) 11 4 18-20 106.44 ± 14.54 SIII (40 yrs.) V. pinnata 10 5 20-25 148.44 ± 48.14 CRI (10 yrs.) Melanolepis glandulosa (Euphorbiaceae) 8 — 6-8 83.16 ± 9.89 CRII (20 yrs.) M. glandulosa 7 — 18-20 107.43 ± 14.54 CRIII (50 yrs.) M. glandulosa 9 — 18-25 122.89 ± 23.20 adult spiders, which are stored in the collec- tion of C. Deeleman. Data analysis, — Spider com.munities in forest types were compared using alpha- and beta diversity indices (Magurrae 1988). Wil- liam’s alpha (after Fisher et al. 1943) is a widely used parametric index of diversity, which is largely independent of sample size. Simpson’s index describes the probability that a second individual drawn from a population should be of the same species as the first. It is mainly influenced by common species and therefore a measure of equitability (the larger the value the greater the equitability). Sample sizes were standardized by using rarefaction statistics (Hurlbert 1971; Hayek & Buzas 1997). For this purpose, spiders of all fogged trees per forest type were pooled and diversity was expressed as the number of expected spe- cies within an equal sub-sample size (this cor- responded with the 65 species identified from all 306 specimens in the isolated forest CRI). If rarefaction values are computed for increas- ing sub-samples and plotted graphically, the resulting curve can be interpreted as a species accumulation curve, which gives information on the structure of spider communities in each forest type (Achtziger et al. 1992). Shinozaki curves were calculated to compare commu- nities on the beta-diversity level (Shinozaki 1963; Achtziger et al. 1992). They are ex- pected species accumulation curves based on qualitative (presence / absence) data of spe- cies. Their steepness provides information about the overall completeness of the sam- pling effort. Furthermore, Soereesen’s quan- titative index of similarity was calculated. Dif- ferences in means of beta-diversity between forest types were tested with a Mantel test us- ing a randomization test (Monte Carlo), The number of randomized runs was 1000. For a between forest comparison, the fogging data were standardized for a crown projection of Im^ and a leaf cover of 100%. RESULTS From all 102 foggings, 6999 spiders were collected and sorted to 578 species in 29 fam- ilies (Appendix 1). Scientific names were found for 107 species of which 75 species (12.9%) were new for Borneo. The five most abundant families, declining in rank-order, were Theridiidae, Salticidae, Araneidae, Thomisidae, and Clubionidae, together repre- senting between 73% and 94% of all spiders in each forest. These families contributed also between 75% and 84% of all species. Theri- diidae represented 153 species, Salticidae 111 species, Araneidae 80 species, Thomisidae 74 species, and Clubionidae 31 species. Spiders provided on average between 4.6% and 9.8% of all arthropods in a community (Table 2). Differences in the relative proportion of spi- ders per tree were detected only between the youngest isolated forest CRI and the primary forest, CRI and SII, and CRI and SIII (AN- OVA, F = 4.235, df = 6, F < 0.01, Tamhane post-hoc tests for unequal variances were car- ried out, P < 0.05). The number of collected spider individuals, standardized on Im^ col- lecting sheets and 100% leaf cover, differed not significantly between tree species or forest types, only between the primary forest and SII where spider numbers were lowest (ANOVA, 326 THE JOURNAL OF ARACHNOLOGY C/) 0 ■q 0 Q. 0 ■0 0 *0 0 Q. X LU 150 H 250 n 200 - 100 T 1 1 1 1 0 200 400 600 800 1000 Size of subsample Figure 2. — Rarefaction curves of spider communities based on all foggings. 1200 Sill (40 yrs.) PRIMARY FOREST Si (5 yrs.) CRH (20 yrs,) CRIII (50 yrs.) CRI(10yrs.) F = 2.358, P < 0.05, Tainhane post-hoc test, df = 6, P < 0.05). Most abundant in the for- ests close to the primary forest was Talaus nanus (Thorell 1892) (Thomisidae) with 297 individuals, followed by Ogulnius sp. (Theri- diosomatidae) with 103 individuals, and Mo- lione kinabalu (Yoshida 2003) (Theridiidae) with 74 individuals. The isolated forests were numerically dominated by Ocyllus sp. (Thom- isidae) with 75 individuals followed by Te- tragnatha hasselti (Thorell 1890) (Tetragnath- idae) with 51 specimens. Family diversity depended on sample size and was highest in the primary forest. Rarefaction statistics allow comparison of forest types that have been sampled with different efficiency. On a rare- fied sub-sample of 306 individuals (corre- sponding to the number of spiders of the smallest sample CRI) a similar number of families, namely 22, were observed in SII and the primary forest. The number of spider fam- ilies was least in the isolated forest plots. Rar- ified species numbers and, correspondingly, William's alpha were highest in the primary forest and in SIIL Again these indices were clearly lower in the isolated forests where al- pha-diversity had changed little even after 50 years compared to the 'Sorinsim-forests'. Rar- ified species numbers were higher both in SI and SII than in CRI and CRII (related to the gradient forests these were 30.9% and 33.0%, respectively), and 41,7% more species were collected in Sill compared to CRIII. An ap- proximate value for the loss of species follow- ing anthropogenic disturbance is the relation of species numbers to primary forest species numbers (Floren & Linsenmair 2005). Only 22.0% of the primary forest species number was collected in the isolated forest Cl, the most severely disturbed forest with the lowest number of species. Relative proportion of sin- gletons in each forest type was lowest in the primary forest (32,4%) and increased in the disturbed forests. Despite the large sampling effort in the primary forest, there were still 96 species represented by only one individual. In contrast to the proportion of singletons per forest type, the mean proportion of singletons of all tree specific communities per forest type was a better discriminator between primary and disturbed forests despite high variance be- tween tree specific communities. The average proportion of singletons was highest and not significantly different in the primary and the FLOREN & DEELEMAN-REINHOLD— SPIDER DIVERSITY IN MALAYSIA 327 old secondary forest Sill (Table 2). The pri- mary forest (mean 25.8 ± 1 1.8) differed from the isolated forests CRI (mean 10.3 ± SD 4.2) and CRIII (mean 16.2 ± SD 3.6) (ANOVA, F = 7.647, df = 6, P < 0.001, Tamhane post- hoc test, P < 0.001 and P < 0.01, respective- ly). In all other forests, spider communities differed not significantly in respect to the pro- portion of singletons per community. Equita- bility, as measured by Simpson’s index, was least in the disturbed forests and most in SII and SIIL Due to the numerical dominance of an individual species, unknown genus cf. Pyc- naxis sp. (Thomisidae), which occurred on 13 out of 27 trees with maximum 71 and 86 in- dividuals per tree, the primary forest evenness was lower. However, excluding this species from the analysis resulted in an index of 84.0, confirming high evenness for all other species. Rarefaction curves did not level off with increasing size of sub-samples (Fig. 2). How- ever, in contrast to the isolated forests, the curves were much steeper in the primary and the connected Sorinsim-forests indicating that spider communities were not collected repre- sentatively. Also Fig. 2 shows the clear sep- aration between forest types, indicating that spider communities recovered much faster in the Soriesim forests, which were adjacent to the primary forest, than in the isolated forests (see also Table 2). The increase of the rare- faction curve of SII indicates that the rate of species collection was similar to that of the primary forest. Prominent was the high spe- cies diversity of the 40 year-old forest SIIL Figure 3 shows the species frequency distri- bution of all spiders from all pooled foggings. Increasing the sample size always resulted in many new species indicating that the regional species pool was not sampled representatively by the 80 first foggings. Computing Shino- zaki-curves for the four largest families, how- ever, showed that they were collected reliably by fogging. Comparing mean similarities of tree-specif- ic spider communities (expressed by the So- erensen index, Fig. 4) showed clear differenc- es between forest types (Mantel-test, Monte Carlo randomization, z = —0.583362, P < 0.001). In the primary forest and also in SI, SII, and Sill, 70% to 80% of all species were found only on one tree. In contrast, tree spe- cific spider communities in the isolated forests shared many more species and consequently communities showed a significantly higher overlap in species. Most spiders were found only in one forest type: 155 species (52%) of primary forest spe- cies were restricted to the primary forest, 149 species (48%) and 62 species (38%) respec- tively were only found in the adjacent and the isolated forests. Changes in spider communi- ties came along with drastic faunistic changes. For example, 96 widespread ubiquitous spe- cies (species distributed in the Malay Archi- pelago) were identified. Their proportion was highest in the isolated forests representing 56 of all 160 species (35%), 69 ubiquitous spe- cies (21.9%) were collected in the primary forest and 63 species (18.9%) in the connected ‘Sorinsim’ forests (Appendix 1). Most of the ubiquitous species were Araneidae, Theridi- idae and Tetragnathidae and could be identi- fied to the species level, like Neoscona vigi- lans (Blackwall 1865), N. punctigera (Doleschall 1857), common Cyclosa and Gas- teracantha species, Chrysso spiniventris (O.P- Cambridge 1869), Takayus lyricus (Walcken- aer 1842), Tetragnatha hasselti (Thorell 1890) and Mesida gemmea (van Hasselt 1882) (Plat- nick 2005; Yin et al. 1997; Zhu 1998; Zhu et al. 2003; Yoshida 2003). DISCUSSION Anthropogenic destruction of tropical rain forests makes it necessary to assess the im- mediate and the long-term consequences for man and nature. Only on the basis of such knowledge is a sound nature protection plan possible. This, however, requires a high effort of basic research because even the extent of species richness is not known for most taxa (Basset et al. 2003). In this paper we present such a basic study for arboreal spiders, which we collected by pyrethrum knockdown fog- ging in primary and secondary lowland rain forests of Sabah, Malaysia on Borneo. Next to Formicidae, spiders are the most abundant group of predators in tropical lowland forest canopies (Adis et al. 1984; Stork 1991; Floren & Linsenmair 1997, 2001). Our study con- firmed high species diversity of arboreal spi- ders. Despite a total of 102 foggings, the re- gional species pool was not sampled representatively and new species are still be- ing found in new samples (Deeleman pers. obs.). There is a need to extend investigations, including further yet unsampled habitats, and 328 THE JOURNAL OF ARACHNOLOGY Table 2. — ^Comparison of spider communities between forest types. Analysis is based on first foggings only. Means are given with standard deviations. * = Data are standardized for a crown projection of Im^ and a leaf cover of 100%. SI, SII, Sill = secondary forests connected with primary forests; CRI, CRII, CRIII = isolated secondary forests. Prim, forest SI 5 yrs. SII 15 yrs. Sill 40 yrs. CRI 10 yrs. CRII 20 yrs. CRIII 50 yrs. Mean rel. 5.6 7.2 4.6 5.9 9.8 6.3 6.4 prop, of spi- ders per for- est No. of families 28 15 24 24 11 15 19 Rarefied no. of 21.6 14.5 22.2 20.5 11 14.1 16.2 families (m = 306) No. of species 296 120 127 230 65 97 102 Rarefied no. of 122 94 1 15 132 65 77 77 species (m = 306) William’s al- 87.5 48.6 67.1 91.0 25.3 35.4 34.6 pha Total number 2488 525 365 1048 306 523 625 of spiders collected Standardized 19.6 ± 15.63 13.0 ± 7.8 6.4 ± 2.9 14.8 ± 5.3 15.8 ± 6.2 1 1.7 ± 8.1 11.6 ± 6.6 mean abun- dance* Singletons 96 (32.4%) 46 (38.3%) 61 (48.0%) 86 (37.4%) 27 (41.5%) 40 (41.2%) 35 (34.3%) Mean propor- 25.8 ± 1 1.8 17.4 ± 7.9 17.5 ± 6.3 33.8 ± 10.3 10.3 ± 4.2 17.6 ± 6.9 16.2 ± 3.6 tion of sin- gletons of all trees Simpson-index 30.6 20.3 41.3 44.1 16.3 18.0 22.6 compare spider diversity with that reported in the few studies that have been carried out so far in the region (Russell-Smith and Stork 1994, 1995; Deeleman-Reinhold 2001) in or- der to assess the extent of diversity and to investigate the role spiders play in ecosystem functioning (New 1999). Primary forests differ conspicuously from disturbed forests in habitat complexity. As a consequence, the diversity, structure, and dy- namics of arthropod communities also change in disturbed forests (Floren et al. 2001). This was also confirmed, convincingly, for arboreal spiders. Using the primary forest as a basis, we investigated how spider communities changed in various secondary forests of dif- ferent ages; that is to say in forests of different disturbance levels. Our data do not allow us to perform a full community level analysis be- cause local species pools have not been sam- pled representatively and differences between communities might simply be due to collect- ing new species. However, we can compare data after standardization, e.g. by using rare- faction statistics, comparing relative propor- tions of a parameter or by looking for changes in community structure and faunistic compo- sition. On the basis of such comparisons, pri- mary forests are clearly distinguishable from the adjacent secondary forests (SI, SII, Sill) merging into the primary forest, which are, themselves, clearly separated from the isolat- ed forests (CRI, CRII, CRIII). As demonstrat- ed by our data, the comparatively small dis- tance of 10 km to the primary forest forms an effective barrier preventing species recoloni- zation when the surroundings are cultivated land. Spider density was similar in all forests in- dicating that the number of spiders collected by fogging did not depend on the tree species or the level of disturbance of the secondary FLOREN & DEELEMAN-REINHOLD— SPIDER DIVERSITY IN MALAYSIA 329 C/5 ■q 0 CL CO TD 0 •*—» O 0 Q- X UJ 600 n 500 400 300 200 100 0 Thonnisidae All spiders Theridiidae Salticidae Araneidae Number of loggings Figure 3. — Shinozaki curves of spider communities based on all foggings. forest. Changes in spider communities oc- curred already at the family level (number of families collected per forest type) and were recognizable especially by the dominance of species from the families Theridiidae, Thom- isidae, Salticidae, Araneidae, and Clubioni- dae. Dominance of individual species was highest in the most disturbed forests. Above all, high species diversity in SII and Sill in- dicate that the spider fauna recovered much faster in the forests close to the primary forest than in the isolated forests. An approximation to the conditions of the primary forest during the course of forest succession is obvious in most parameters analyzed and is in correspon- dence with similar findings for Formicidae and Coleoptera (Floren et al. 2001; Floren & Linsenmair 2001). In contrast, diversity in the isolated forests was significantly lower and changed only a little during forest regenera- tion. Species numbers give an impressive ex- ample: even in the 5 year-old pioneer forest SI, we found more species than in the 50 year- old isolated forest CRIIL Interestingly, the 40 year-old forest SIII was richer in species than the primary forest. A probable reason for this is that many primary forest species had al- ready become established in SIII and were able to coexist with species that were more successful under the disturbance regime. Although spider communities of the con- nected forests SII and SIII resembled those of the primary forest in many respects, there were still clear differences. While the propor- tion of singletons was larger than 30% in each forest type and did not correlate with the de- gree of disturbance, the mean number of sin- gletons per tree community distinguished the primary and the old secondary forest SIII from all other disturbed forests. The number of singletons per tree-specific community was lowest in CRI, the youngest isolated and most disturbed forest fragment investigated. The low proportion of singletons per community corresponded with low overall diversity in the disturbed forest fragments and seems to be a good discriminator between primary, old-sec- ondary and more severely disturbed forests. Community equitability also changed with forest disturbance from even communities in the primary forest to uneven communities in the disturbed forests. Similar changes are usu- ally observed as a consequence of anthropo- genically disturbance of forests (e.g., Leigh et al. 1993; Laurance 1994; Daily & Ehrlich 1995). The reason for the high abundance of 330 THE JOURNAL OF ARACHNOLOGY 0,70 0,60 0,50 C 0) m 0,40 c 0 g 0,30 (f) 0,20 0,10 0,00 Figure 4. — Mean beta-diversities (measured by the Soerensen-index) between all spider communities of each forest type, expressed as Box Plots. The boxes cover 50 percent of all values (wiskers 75%) and show the median. A circle indicates outlier values between one and three times the box length. O Primary Sll CRI CRII! SI SIM CRII the thomisid new genus & species cF Pyc- naxis, a species which seems unrelated to any other species and which has been found ex- clusively in the primary forests in Sabah, is not currently understood. It might be con- nected to the El Nino droughts of the year of collection in 1998. Spider communities be- came structurally simpler in the disturbed for- ests, because fewer species were found with median abundance classes. In the isolated for- ests we found a dominance of a number of common widespread web-building spider spe- cies: for instance several Gasteracantha and Tetragnatha species, Mesida gemmea (Hasselt 1882), and a number of smaller theridiid spe- cies. Several tiny (2 — 3 mm) widespread oon- opid and theridiid species were found exclu- sively in the primary forests; these species probably live among the roots of epiphytic plants. Our results led us to conclude that recolo- nization from primary forests is absolutely necessary for the restoration of species diver- sity. If such species-sources are lacking, the restoration of spider diversity and spider com- munities proceeds only slowly if at all. These data indicate that the time necessary for re- covery of arthropod diversity is usually great- ly underestimated. The process of recoloni- zation needs decades even under optimal conditions. In contrast, we sampled only ru- dimentary spider communities in the isolated secondary forest stands where no recoloniza- tion occurred. Even after 50 years of forest regeneration, spider communities were of low diversity and dominated by common species characteristic of open vegetation and shrub. Today small forest fragments dominate the landscape and already 40 year old forests are under high pressure by local people and the wood industry. Our study led us to suspect that the loss of spider species diversity will be FLOREN & DEELEMAN-REINHOLD— SPIDER DIVERSITY IN MALAYSIA 331 immense with primary forests lacking as spe- cies-sources from which recolonization can start. ACKNOWLEDGMENTS We thank the director of Sabah Parks, Da- tuk Ali Lamri, for permission to work in the Kinabalu Park. We are very much indebted to Jameli Nais and Alim Biun for generous sup- port, Andre Kessler and Stefan Otto helped in the field work. Helmut Stumpf, Domir De Bakker, Daiqin Li and one anonymous referee gave valuable comments on the manuscript. John Murphy kindly checked the English. Fi- nancial support for this study came from the German Science Foundation (DFG), Li 150/ 13-4. LITERATURE CITED Achtziger, R., U. Nigmann & H. Zwolfer. 1992. Rarefaction-MethO'den und ihre Einsatzmogli- chkeiten bei der zoookologischen Zustandsana- lyse und Bewertung voe Biotopen. Zeitschrift fiir Okologie und Naturschutz 1:89-105. Adis, J., Y. Basset, A. Floren, P.M. Hammond & K.E. Linseemair. 1998. Canopy fogging of an overstorey tree — Recommendations for standard- ization, Ecotropica 4:93-97. Adis, J., Y.D. Lubin & G.G. Montgomery. 1984. Arthropods from the canopy of inundated and terra firme forests near Manaus, Brazil, with crit- ical considerations on the pyrethrum-fogging technique. Studies on Neotropical Fauna and En- vironment 4:223-236. Basset, Y, V. Novotny, S. Miller & R. Kitchieg. 2003. Arthropods of Tropical Forests: Spatio- Temporal Dynamics and Resource Use in the Canopy. Cambridge University Press, Cam- bridge. 474 pp Brehm, G., J. Homeier & K. Fiedler. 2003. Beta diversity of geometrid moths (Lepidoptera: Geo- metridae) in an Andean montane rainforest. Di- versity and Distributions 9:351-366. Daily, G.C. & RR, Ehrlich. 1995. Preservation of biodiversity in small rainforest patches: rapid evaluations using butterfly trapping. Biodiversity and Conservation 4:35-55. Davidson, D.W., S.C. Cook, R.R. Snelling. & TH. Chua. 2003. Explaining the abundance of ants in lowland tropical rainforest canopies. Science 300:969-972. Deeleman-Reinhold, C.L. 2001. Forest spiders of South East Asia. Brill, Leiden. 591 pp, Erwin, TL. 1983. Tropical forest canopies: the last biotic frontier. Bulletin of the Entomological So- ciety of America 30:14-19. Fisher, R.A., A. Steven Corbet & C.B. Williams. 1943. The relation between the number of spe- cies and the number of individuals in a random sample of an animal population. Journal of An- imal Ecology 12:42-58. Floren, A., A. Biun & K.E. Linsenmair. 2002. Ants as key predators in tropical lowland rain forests. Oecologia 131:137-144. Floren, A., A. Freking, M. Biehl & K.E. Linsen- mair. 2001. Anthropogenic disturbance changes the structure of arboreal tropical ant communi- ties. Ecography 24:547-554. Floren, A. & K.E. Linsenmair. 1997. Diversity and recolonisation dynamics of selected arthropod groups on different tree species in a lowland rain forest in Sabah, Malaysia with special reference to Formicidae. Pp. 344-381. In N. E. Stork, J. A. Adis & R. K. Didham, editors. Canopy Ar- thropods. Chapman & Hall, London. Floren, A. & K.E. Linsenmair. 2001. The influence of anthropogenic disturbances on the structure of arboreal arthropod communities. Plant Ecology 153:153-167. Floren, A. & K.E. Linsenmair. 2005. The impor- tance of primary tropical rain forest for species diversity: an investigation using arboreal ants as an example. Ecosystems. Hayek, L.A.C. & M.A. Buzas. 1997. Surveying nat- ural populations. Columbia University Press. Hoefer, H., A.D. Brescovit, J. Adis & W. Paarmann. 1994. The spider fauna of Neotropical tree can- opies in Central Amazonia: first results. Studies on Neotropical Fauna and Environment 29:23- 32. Horstmann, K,, A. Floren & K.E. Linsenmair. In press. Ichneumonidae (Hymenoptera) from the canopy of tropical forests in Sabah, Malaysia: a comparison between primary and secondary for- ests. Ecotropica. Hurlbert, S.H. 1971. The non-concept of species di- versity: a critique and alternative parameters. Ecology 52:577-586. Laurance, W.F. 1994. Rainforest fragmentation and the structure of small mammal communities in tropical Queensland. Biological Conservation 69: 23-32. Leigh, E.G., S.J. Wright, E.A. Herre & F. E. Putz. 1993, The decline of tree diversity on newly iso- lated tropical islands: a test of a null hypothesis and some implications. Evolutionary Ecology 7: 76-102. Linsenmair, K.E., A.J. Davis, B. Fiala & M.R. Speight. 2001. Tropical Forest Canopies: Ecolo- gy and Management. Kluwer Academic Publish- ers, Dordrecht. 370 pp. Magurran, A.E. 1988. Ecological Diversity and its Measurement. Princeton. 179 pp. Morse, D.R., N.E. Stork & J.H. Lawton. 1988. Spe- cies number, species abundance, and body length relationships of arboreal beetles in Bornean low- 332 THE JOURNAL OF ARACHNOLOGY land rain forest trees. Ecological Entomology 13: 25-37. New, T.R. 1999. Untangling the web: spiders and the challenges of invertebrate conservation. Jour- nal of Insect Conservation 3:251-256, Platnick, N.I. 2005. The world Spider Catalog, ver- sion 5.0. American Museum of Natural History, online at http://research.amnh.org/entomology/ spider/catalogS 1-87/index. html. Russell-Smith, A. & N.E. Stork. 1994. Abundance and diversity of spiders from the canopy of trop- ical rainforests with particular reference to Su- lawesi, Indonesia. Journal of Tropical Ecology 10:545-558. Russell-Smith, A. & N.E. Stork 1995. Composition of spider communities in the canopies of rain- forest trees in Borneo. Journal of Tropical Ecol- ogy 11:223-235. Santos, G.B., M.I. Marques, J. Adis & C.R. De Mu- sis. 2003. Arthropodos associados a copa de At- talea phalerata Mart. (Arecaceae), na regiao do Pantanal de Pocone, Mato Grosso, Brasil. Revis- ta Brasileira de Entomologia 47:21 1-221. Shinozaki, K. 1963. Note on the species-area curve. Proceedings of the first annual meeting of the ecological society of Japan 5. Tokyo, p. 5. Stork, N.E. 1991. The composition of the arthropod fauna of Bornean lowland rain forest trees. Jour- nal of Tropical Ecology 7:161-180. Wagner, T. 1997. The beetle fauna of different tree species in forests of Rwanda and east Zaire. Pp. 169-183: In N. E. Stork, J. Adis & R. K. Did- ham, editors. Canopy Arthropods. Chapman & Hall, London. Yin, C.M., J.E Wang, M.S. Zhu, L.P. Xie, X.J. Peng & Y.H. Bao. 1997. Fauna Sinica: Arachnida: Ar- aneae: Araneidae. Science Press, Bejing, xiii -h 460pp. Yoshida, H. 2003. The Spider Family Theridiidae (Arachnida: Araneae) from Japan. Arachnologi- cal Society of Japan, 224 pp. Zhu, M.S. 1998. Fauna Sinica: Arachnida: Araneae: Theridiidae. Science Press, Bejing, xi -!- 436pp. Zhu, M.S., D.X. Song & J.X. Zhang. 2003. Fauna Sinica: Invertebrata Vol.35: Arachnida: Araneae: Tetragnathidae. Science Press, Bejing, vii -h 418pp. Manuscript received 1 February 2005, revised 17 September 2005. FLOREN & DEELEMAN-REINHOLD— SPIDER DIVERSITY IN MALAYSIA 333 Appendix 1. — Number of species per family and number of widespread species in all fogging samples. Sampling area Primary forest Secondary forests close to prim, forest Secondary isolated forests Total No. of foggings 30 48 24 102 No. of families 26 23 20 29 Ind. Sp. Ind. Sp. Ind. Sp. Total sp. Oonopidae 150 6 47 5 75 5 7 Pholcidae 147 10 3 2 7 2 11 Scytodidae 6 1 0 0 1 1 1 Clubionidae Clubioninae 230 17 54 15 21 4 24 Systariinae 18 3 0 0 0 0 3 Eutichurinae 14 2 12 2 5 1 4 Corinnidae Castianeirinae 40 9 64 7 10 3 10 Trachelinae 34 3 31 2 0 0 4 Phrurolithinae 2 1 17 1 2 1 3 Gnaphosidae 8 3 11 4 0 0 5 Sparassidae 71 6 59 7 42 3 10 Ctenidae 2 1 1 1 0 0 1 Selenopidae 3 1 0 0 0 0 1 Salticidae 334 59 426 65 72 19 111 Zodariidae 9 2 11 3 0 0 3 Oxyopidae 41 6 45 8 0 0 8 Pisauridae 0 0 21 1 1 1 1 Thomisidae 570 31 767 54 223 19 74 Philodromidae 15 2 5 2 12 1 2 Hahniidae 21 1 0 0 0 0 1 Hersiliidae 65 6 28 3 14 1 6 Linyphiidae 47 6 33 4 4 2 8 Theridiidae 562 80 600 83 631 53 153 Mimetidae 24 2 2 1 0 0 2 Theridiosomatidae 1 1 109 5 7 2 6 Tetragnathidae 98 10 117 10 171 9 19 Araneidae 151 36 307 44 97 29 80 Mysmenidae 0 0 10 6 2 1 7 Anapidae 0 0 5 2 1 1 3 Uloboridae 32 4 41 3 32 1 5 Dictynidae 43 1 0 0 0 0 1 Psechridae 2 2 0 0 1 1 2 Deinopidae 2 2 0 0 0 0 2 Total species 314 332 160 578 Identified widespread species 69 63 56 Percentage widespread species 21.9% 18.9% 35.0% 2005. The Journal of Arachnology 33:334-346 GENDER SPECIFIC DIFFERENCES IN ACTIVITY AND HOME RANGE REFLECT MORPHOLOGICAL DIMORPHISM IN WOLF SPIDERS (ARANEAE, LYCOSIDAE) Volker W. Frameeau^: Department of Zoology, The University of Melbourne, Parkville, Victoria, 3010, Australia. E-mail: volker.framenau@museum.wa.gov.au ABSTRACT. Sexual dimorphism of locomotory organs appears to be common in a variety of arthro- pods, however, the underlying evolutionary mechanisms remain poorly understood and may be the con- sequence of natural or sexual selection, or a combination of both, I analyzed the activity pattern of seven cohorts of a wolf spider, Venatrix lapidosa, over four consecutive years. Males appear to be the more active sex in search for a mate as they show temporarily higher activity prior to the periods of female brood care. Morphometric data on leg length showed comparatively longer legs for males than females. Allometric leg elongation in all four legs of males arises only after the final molt suggesting its significance in reproductive behavior such as mate search. A comparative analysis of two Australasian wolf spider genera with different activity profile of females, Venatrix (sedentary females) and Arfor/n (vagrant females) provides further evidence that limb elongation in males mainly arises due to indirect male mate compe- tition. Keywords: Sexual dimoiphism, locomotion, leg length, mark and recapture, minimum convex polygon Sexual dimorphism is thought to have evolved through sexual selection, ecological niche partitioning, differences in reproductive roles or a combination of these factors (e.g., Selander 1972; Hedrick & Temeles 1989; Shine 1989; Reynolds & Harvey 1994; Fair- bairn 1997). Sexual selection arises through competition between members of one sex for reproduction with the other sex (Andersson 1994). Ecological niche partitioning may re- sult in sexual dimorphism if each sex develops different structures as adaptations to different resources (Shine 1989; Walker & Rypstra 2001). Different reproductive success primar- ily arises through a fecundity advantage of large body size in females and is particularly evident in insects and spiders in which a com- mon finding is that, throughout a wide range of sizes, female fecundity varies directly with mass (e.g., Head 1995; Prenter et al. 1999). Selection for early maturation of males (pro- tandry) may also favor smaller male body size and thus result in sexual dimorphism (Bulmer 1983; Gunnarsson & Johnsson 1990). These explanations are not mutually exclusive and * Current address; Department of Terrestrial In- vertebrates, Western Australian Museum, Locked Bag 49, Welshpool DC, Western Australia 6986, Australia thus sexual dimorphism could evolve in a spe- cies through both sexual and natural selection. Therefore, it is often difficult to determine what mix of influences has resulted in sexual dimorphism in a particular species (Hedrick & Temeles 1989). The difficulty of identifying selective pres- sures is especially evident in the sexual di- morphism of locomotory structures, like wings or legs, which is a common phenome- non in many arthropods (Montgomery 1910; Thornhill & Alcock 1983). The evolution of gender specific differences in locomotory or- gans may be favored by both selection on male mate searching behavior and natural se- lection on female movements in relation to foraging or oviposition. Therefore, sexual di- morphism of locomotory structures has gen- erally not been considered in studies of sexual selection (Darwin 1871; Andersson 1994). Gender specific differences in locomotory structures have usually been attributed to a more active behavior of one sex, typically males, in search for mates (Thornhill & Al- cock 1983; Gasnier et al. 2002). Higher mo- bility may increase encounter rates of males with females and therefore increase fertiliza- tion success. However, gender specific elon- gation of limbs, even if under the influence of 334 framenau-activity and sexual dimorphism in wolf spiders 335 sexual selection, may not indicate an advan- tage in locomotion. Male elongated legs are important for direct male competition for ma- tes in water striders (Tseng & Rowe 1 999) and megalopodine beetles (Eberhard & Marin 1996), in grasping females during mating in mayflies or calanoid copepods (Peters & Campbell 1991; Ohtsuka & Huys 2001), in newt courtship displays (Malmgren & Thol- lesson 2001) and to reduce the risk of sexual cannibalism in some orb-web spiders (Elgar et al. 1990). In cursorial spiders, elongated segments of legs, in particular the first pair, have also been reported in combination with ornamentations in species with visual court- ship display (Kronestedt 1990; Hebets & Uetz 2000). Therefore, it is vital to correlate activ- ity and mobility patterns with sexual dimor- phism of leg length to provide evidence of sexual selection acting on locomotion itself. Sexual dimorphism, in spiders has been studied extensively, however, the evolution of sexual size dimorphism remains controversial (e,g,, Elgar 1991; Vollrath & Parker 1992; Head 1995; Hormiga et al. 1995). There are two main explanations for patterns of sexual size dimorphism in spiders (see Elgar 1998). Firstly, fecundity selection may favor larger females (Preeter et al. 1997, 1998, 1999). Al- ternatively, Vollrath & Parker (1992) suggest that sexual dimorphism may arise from dif- ferences in male and female lifestyles. In spe- cies with sedentary females, an increase in male mortality through mate searching behav- ior relaxes selection for large male body size and thus selection for protandry will favor smaller males. Ground living spiders are gen- erally less size dimorphic than web-building species, which has been explained by their differing reproductive and foraging strategies (Enders 1976; Prenter et al. 1999). There is some evidence for sexual dimorphism in lo- comotory structures in ground living spiders (e.g., Gasnier et al. 2002). Montgomery (1910) reported that males have relatively lon- ger legs than females, which he suggested is a result of the nomadic behavior of males after attaining sexual maturity. This idea is sup- ported by a number of short term studies on the locomotor y activity of wolf spiders, in which males were the more active sex (e.g., Hallander 1967; Richter et al. 1971; Cady 1984; Frameeau et al. 1996a). However, wolf spiders differ in activity profiles due to vary- ing life strategies that range from permanently burrowing (e.g., Geolycosa or Lycosa s. str.), to permanently vagrant animals (e.g., Pardosa and Pirata; e.g., Dondale & Redner 1990). These different lifestyles are reflected in me- chanics of locomotion and activity response to variation in food supply (Ward & Hum- phries 1981; Walker & Rypstra 2001). There- fore, it is important to analyze sexual dimor- phism in locomotory organs in conjunction with data on the general activity pattern over an adult spider’s life span. The goal of this study was to relate the ac- tivity profile of males and females of a cur- sorial wolf spider, Venatrix lapidosa (McKay 1974), to gender specific differences in the morphology of their locomotory organs. The activity profile of these spiders was generated by conducting a fortnightly mark and recap- ture survey over a period of more than three years, covering seven generations of adult spi- ders. This allowed an analysis of both the var- iation of spider activity over their entire adult life, and incorporated seasonal variation, thus contrasting with all previous studies of wolf spiders that typically observed individuals for only up to a day (Richter et al. 1971; Cady 1984). I was not only interested in each in- dividuaFs activity (i.e. movement per unit time), but also the spatial aspect of movement (home range). Increases of both variables have the potential to augment fertilization success of males by increasing their encounter rates with females. However, these variables may not co-vary and higher activity may not nec- essarily increase home range size. Differential spatial use by males and females, as inferred from their home range, may also influence the operational sex ratio, thereby affecting the po- tential for male-male competition. Lastly, I analyzed locomotory structures of two Aus- tralasian genera of wolf spiders, Venatrix and Artoria, with different activity profiles of fe- males to determine if differences in behavior are reflected in leg length dimorphism across a higher taxonomic level. METHODS Study species. — Venatrix lapidosa is a nocturnal wolf spider inhabiting riparian grav- el banks in southeastern Australia (McKay 1974; Framenau & Vink 2001). It is a vagrant species, but brood caring females, and all spi- ders during overwintering, dig excavations 336 THE JOURNAL OF ARACHNOLOGY penultimate adult penultimate adult penultimate adult penultimate adult Figure 1 . — Relative leg length (residuals of leg length on cephalothorax width) (mean ± s.e.) of female and male penultimate and adult Venatrix lapidosa. For statistical analysis (ANOVA) see Table L under rocks that they line with a thin layer of silk (Framenau 1998, 2002a). Venatrix lapi- dosa is biennial, with juvenile development requiring up to 16 months. Adult life span of females may be up to eleven months and that of males up to ten months (this study; also Framenau & Elgar 2005). However, the av- erage life span of adults of both sexes does not generally exceed 6 months. The life cycle of V. lapidosa in the Victorian Alps is char- acterized by the maturation of two distinct co- horts within each year (Framenau & Elgar 2005). In autumn maturing cohorts, most in- dividuals mature between March and May, en- ter winter diapause, reproduce only after over- wintering and die by December. In spring maturing cohorts, spiders molt to maturity be- tween November and January, reproduce im- mediately and most spiders die by May. Over- lap between adult individuals of both cohorts is minimal and generally limited to a low number of long-lived individuals that reach the maturation period of the following cohort. Laboratory-reared males and females of dif- ferent cohorts readily mate (Cutler 2002) and this overlap permits gene flow between the different cohorts. Winter covers a large period of the adult life span of autumn maturing spi- ders, whereas adult individuals of the spring maturing cohorts live over summer, so cohorts were expected to differ considerably in their activity profile. Morphology of V. lapidosa, — I collected adult (9 females, 15 males) and penultimate (15 females, 12 males) V. lapidosa from a va- riety of populations at ten rivers in seven ma- jor catchments during a survey of riparian gravel banks in the Victorian Alps between^ November 1999 and January 2000 (Framenau et al. 2002). Leg length (sum of all segments measured dorsally) of all four pairs of legs was determined under a stereomicroscope to the nearest 0.1 mm. Carapace width was mea- sured above the coxae of the second pair of legs as an indicator of spider size (Hagstrum 1971; Jakob et al. 1996). I included penulti- mate spiders in the analysis to establish if an allometric increase in leg length occurs during the last molt, potentially indicating the impor- tance of dimorphism for mature, sexually ac- tive spiders. Younger than penultimate spiders were not used as it was impossible to establish their sex. Gender specific differences in leg length were analyzed using the residuals of a least squares regression using leg length on , cephalothorax width over both sexes and adult , and penultimate spiders. This gives rise to ^ measures of relative leg length, which were independent of body size. Residuals between ! the sexes and adult and penultimate spiders | were compared by two-way ANOVA. ■ i Mark and Recapture. — The mark and re- capture study of y. lapidosa was conducted on a gravel bank at the Avon River near Va- lencia Creek in Victoria, southeastern Austra- lia (37°48'S, 146°27'E). The climate of the re- gion is moderate with mean daily maximum . and minimum temperatures of 20.0 °C and 8.0 °C, respectively. Annual rainfall averages 594 ^ mm (Data from Maffra Forestry Office; Bu- reau of Meteorology, Melbourne). The gravel bank studied was bordered by the Avon River FRAMENAU-ACTIVITY AND SEXUAL DIMORPHISM IN WOLF SPIDERS 337 m '-o a 'm 0 -o Q. 0 m S ® 1 ® E > — - _ 0 ro m ■g S > '•o c 03 □ autumn maturing cohorts H spring maturing cohorts females males Figure 2. — Individual average daily distances (meters per day; mean ± s.e.) of female and male Venatrix lapidosa from the autumn and spring ma- turing cohorts at the Avon River. Average daily dis- tance moved based on two-week interfix intervals. To determine individual average daily distances, all average daily distances measured during the life span of one individual were averaged. on the southern side and a dense cover of veg- etation (wattle, Acacia trilobata, willow, Saiix sp. and blackberry, Bromus sp.) on a steep slope on the northern side, keeping spider im- migration and emigration minimal. With the exception of a few Acacia and Saiix shrubs, the gravel bank was bare of vegetation. When the survey commenced in 1996, the surface area of the bank was 1,830 m^. In August 1998, it diminished in size to 1,540 m^ due to a severe flood. A 5 m X 5 m grid was established on the gravel bank using wooden pegs. One co-or- dinate of this grid (‘Y’ -value) was perpendic- ular to the river and therefore expressed the relative distance from the water. Surveys were conducted fortnightly, from 8 November 1996 to 2 May 2000. This observation period cov- ered four spring maturing cohorts (1996- 1999) and three autumn maturing cohorts (1997-1999) (Table 2). All rocks large enough to provide shelter for spiders were overturned and replaced. Each survey started randomly at either end of the gravel bank. Each grid was examined in a spiral from exterior to interior, in order to prevent spiders from leaving a grid while it was searched. Spider locations were determined to within an accuracy of 1 m. New adult spiders in the population were individ- ually marked with a bee tag glued to their cephalothorax using a cyane-acrylate based adhesive (Supaglue Gel®). Cephalothorax width and body length were determined with vernier callipers to the nearest 0.1 mm. When returned, most spiders either remained without any movement under the same rock, or found shelter under the next available rock. Initial disturbance was therefore considered minimal. On subsequent encounters, only a spider's po- sition was recorded to avoid further distur- bance. Activity was determined using the average daily distance (Velocity’ in Samietz & Berger 1997), which is defined as the distance be- tween two consecutive fixes divided by the days between both observations. Not all spi- ders were recaptured every survey and aver- age daily distances significantly decreased with the time lapsed between two consecutive fixes (two-, four-, six-, eight-weekly interfix intervals; = 0.072, P < 0.001, n = 2,706). Thus, only average daily distances based on recaptures within two weeks of a previous one were considered in the analysis. In addition, these shorter intervals provided the most ac- curate picture of a spider’s movement. I com- pared individual activity of males and females ('individual average daily distance’) and au- tumn and spring maturing cohorts by their mean average daily distances over the whole observation period. To analyze seasonal vari- ability of activity, average daily distances were also determined for each month of the year pooled over all individuals of each sex but analyzed separately for autumn and spring maturing cohorts. In this case, the average dai- ly distance was obtained by analyzing cap- tures within two consecutive months were as- signed to the month that contained most days of the interfix interval. I pooled monthly data over all years after establishing that there was no between year variation. Home range. — I estimated home ranges using 100% minimum convex polygons (MCP; Mohr 1947). For low capture numbers, MCPs increase with each additional fix until a stable home range is reached. Regression analysis identified nine as the minimum num- ber of captures from which an increase in fixes did not result in a further, significant increase in home range size {R^ = 0.008, P = 0.397, n = 91). Increment analysis of home ranges (Kenward & Hodder 1996) showed that nine fixes provided an average of 90% of the full home range. This conforms to results of Sam- 338 THE JOURNAL OF ARACHNOLOGY ietz & Berger (1997), who show that home ranges (100% MCP) for insects appear to be stable from 10 captures. Two-week observa- tion periods guaranteed temporal indepen- dence of subsequent fixes, which is assumed if an animal can cross its home range within this period (White & Garrott 1990). As a mea- surement of home range shape, I calculated the range span as the distance of the furthest two points in a home range. Range centers were calculated as the mean of the X- and Y- values, corresponding to the established grid, of all fixes in a home range. Only one range in the spring maturing cohorts was based on more than eight fixes (Table 2). Therefore, be- tween cohort analysis of home range size and range span was not possible. Home range overlap was calculated for pairs of spiders that belonged to the same co- hort and so could potentially meet. Two val- ues of home range overlap could be deter- mined for each pair of spiders, i.e. how much of the home range of spider A was overlapped by the range of spider B, and vice versa. I used the average of both values as the mea- surement of range overlap. Comparative morphology. — The taxono- my of only two Australasian wolf spider gen- era, Venatrix Roewer (Framenau & Vink 2001; 23 species) and Artoria Thorell (Fra- menau 2002b; 1 1 species) is known sufficient- ly to allow interspecific comparative analyses. Both genera differ considerably in their mo- bility pattern, as most females of Venatrix dig permanent burrows or construct temporary ex- cavations during brood care, whereas females of Artoria are vagrant throughout their life (Framenau 2002b; also pers. obs). Least squares regression of leg length on cephalo- thorax width for all pairs of legs over all spe- cies derived from the primary taxonomic lit- erature of both genera (Framenau & Vink 200 1 ; Framenau 2002b) provided measures of relative leg length (residuals) compared to a TypicaT lycosid. To test for sexual dimor- phism within both genera, these residuals were compared between the sexes using two- sample t-tests. Statistical analysis. — Home range and ac- tivity parameters were calculated using the software package ‘RANGES V’ (Ken ward & Hodder 1996). Subsequent statistical analyses were performed with ‘SYSTAT Version 9' (SPSS Corp. 1998). Data that did not comply with ANOVA assumptions were log-trans- formed, in case of average daily distances (log + 1 )-transformed (Quinn & Keough 2002). If normality of data could not be achieved, non- parametric tests (Mann- Whitney U Test) were used to compare sexes. Measurements are giv- en as mean ± standard error (s.e.) unless oth- I erwise indicated. Voucher specimens of V. lapidosa were deposited at the Museum Vic- toria, Melbourne, and the Western Australian Museum, Perth. RESULTS Morphology of V, lapidosa. — The carapace width (± s.e.) of adult female V. lapidosa (6.66 ± 0.18 mm, n = 9) was significantly larger than that of males (5.91 ± 0.06 mm, n = 15; separate t = 4.019, d.f. = 10.2, P = 0.002). The length of all legs was positively correlated with cephalothorax width and resid- uals of these regressions yielded measures of relative leg length (Table 1). All legs were comparatively longer in males than in females for adult and penultimate spiders (Table 1, Fig. 1). A significant interaction between age and sex for all legs indicates a proportionally higher elongation (allometric growth) for male legs during their last molt compared with fe- males (Fig. 1). Mark and recapture survey. — A total of 741 males and 712 females were individually marked over a period of 3.5 years, yielding an i overall even sex ratio (x^ — 0.802, P = 0.37) 1 (Table 2). However, recapture rates, i.e. how j often individual spiders were caught, differed | significantly for males and females (Mann- i Whitney U = 314226.5, P = 0.008) due to a higher number of males encountered only once. The total number of fixes analyzed was , 4,963 yielding a detailed life cycle profile for . each cohort in each year (see Framenau & El- ; gar 2005). Activity. — Mean average daily distances of individual spiders, based on two-week interfix intervals, were significantly higher for males than females, and higher for individuals of the spring mating cohorts than of the autumn ma- turing cohorts (two-way ANOVA; sex: Fj^ggg = 6.045, P = 0.014; cohort: F,,699 = 4.816," P — 0.029; interaction: F,_699 = 0,384, P = 0.536) (Fig. 2). Monthly average daily distances in the au- i tumn cohort showed no significant difference j between sexes (two-way ANOVA; F]_||8(, = FRAMENAU-ACTIVITY AND SEXUAL DIMORPHISM IN WOLF SPIDERS 339 ■o m ■o 05 © < Autumn maturing cohorts (February 1997 - January 2000) ® 2 1 - ^ 3 FMAMJ JASONDJ Spring maturing cohorts (November 1996 - May 2000) 2 - 1 - H femaies ■ males IliliSillSI 4.8 H M M Figure 3. — Monthly average daily distances (meters per day; mean ± s.e.) of female and male Venatrix lapidosa from the autumn and spring maturing cohorts during the survey at the Avon River. The month on the far left in each graph represents the maturation of each cohort. Average daily distance moved is based on two-week interfix intervals. To determine monthly average daily distances, all average daily distances within a month were averaged over all individuals. Asterisk (*) indicates significant difference between sexes. 0.087, P = 0.769), however, there were sig- nificant differences between months (Fiojjg^ = 7.908, P < 0.001; interaction: = i.055, P - 0.394; January excluded due to missing variation between sexes; Fig. 3). In the spring cohort, there was also no overall gender spe- cific difference in monthly average daily dis- tances (two-way ANOVA; Fj^gg = 1.514, F = 0.219) and, in contrast to the autumn cohort, there was no differences between months (^5,389 ” 2.188, P = 0.055; interaction: F53g9 = 0.892, P = 0.486; May-October excluded due to missing variance in sex or months; Fig. 3). However, a within months comparison of average daily distances between males and fe- males revealed significantly higher male activ- ity in August (pooled t = 3.048, d.f. = 28, P = 0.005) and September (pooled t = 2.199, Table 1 , — Comparison of relative leg length (residuals of leg length on carapace width) between male and female and adult and penultimate Venatrix lapidosa. Regression of leg length on carapace width: Leg 1: R2 = 0.543, slope = 2.978, P < 0.001, n = 51; leg 2: = 0.608, slope = 2.972, P < 0.001, n = 51; leg 3: = 0.689, slope = 2.909, P < 0.001, n = 51; leg 4: R2 = 0.672, slope = 3.457, P < 0.001, « = 51. Given are the Fi^47-values and significance level (*F < 0.05, < 0.01, < 0.001) of a two-way ANOVA. Factor Leg 1 Leg 2 Leg 3 Leg 4 Sex 60.918*** 55.644*** 35.171*** 26.328*** Age 74.322*** 76.209*** 39.534*** 39 378*** Interaction: Sex * Age 22.746*** 18.380*** 13.416** 7.923** 340 THE JOURNAL OF ARACHNOLOGY ^ 60 Q. ® 50 o ® c 40 2 E 30 o X 20 Figure 4. — Intra- and intersexual home range overlap (mean ± s.e.) in Venatrix lapidosa within the 1997 autumn maturing cohort. No other cohort provided a sufficient number of home range esti- mates to compare between and within sexes (see Table 1). femaie female male d.f. = 221, P = 0.029) for the autumn cohort and in January (pooled t = 2.673, d.f. = 120, P = 0.026) for the spring cohort (Fig. 3). Home range* — Due to the major flood in 1998, home range overlap analysis was re- stricted to the autumn 1 997 cohort when males and females were caught in sufficient numbers (minimum of nine fixes) to allow a comparison between the sexes (Table 2). Therefore, a comparison between cohorts was not possible. Home range estimates (100% MCP ± s.e.) did not differ significantly be- tween males (302 ± 16 m^, n = 42) and fe- males (311 ± 17 m^, n = 49; pooled t = 0.375, d.f. = 89; F = 0.709). Home range span also did not differ between males (55.4 ± 2.1 m, « = 42) and females (53.7 ± 2.2 m, n = 50; pooled t = 0.548, d.f, = 89; P = 0.585). Home range centers did not show a significant difference between sexes along the river (X-coordinate; pooled t = 1.732, d.f. = 89; P = 0.087), but the relative distance from the river (Y-value) was significantly higher for females (10.1 ± 0.2, n = 49) than males (9.3 ± 0.2, n = 42; pooled t = 3.156, d.f. = 89; P = 0.002). In addition, females carrying an eggsac were found significantly further away from the water (Y-coordinate ± s.e; 9.8 ± 3.3, n = 124) than females not caring for brood (9.0 ± 3.0, n = 1,179; pooled t = 2.372, d.f. = 1301; P = 0.018). Overall, range overlap was high (> 50%, Fig. 4), but it differed sig- nificantly between sexes, with male-male overlap highest and female-female overlap lowest (ANOVA; F23743 = 19.315, P < 0.001) (Fig. 4). Comparative morphology. — There was a positive correlation between cephalothorax width and leg length in both Venatrix and Ar- toria and measures of relative leg length were obtained from the residuals of the respective regressions (Table 3, Fig. 5). Males had com- paratively longer legs than females within the genus Venatrix, but there was no gender spe- cific difference in the relative leg length in Artoria (Table 3, Fig. 5). DISCUSSION There was a considerable difference in the activity and mobility pattern of male and fe- male y. lapidosa, which corresponded to a pronounced sexual dimorphism in the length of their legs. The evolution of longer legs in Table 2. — Capture statistics of the mark and recapture survey of Venatrix lapidosa at the Avon River, distinguished by cohorts. Average daily distance based on two-weekly interfix intervals only (see text). §Considered in analysis of average daily distances; f Considered in home range analysis. Cohort Spring 1996 Autumn 1997 Spring 1997 Autumn 1998 Spring 1998 Autumn 1999 Spring 1999 Males: Total number marked 115 179 80 168 86 84 29 Recaptured at least once§ 58 157 35 121 43 67 9 Minimum of nine captures^ 0 38 0 2 0 2 0 Females: Total number marked 127 176 105 140 77 66 21 Recaptured at least once§ 88 158 72 96 48 56 7 Minimum of nine captures^ 1 48 0 0 0 0 0 framenau-activity and sexual dimorphism in wolf spiders 341 Artoria Venatrix Artoria Venatrix Artoria Venatrix Artoria Venatrix Eigure 5. — Relative leg length (residuals of leg length on cephalothorax width) (mean ± s.e.) of female and male species of Venatrix and Artoria. For statistical analysis between sexes in each genus (t-test) see Table 3. males may be the result of an increased like- lihood to encounter more stationary females assuming a higher energy efficiency or speed as a result of leg elongation. The lack of sex- ual dimorphism in leg length in juvenile V. lapidosa and the genus Artoria (in which fe- males are vagrant) supports this argument. Activity. — Venatrix lapidosa is a compar- atively immobile spider. Similar low activity occurs in other cursorial spiders inhabiting terrestrial-aquatic ecotones (Framenau et al. 1996a; Kreiter & Wise 2001). Limited mobil- ity of riparian species may be a result of their fragmented habitat consisting of generally small isolated gravel banks. In addition, high prey availability near the water edge may ren- der it unnecessary to move (Greenstone 1983). As expected for poikilothermic ani- mals, activity between cohorts differed, most likely reflecting seasonal patterns. Activity was lower for individuals of the autumn mat- ing cohort, due to a drop in movement over winter. Individuals of the spring mating co- horts, although more active than the autumn mating cohorts, showed no significant differ- ence in activity between months. These indi- viduals are adults mainly in summer. Temper- ature dependent movement patterns have also been reported in other wolf spiders, such as Pardosa amentata (Clerck 1757) (Ford 1978). Activity patterns of males and females are similar within both cohorts, with males the more active sex. A variety of studies on wolf spiders have shown that an increase in male activity reflects mate searching (e.g., Hallan- der 1967; Framenau et al. 1996a). In V. lapi- dosa, significantly higher male activity ap- pears to be temporary, emerging about three months after maturation (delayed in the au- tumn maturing cohort by winter diapause). In the autumn maturing cohort, males emerge earlier from diapause and are more active than females two months prior to female egg pro- duction, suggesting that males are searching for mates. Higher male activity in the spring mating cohort cannot be as easily explained in terms of mate searching, as male activity was particularly high in January, when fe- males had already commenced egg produc- tion. Since higher male activity is observed for only a few months, females appear to move more than suggested by previous studies on wolf spiders (Richter et al. 1971; Hallander 1967). Initial female activity may be high due to increased foraging effort to meet energetic requirements for egg production (Kreiter & Wise 2001). Movement in female V. lapidosa may also be induced by the apparent prefer- ence of females to oviposit some distance from the water. Activity may subsequently drop, as females become sedentary to care for their brood (Hackman 1957; Hallander 1967; Framenau et al. 1996a, b; Nyffeler 2000). An unusually low proportion of ovipositing fe- males in V. lapidosa (Framenau & Elgar 2005) compared to other lycosids (e.g., Fra- menau 1996a; Humphreys 1976) suggests a comparatively low number of stationary, broodcaring females and may partly explain why differences in activity between males and females is limited. Home range. — Despite the temporary in- crease in male activity, home range size did not differ between males and females in V. lapidosa. The movements of V. lapidosa may have been restricted by the size of the study site itself, but average home range size and range span for both sexes were considerably smaller than the surface and length of the in- vestigated gravel bank. 342 THE JOURNAL OF ARACHNOLOGY Table 3. — Comparison of relative leg length (residuals of leg length on carapace width) between males and females in species of the genera Venatrix and Artoria. Regression of leg length on carapace width: Leg 1: R2 = 0.899, slope = 3.292, P < 0.001, « = 51; leg 2: R^ = 0.912, slope = 3.037, P < 0.001, n = 58; leg 3: R^ = 0.903, slope = 2.772, P < 0.001, n = 58; leg 4: R^ = 0.892, slope - 3.512, P < 0.001, n — 56. Given are the r-values (pooled variance) with significance level (n.s. non significant; *P < 0.05, < 0.01, < 0.001) and degrees of freedom (d.f.). Factor Leg 1 Leg 2 Leg 3 Leg 4 Venatrix Sex 4. 140*** 3.056** 3.182** 2.039* d.f. 33 34 34 34 Artoria Sex 1.832 n.s. 1.529 n.s. 1.642 n.s. 0.770 n.s. d.f. 19 20 20 18 Although home range size was similar be- tween the sexes, there was a significant dif- ference in the distribution of range centers be- tween females and males. Female range centers were, on average, located further away from the water as ovipositing females retreat from the border of the gravel bank. Females may not tolerate a high degree of soil mois- ture, or they may look for more protected ar- eas from varying water levels, before exca- vation of brood chambers. Site specificity in relation to abiotic factors occurs frequently in lycosids that build permanent burrows (e.g., Humphreys 1976; Milasowszky & Zulka 1998). Females of A. cinerea (Fabricius 1777) and Trochosa ruhcola (DeGeer 1778), two ly- cosids inhabiting shore habitats, also move away from the water prior to brood care (Hackman 1957; Framenau et al. 1996b). Dif- ferential microhabitat preferences can have a strong influence on the activity and distribu- tion patterns of individuals. In wolf spiders, intraspecific habitat preferences not only dif- fer between females with and without eggsacs (Edgar 1971; Hallander 1967; Greenstone 1983), but also between sexes (Cady 1984) and adults and juveniles (Edgar 1969, 1971; Kronk & Riechert 1979). The utilization of areas further away from the water, together with equal home range size, may explain the lower female-female range overlap compared to males. In wolf spi- ders, males and females do not encounter each other haphazardly. Males follow silk draglines laid by receptive females which contain sex- attracting pheromones (Hedgekar & Dondale 1969; Tietjen & Rovner 1982). In addition, strong agonistic behavior within sexes has been reported in wolf spiders (Aspey 1977a, b; Fernandez-Montraveta & Ortega 1991). The effect of different habitat requirements, i.e. the search of a favorable location for brood care, appears to be stronger than male- female attraction or intrasexual aggression. Sexual dimorphism. — Venatrix lapidosa is sexually dimorphic. Females are generally larger than males, but males have compara- tively longer legs. The sex ratio of V. lapidosa was not biased in the autumn mating cohort, and limited to later months in the spring co- horts when many females were already caring for their brood (Framenau & Elgar 2005). Fur- ther, higher home range overlap between i males suggests greater rather than less oppor- ' tunity for male-male competition. Therefore, my data are not consistent with the underlying assumptions of the model developed by Voll- i rath & Parker (1992) that relates sexual size dimorphism in spiders to reduced male-male competition due to an increase in mortality caused by mate search. Sexual dimorphism in V. lapidosa most likely evolved through a fe- cundity advantage for larger females (Prenter et al. 1997, 1998, 1999); clutch size increases j with body size in many wolf spider species , (Marshall & Gittleman 1994; Simpson 1995). | While increased female fecundity may ex- | plain size differences between males and fe- | males, sexual selection through indirect male- j male competition may explain the i comparatively longer legs of males. Allome- | trie growth leading to relatively longer legs ' only takes place in males and mainly during the final molt supporting an evolutionary hy- pothesis of leg elongation in males rather than leg shorting in females due to burrowing be- j FRAMENAU-ACTIVITY AND SEXUAL DIMORPHISM IN WOLF SPIDERS 343 havion The production of longer legs may be ontogenetically costly and thus would be off- set by energetically more efficient movement. There are no experimental or comparative data of increased movement efficiency with longer legs in arthropods (J. Shultz pers. Comm.) and the relationship between leg di- mensions (length and thickness) and metabol- ic rate are complex and also entail the mass of the spider. Simple lever mechanics predicts that if the length of the output lever arm in- creases, the velocity and excursion at the end of the lever will increase (and thus speed and distance moved per stride), but that the force the lever will exert will decrease (Manton 1977; Alexander 1982; Hildebrand & Goslow 2003). Males can compensate the loss of force by reducing their own mass which, in turn, augments selection for smaller males, provid- ing a novel aspect in the explanation for sex- ual size dimorphism in vagrant spiders. Over- all, longer-legged, smaller males are able to search faster and more extensively for females and potentially increase their encounter rates with females. This advantage would be fa- vored by sexual selection if it provided males with a competitive edge in terms of fertiliza- tion success. A limb elongation due to more efficient lo- comotion is also supported by the fact that all four legs show the same allometric pattern which was not required if the difference in leg length between sexes arose through sexual cannibalism (Elgar et al. 1990), male-male combats (Tseng & Rowe 1999), or in combi- nation with leg ornamentation in used in courtship display (Kroeestedt 1990; Hebets & Uetz 2000). In addition, sexual cannibalism and male-male combats are extremely rare in wolf spiders (Aspey 1977a). Alternatively, different foraging behavior between males and females could provide an explanation of sex- ual dimorphism based on different locomotory patterns (Givens 1978). However, due to low- er metabolic requirements male wolf spiders attack considerably fewer prey than females (Walker & Rypstra 2001). Longer legs may also provide a sensory advantage due to an increased radius to mount olfactory chemo- receptors or trichobothria. In wolf spiders, ol- faction plays some role in mate search, how- ever, the main senses used by males to follow trail lines of females are situated on the dorsal side of the cymbium of the pedipalps (Tietjen an Rovner 1982). The comparison in the pattern of leg-length dimorphism in Venatrix and Artoria provide further evidence that male mate-searching be- havior favors relatively longer legs in males. Although males in Artoria also tend to have longer legs than females, this difference is not significant within the genus and is far less pro- nounced than in Venatrix. It appears unlikely that leg length dimorphism arises through shortening of female legs due to burrowing behavior, as allometric growth occurs between penultimate and adult spiders. In addition, there is no evidence that female Venatrix have comparatively shorter legs than vagrant Arto- ria. This study provides evidence that longer legs in male wolf spiders are mainly caused by sexual selection through indirect competi- tion with increased male activity in searching for a mate. To further elucidate the selective forces responsible for the elongation of male legs, future work should focus on three ques- tions. Firstly, it is important to experimentally confirm the assumption that longer legs are more energy efficient in spider movement. Secondly, evidence that higher male activity will ultimately lead to higher fertilization suc- cess is required. This is strongly dependent on the species mating system in question, and re- quires an understanding of multiple mating and sperm priority patterns of wolf spiders (Austad 1984; Elgar 1998). Although V. lap- idosa has been reported to mate multiply in- creasing the chance of male mate competition (Cutler 2002), there is no information on sperm priority patterns in this and other Ly- cosidae. Lastly, comparative studies in leg length dimorphism in comparison with the life time activity patterns of cursorial spiders on a broader taxonomic base may help us under- stand to what extent sexual dimorphism of limbs are under natural or sexual selection. ACKNOWLEDGMENTS I am grateful to Melissa Thomas, Mark El- gar, Douglas Morse, and Mark Harvey for in- valuable comments on earlier drafts of this manuscript. Jeff Shultz and Ken Prestwich provided important information on arthropod locomotion and energetics. Melissa Thomas, Fleur de Crespigny, Shar Ramamurthy, Eliz- abeth Dalgleish, Romke Kats, Karen Blaak- 344 THE JOURNAL OF ARACHNOLOGY meer, and Jenny Maupin provided assistance in the field and laboratory. Casey and Elisa- beth Klomp offered friendly accommodation at the Caravan Park in Stratford on the River Avon. Cath Handasyde provided access to her copy of ‘RANGES V% and Graham Coulson provided further assistance in the analysis of home range data. This study was funded by the University of Melbourne (MRS), The Eco- logical Society of Australia, The River Basin Management Society, and the Donors Asso- ciation for the Promotion of Science and Hu- manities in Germany (Stifterverband fiir die Deutsche Wissenschaft). LITERATURE CITED Alexander, R.McN. 1982. Size, shape and structure for running and flying. Pp. 309-324. In A Com- panion to Animal Physiology. (C.R. Taylor, V. Johanson & L. Bolis, ed.). Cambridge University Press, Cambridge. Andersson, M. 1994. Sexual Selection. Princeton University Press, Princeton. Aspey, W.P. 1977a. Wolf spider sociobiology: I. Agonistic display and dominance-subordinance relations in adult male Schizocosa crassipes. Be- haviour 62: 103-141. Aspey, W.P. 1977b. Wolf spider sociobiology: 11. Density parameters influencing agonistic behav- iour in Schizocosa crassipes. Behaviour 62:142- 163. Austad, S.N. 1984. Evolution of sperm priority pat- terns in spiders. Pp. 223-249. In Sperm Com- petition and the Evolution of Animal Mating Systems (R.L. Smith, ed.). Academic Press, New York. Bulmer, M.G. 1983. The significance of protandry in social Hymenoptera. American Naturalist 121: 540-551. Cady, A.B. 1984. Microhabitat selection and loco- motory activity of Schizocosa ocreata (Walcken- aer) (Araneae: Lycosidae). Journal of Arachnol- ogy 11:297-307. Cutler, A.R. 2002. Courtship in the Wolf Spider Venatrix lapidosa (Lycosidae: Araneae). Hon- ours Thesis, The University of Melbourne, Mel- bourne. Darwin, C, 1871. The Descent of Man, and Selec- tion in Relation to Sex. J. Murray, London Dondale, C.D. & J.H. Redner. 1990. The Insects and Arachnids of Canada, Part 17. The Wolf Spi- ders, Nursery Web Spiders, and Lynx Spiders of Canada and Alaska (Araneae: Lycosidae, Pisaur- idae, and Oxyopidae). Biosystematics Research Centre, Ottawa, Ontario. Eberhard, W.G. & M.C. Marin. 1996. Sexual be- havior and the enlarged hind legs of male Me- galopus armatus (Coleoptera, Chrysomelidae). Journal of the Kansas Entomological Society 69: 1-8. Edgar, W.D. 1969. Prey and predators of the wolf spider Lycosa lugubris. Journal of Zoology, Lon- don 159:405-411. Edgar, W.D. 1971. The life-cycle, abundance and seasonal movement of the wolf spider, Lycosa (Pardosa) lugubris, in Central Scotland. Journal of Animal Ecology 40:303-322. Elgar, M.A. 1991. Sexual cannibalism, size dimor- phism, and courtship behaviour in orb-weaving spiders (Araneidae). Evolution 45:444-448. Elgar, M.A. 1998. Sperm competition and sexual selection in spiders and other arachnids. Pp. | 307 — 337. In Sperm Competition and Sexual Se- i lection (TR. Birkhead & A.P. Mpller, ed.). Aca- | demic Press, New York Elgar, M.A., N. Ghaffar & A.F. Read, 1990. Sexual I dimorphism in leg length among orb-weaving ; spiders: a possible role for sexual cannibalism. Journal of Zoology, London 222:455-470. , Enders, F. 1976. Clutch size related to hunting man- ' ner of spider species. Annals of the Entomolog- ! ical Society of America 69:991-998. I Fairbairn, D.J. 1997. Allometry for sexual size di- i morphism: pattern and process in the coevolution ' of body size in males and females. Annual Re- j views in Ecology and Systematics 28:659-687. | Fernandez-Montraveta, C. & J. Ortega. 1991. Own- ! er-biased agonistic behavior in female Lycosa ta- rantula fasciiventris (Araneae, Lycosidae). Jour- nal of Arachnology 19:80-84. - Ford, M.J. 1978. Locomotory activity and the pre- dation strategy of the wolf spider Pardosa amen- tata (Clerck) (Lycosidae). Animal Behaviour 26: 31-35. Framenau, V.W. 1998. Life cycles of Lycosa lapi- dosa McKay, 1974, and Lycosa arenaris Hogg, 1905, two riparian wolf spiders from south east- ern Australia. Pp. 227 — 234, In Proceedings of the 17'h European Colloquium of Arachnology, Edinburgh 1997 (P.A. Selden, ed.). British Ar- achnological Society, Burnham Beeches, Bucks. Framenau, V.W. (2002a). Taxonomy, Life History Characteristics, and Ecology of Riparian Wolf Spiders (Araneae, Lycosidae) in the Victorian Alps, South-East Australia. Ph.D. thesis. The j University of Melbourne, Melbourne. Framenau, V.W. (2002b). Review of the wolf spider j genus Artoria Thorell (Araneae, Lycosidae). In- vertebrate Systematics 16:209-235. Framenau, V.W, & M.A. Elgar. 2005. Cohort de- pendent life history traits in a wolf spider (Ara- neae, Lycosidae) with bimodal life cycle. Journal of Zoology, London 265:179-188. Framenau, V., M. Dieterich, M. Reich & H. Plach- ter. 1996a. Life cycle, habitat selection and home : ranges of Arctosa cinerea (Fabricius, 1777) (Ar- : aneae: Lycosidae) in a braided section of the Up- i FRAMENAU-ACTIVITY AND SEXUAL DIMORPHISM IN WOLF SPIDERS 345 per Isar (Germany, Bavaria). Revue Suisse Zool- ogie vol. hors serie 1:223-234. Framenau, V.W., R Manderbach & M. Baehr. 2002, Riparian gravel banks of upland and lowland riv- ers in Victoria (South-east Australia): Arthropod community structure and life history patterns along a longitudinal gradient. Australian Journal of Zoology 50:103-123. Framenau, V., M. Reich, & H. Plachter. 1996b. Zum Wanderverhalten und zur Nahrungsokologie von Arctosa cinerea (Fabricius, 1777) (Araneae: Ly- cosidae) in einer alpinen Wildflusslandschaft. Verhandlungen der Gesellschaft fiir Okologie 26: 369-376. Framenau, V.W. and Vink, CJ. 2001. Revision of the wolf spider genus Venatrix Roewer. Inverte- brate Taxonomy 15:927-970. Gasnier, T.R., C.S. de Azevedo, M.R Torres-San- chez & H. Hofer. 2002, Adult size of eight hunt- ing spider species in central Amazonia: temporal variations and sexual dimorphism. Journal of Ar- achnology 30:146-154. Givens, R.P. 1978. Dimorphic foraging strategies of a salticid spider (Phidippus audax). Ecology 59: 309-321. Greenstone, M.H. 1983. Site-specifity and site te- nacity in a wolf spider: a serological dietary analysis. Oecologia 56:79-83. Gunnarsson, B. & J. Johnsson. 1990. Protandry and moulting to maturity in the spider Pityohyphan- tes phrygianus. Oikos 59:205-212. Hackman, W. 1957. Studies on the ecology of the wolf spider Trochosa ruricola Deg. Commenta- tiones Biologicae 16:1-34. Hagstrum, D.W. 1971. Carapace width as a tool for evaluating the rate of development of spiders in the laboratory and the field. Annals of the En- tomological Society of America 64:757-760. Hallander, H. 1967. Range and movements of the wolf spiders Pardosa c he lata (O. F. Miiller) and P. pullata (Clerck). Oikos 18:360-364, Head, G. 1995. Selection on fecundity and variation in degree of sexual size dimorphism among spe- cies (class: Araneae). Evolution 49:776-781. Hebets, E.A. & G.W. Uetz. 2000. Leg ornamenta- tion and the efficacy of courtship display in four species of wolf spider (Araneae: Lycosidae). Be- havioral Ecology and Sociobiology 47:280-286. Hedgekar, B.M. & C.D. Dondale. 1969. A contact sex pheromone and some response parameters in lycosid spiders. Canadian Journal of Zoology 47: 1-4. Hedrick, A.V. & E.J. Temeles. 1989. The evolution of sexual dimorphism in animals: hyptheses and tests. Trends in Ecology and Evolution 4:136- 138. Hildebrand, M. & G.E. Goslow. 2003. Analysis of Vertebrate Structure. John Wiley & Sons, New York. Hormiga, G., W.G. Eberhard & J.A. Coddington, 1995. Web-construction behaviour in Phonog- natha and the phylogeny of Nephiline and Te- tragnathid spiders (Araneae: Tetragnathidae). Australian Journal of Zoology 43:313-364. Humphreys, W.F. 1976. The population dynamics of an Australian wolf spider, Geolycosa godejfroyi (L. Koch 1865) (Araneae: Lycosidae). Journal of Animal Ecology 45:59-80. Jakob, E.M., S.D. Marshall & G.W. Uetz. 1996. Es- timating fitness: a comparison of body condition indices. Oikos 77:61-67. Ken ward R.E. & K.H. Hodder. 1996. Ranges V. An Analysis System for Biological Location Data. Natural Environment Research Council, Dorset. Kreiter, N. & D.H. Wise. 2001. Prey availability limits fecundity and influences the movement pattern of female fishing spiders. Oecologia 127: 417-424. Kronestedt, T. 1990. Separation of two species standing as Alopecosa aculeata (Clerck) by mor- phological, behavioural and ecological charac- ters, with remarks on related species in the pul- verulenta group. Zoologica Scripta 19:203-225. Kronk, A.E. & S.E. Riechert. 1979. Parameters af- fecting the habitat choice of a desert wolf spider, Lycosa santrita Chamberlin and Ivie. Journal of Arachnology 7:155-166. Malmgren, J.C. & M. Thollesson. 2001. Sexual size and shape dimorphism in two species of newts, Triturus cristatiis and T. vulgaris (Caudata: Sal- amandridae). Journal of Zoology, London 249: 127-136. Manton, S.M. 1977. The Arthropoda. Habits, Func- tional Morphology, and Evolution. Clarendon Press, Oxford. Marshall, S.D. & J.L, Gittleman. 1994. Clutch size in spiders: is more better? Functional Ecology 8: 118-124. McKay, R.J. 1974. The wolf spiders of Australia (Araneae: Lycosidae): 2. The arenaris group. Memoirs of the Queensland Museum 17:1-19. Milasowszky, N, & K.P. Zulka. 1998. Habitat re- quirements and conservation of the “flagship species” Lycosa singoriensis (Laxmann 1770) (Araneae: Lycosidae) in the National Park Neu- siedler See-Seewinkel (Austria). Zeitschrift fiir Okologie und Naturschutz 7:111-119. Mohr, C.O. 1947. Table of equivalent populations of North American small mammuls. American Midland Naturalist 37:223-249. Montgomery, T.H. 1910. The significance of the courtship and secondary sexual characters of ar- aneads. The American Naturalist 44:151-177. Nyffeler, M. 2000. Do adult female lycosids feed during the period of maternal care? Bulletin of the British Arachnological Society 11:388-390. Ohtsuka, S. & R. Huys. 2001. Sexual dimorphism 346 THE JOURNAL OF ARACHNOLOGY in calanoid copepods: morphology and function. Hydrobiologia 453:441-466. Peters, W.L. & I.C. Campbell. 1991. Ephemeroptera (Mayflies). Pp. 279 — 293. In The Insects of Aus- tralia. A Textbook for Students and Research Workers (CSIRO, ed.). Melbourne University Press, Carlton. Prenter, J., W.L Montgomery & R.W. Elwood. 1997. Sexual dimorphism in northern temperate spiders: implications for the differential mortality model. Journal of Zoology, London 243:341- 349. Prenter, J., R.W. Elwood & W.L Montgomery. 1998. No association between sexual size dimor- phism and life histories in spiders. Proceedings of the Royal Society of London, Series B 265: 57-62. Prenter, J., R.W. Elwood & W.L Montgomery. 1999. Sexual size dimorphism and reproductive investment by female spiders: a comparative analysis. Evolution 53:1987-1994. Quinn, G.P. & M.G. Keough. 2002. Experimental Design and Data Analysis for Biologists. Cam- bridge University Press, Cambridge. Reynolds, J.D. & PH. Harvey. 1994. Sexual selec- tion and the evolution of sex differences. Pp. 53 — 70. In The Differences between the Sexes (R.V. Short & E. Balaban, ed.). Cambridge Uni- versity Press, Cambridge. Richter, C.J.J., J. den Hollander & L. Vlijm. 1971. Differences in breeding and motility between Pardosa pullata (Clerck) and Pardosa prativaga (L. Koch), (Lycosidae, Araneae) in relation to habitat. Oecologia 6:318-327. Samietz, J. & U. Berger. 1997. Evaluation of move- ment parameters in insects — bias and robustness with regard to resight numbers. Oecologia 110: 40-49. Selander, R.K. 1972. Sexual selection and dimor- phism in birds. Pp. 180 — 230. In Sexual Selec- tion and the Decent of Man, 1871-1971 (D.C. Campbell, ed.). Aldine, Chicago. Shine, R. 1989. Ecological causes for the evolution of sexual dimorphism: a review of the evidence. Quarterly Reviews in Biology 64:419-461. Simpson, M.R. 1995. Covariation of spider egg and clutch size: the influence of foraging and parental care. Ecology 76:795-800. Thornhill, R. & J. Alcock. 1983. The Evolution of Insect Mating Systems. Cambridge University Press, Cambridge. Tietjen, W.D. & J.S. Rovner. 1982. Chemical com- munication in lycosids and other spiders. Pp. 249 — 279. In Spider Communication: Mecha- nisms and Ecological Significance (P.N. Witt & J.S. Rovner, ed.). Princeton University Press, Princeton. Tseng, M. & L. Rowe. 1999. Sexual dimorphism and allometry in the giant water strider Gigan- tometra gigas. Canadian Journal of Zoology 77: 923-929. Vollrath, F. & G.A. Parker. 1992. Sexual dimor- phism and distorted sex ratios in spiders. Nature 360:156-159. Walker, S.E. & A.L. Rypstra. 2001. Sexual dimor- phism in functional response and troph mor- phology in Rabidosa rabida (Araneae: Lycosi- dae). American Midland Naturalist 146:161-170. Ward, TM. & W.E Humphreys. 1981. Locomotion in burrowing and vagrant wolf spiders (Lycosi- dae). Journal of Experimental Biology 92:305- 321. White, G.C. & R.A. Garrott. 1990. Analysis of Wildlife Radio-Tracking Data. Academic Press, San Diego/London. Manuscript received 23 August 2004, revised 31 March 2005. 2005. The Journal of Arachnology 33:347-376 EVOLUTION OF ORNAMENTATION AND COURTSHIP BEHAVIOR IN SCHIZOCOSA: INSIGHTS FROM A PHYLOGENY BASED ON MORPHOLOGY (ARANEAE, LYCOSIDAE) Gail E. Stratton: Department of Biology, University of Mississippi, University, Mississippi 38677, U.S.A. E-mail: Byges@olemiss.edu ABSTRACT. A phylogenetic analysis for the North American Schizocosa species was undertaken by scoring 49 morphological characters for 31 taxa representing all of the Nearctic species of Schizocosa plus individuals that are hybrids between S. ocreata and S. rovneri. Rabidosa rabida, Allocosa georgicola and Gladicosa pulchra were used as outgroups. Three clades are recognized: a large clade from eastern North America (Clade A) within which is nested the S. ocreata clade; Clade B, which includes the widespread S. avida and the western 5'. mccooki, and a smaller, third clade, Clade C. Sexual ornamentation occurs on the first legs of mature males of several species within the Schizocosa and takes the form of pigmentation and or bristles primarily on the tibia of leg I; there is at least one species with bristles in each of the three main clades. Mapping the occurrence of male ornamentation on the preferred phylogeny suggests that ornamentation evolved 5 or 6 separate times and was subsequently lost 2 or 3 times. The ornamentation is concentrated in the S. ocreata clade, a clade defined by a finger like projection on the paleal process of the male pedipalp. Courtship behavior is known for 20 of the 3 1 taxa. All species studied utilize chemical communication and seismic signals for communication; some species also have distinct visual signals. Seismic signals are produced by palpal drumming (as is seen in several species within Clade B), or by stridulation (seen in Clade A). Visual signals consisting of movements of the first pair of legs are common in species that are distinctly ornamented. This study provides the first phylogenetic study of a North American genus of wolf spider and provides morphometric comparisons of the North American species in Schizocosa. Keywords: Cladistics, sexual selection, secondary sexual characteristics, evolution of behavior, spiders, multimodal signal, seismic signal In spiders, sexual ornamentation is most ev- ident in groups that have exceptional eyesight (e.g., Salticidae and Lycosidae) with orna- ments being found on mature males in places that are visible to females or other males. In contrast to the colorful salticids (e.g. the North American genus Habronattus EO.P Cam- bridge 1901, Peckham & Peckham 1889, 1890; Griswold 1987; Maddison & Hedin 2003), the ornamentation on lycosids tends to be in black or white, and is generally limited to the first pair of legs or to the pedipalps of mature males. Examples of such ornaments in wolf spiders are wide spread and include dark pigmentation on some part of the first pair of legs as seen in Alopecosa aculeata (Clerck 1757) and A. barbipes (Sundevall 1833) (Kro- nestedt 1990) or on the pedipalps as seen in Pardosa wagleri (Hahn 1822) and P. satura- tior Simon 1937 (Barthei & Helversen 1990) or P. saxatilis (Hentz 1844) (Dondale & Red- ner 1984). In addition, many members of the North American Geolycosa Montgomery 1904 have contrasting pigmented hairs on their first legs (Wallace 1942a). And, as re- ported in Dondale & Redner (1978) and in this study, males of several members of the genus Schizocosa Chamberlin 1904 have darkly pigmented legs and or tibial bristles. Revised by Dondale & Redeer (1978), the Nearctic species of the Schizocosa include the 20 species recognized by Dondale & Redner plus S. rovn.eri Uetz & Dondale 1979, S. strP dulans Stratton 1984 and S. uetzi Stratton 1997 and at least one undescribed species. Males of several members of the genus have conspicuous ornamentation in the form of pig- mentation and/or bristles on the first legs of mature males. The ornamentation varies con- siderably from a complete lack of dark pig- ment [e.g., S. saltatrix (Hentz 1844), or S. rov- neri], to slight pigment on the tibia of males 347 348 THE JOURNAL OF ARACHNOLOGY (e.g., S. uetzi), to concentrated tufts of bristles at one end of the tibia (e.g., S, salsa Barnes 1953), to bristles that extend the length of the tibia and to the metatarsus [as in S. ocreata (Hentz 1844) from Florida]. Indeed, it is this variability in ornamentation that makes this genus particularly interesting for behavioral and evolutionary studies. For some species, the ornamentation has proven useful in spe- cies descriptions (Uetz & Dondale 1979; Stratton 1991, 1997a) as well as in mate choice studies (McClintock & Uetz 1996; Scheffer et ah 1996; Hebets & Uetz 1999, 2000). In some cases, tibial bristles are seen in species that have similar genital morphol- ogy, e.g., S. ocreata and S. crassipes, both commonly called “the brush-legged spider'% suggesting a common origin of the trait, but it is also seen in species with different geni- talia [e.g. compare S. ocreata with S. bilineata (Emerton 1885)], suggesting the possibility of independent origins. Closely related species may be very divergent with respect to orna- mentation, as is seen in Y. ocreata and S. rov- neri, two species long considered to be sibling to each other. Ornamentation and courtship behavior in Schizocosa wolf spiders have been the focus of studies addressing sexual selection and sig- nal evolution, fluctuating asymmetry and re- productive isolation (Hebets & Uetz 1999, 2000; Uetz et al, 1996, Uetz & Smith 1999; McClintock & Uetz 1996; Scheffer et ah 1996; Hebets 2003, 2005) . Perhaps most sur- prisingly, the ornamentation on at least one species (S. uetzi) appears to be important in a social learning context (Hebets 2003). Members of the genus Schizocosa use a va- riety of modes of communication during courtship, including chemical, seismic and vi- sual signals. In two comparative studies, He- bets & Uetz (1999, 2000) found that among six species of Schizocosa, females of three species exhibited a vibratory bias during courtship and three showed a visual bias. He- bets & Uetz (2000) found that in species with active visual displays but without ornamen- tation, an artificial increase in male ornamen- tation resulted in increased female receptivity. In an attempt to test the hypothesis that or- namentation evolved secondarily in this fam- ily to enhance pre-existing visual movement displays, they presented a summary of North American lycosid species which complied in- formation on the presence/absence of orna- mentation and leg waving displays. However, a phylogenetic study including more lycosid species, as in the present study, will provide a more thorough test of such a hypothesis. The mixture of ornamented species and non- or- namented species plus the complexity of ' courtship interactions makes the Schizocosa genus particularly well suited for testing ideas concerning multimodal signaling (Uetz & Roberts 2002 ; Hebets 2005) and the evolution of complex behaviors. The relative impor- tance of phytogeny compared with sexual se- lection can be assessed with a robust phylog- eny. McClintock & Uetz (1996) presented evi- dence that females of S. rovneri, a species without leg ornamentation, showed a higher level of response to video images of conspe- cific males that had been visually manipulated to have tufts on their tibia and to courting het- erospecific males than to controls (un-manip- ulated conspecific males). The preliminary phylogeny presented in their 1996 study (in- cluding 16 characters for 7 species) suggested that the female preference for ornamentations may have preceded the evolution of the or- naments themselves and thus be an example of the sensory bias hypothesis (Ryan & Rand 1993). A more complete phylogenetic study of the genus involving more species and more i characters will provide a more robust test of j the sensory bias hypothesis. i Here, I present the results of a comparison ; of the sexual ornamentation found in members ■ of Schizocosa and the results of a phyloge- ; netic study addressing the Nearctic members of this wolf spider genus. Using the results from my phylogenetic analysis, I address the ; following four hypotheses: 1. Ornamentation in the form of tibial bristles arose once within this genus; 2. Monophyletic groups show sim- ilarities in courtship behavior; 3. Schizocosa > ocreata and S. rovneri are sibling species as ; was suggested by Uetz & Dondale (1979) and i supported by the successful interbreeding re- ported in Stratton & Uetz (1986); and 4. Fi- nally, this study presents a test of the sensory bias hypothesis presented by McClintock & Uetz (1996) that female preference seen in S, rovneri females for males with ornamentation preceded the evolution of ornamentation in closely related species. STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 349 Figures 1-6. — Dorsal view of males of several of the species used in the phylogenetic study: 1. Allocosa georgicola\ 2. Gladicosa pulchra; 3. Schizocosa avida\ 4. Schizocosa retrorsa; 5. Schizocosa rovneri; 6. Schizocosa iietzi. METHODS Ornamentation and courtship behav- ior.— As comparative data on male ornamen- tation and courtship behavior were not avail- able for all species of Schizocosa, I initially measured the form and extent of male tibial bristles for all members of the genus and re- corded the courtship behavior for selected (available) species. To document the ornamen- tation of each species, a lateral view of the first pair of legs for males of all species in- cluded in the study was photographed using a dissecting microscope (Olympus SZX12) and dedicated digital camera (Olympus 750). Courtship behavior (for available species not reported in the literature) was documented by collecting subadult males and females, main- taining them in the laboratory until mature (see Stratton et al. 1996, Miller et al. 1998) and then videotaping courtship behavior. The standard protocol for recording behavior was as follows. Twenty-four hours before testing, females were fed an appropriately-sized crick- et and were placed in a recording chamber 350 THE JOURNAL OF ARACHNOLOGY STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 351 with filter paper as substrate. Males were in- troduced to the chamber and all interactions were recorded with either a Panasonic WD- 5000 camera with Kiron 105 mm f2,8 macro- lens on standard VHS tapes or by using a Sony TRV-22 taping to a mini-DV tape. Seis- mic recordings were made by using a sound transducer attached to an EG&G PARC Mod- el 113 pre-amp connecting to the video re- corder. Video and seismic recordings were made for the following species, whose behav- ior is not yet described in the literature (sum- marized in Table 5): S. avida, S. crassipal- pata, S. floridana, S. saltatrix, S. nr saltatrix, Gladicosa pulchra and Allocosa georgicola. As courtship behavior remains unknown for several species, the behavior was not used in the parsimony analysis. Phylogenetic study: choice of taxa and material examined. — There are 63 species currently listed in the genus Schizocosa (Plat- nick 2005) from all over the world. However, in this study, I chose to focus solely on the Nearctic species; explicit in the exclusion the species from the Philippines, South Africa and other localities outside of North America is the assumption that the Nearctic species are monophyletic, an hypothesis not tested in this study. However, the names of several large genera of wolf spiders (e.g., Lycosa, Schizo- cosa and others) were often categories of con- venience rather than hypotheses of relation- ship for some earlier workers. For example, recent taxonomic work of New Zealand wolf spiders suggests that previous placement of species into Holarctic genera is erroneous (Vink 2002) and Schizocosa berenice L. Koch 1877 from Australia actually belongs in the genus Artoria (Framenau pers. comm.). Additionally, I excluded species that were excluded from the genus by Dondale & Red- ner (1978). Many of these, e.g., S. incerta (Bryant 1934), S. perplexa Bryant 1936, 5". puebla Chamberlin 1925, S. tamae (Gertsch & Davis 1940), S. tristani (Banks 1909) and Av- icosa ceratiola (Gertsch & Wallace 1935) have yet to be placed in another genus but were judged to be outside the scope of this project. This study is thus based on direct exami- nation of preserved specimens of the 20 Schi- zocosa species recognized by Dondale & Red- ner (1978), plus S, rovneri, S. stridulans, S. uetzi and one undescribed species. In addition, I scored 3 populations of Schizocosa ocreata: one from Cincinnati, OH that has been used extensively in behavioral research (see refer- ences by Uetz, Hebets, McClintock and Strat- ton), a second from Central Mississippi (see Miller et al. 1998) and a third from Gaines- ville, Florida; and two populations of Schizo- cosa crassipes: (separate populations from Florida and Mississippi). Finally, I also scored individuals that are hybrids between S. ocrea- ta (OH) and S. rovneri (available from a pre- vious study [Stratton & Uetz 1986]). Appen- dix 2 summarizes the specimens of the 3 1 taxa used and their deposition. Voucher specimens from my own collection will be deposited at the AMNH. Outgroups. — The choice of outgroups for this study was difficult due to the incomplete knowledge of lycosid generic relationships. As the sister-group for the Schizocosa is not known, representatives from several different North American lycosid genera were scored and included in this analysis. Gladicosa pul- chra (Keyserling 1877), Allocosa georgicola (Walckenaer 1837) and Rabidosa rabida (Walckenaer 1837) were chosen as outgroups for the final analysis as they all share some characters with Schizocosa but also have clearly distinguishing features (locality, de- position and citations for figures are given in Appendix 2). Allocosa georgicola has affini- ties to the H. helluo species group (Brady pers. comm.). Choice of characters. — Forty nine mor- phological characters were found to be infor- mative and were scored by direct examination of specimens (data matrix in Appendix 1; de- tails of the characters are provided below). During the course of the study numerous ad- ditional characters were identified and scored Figures 7~10. — Venters of selected species; 7. Male of Schizocosa crassipes’, 8. Male of Schizocosa retrorsa; 9. Female of Schizocosa bilineata (epigynum has been dissected); 10. Female of Schizocosa avida. 352 THE JOURNAL OF ARACHNOLOGY STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 353 but discarded as too variable or uninformative. Characters were chosen by comparing the dif- ferent subfamilies of Lycosidae (Dondale 1986), by screening revisions, descriptions and illustrations of other wolf spiders and by direct examination of the specimens. Figures from the following references were found to be useful: Schizocosa (Dondale & Redner 1978), Trochosa (Brady 1979); Gladicosa (Brady 1986); Rabidosa (Brady & McKinley 1994); Geolycosa (Wallace 1942a); Hogna georgicola (Chamberlin & Ivie 1944) and Iso- hogna lenta (Wallace 1942b). As characters were treated as ueordered, hypotheses of polarity are an emergent prop- erty of the analysis. Forty characters are bi- nary; nine are multistate characters. Six of the characters are morphometric (characters 1, 17, 20, 33, 37, 39) with three being expressed as ratios (characters 17, 33, 37). To assign states for the morphometric data, I examined the data for gaps. In several cases, I opted for in- dependent coding of variables for a given structure as opposed to coding as a multistate character with linked states to minimize as- sumptions of congruence (e.g., characters 21- 25). However, a potential problem with this is the duplication of absences (Maddisoe 1993). For the final analysis, the 49 characters were grouped in the following manner: 19 so- matic characters, 14 male palpal characters, 10 female epigynal characters and six male sec- ondary sexual characters. Several additional characters (e.g., behavioral and ecological characters) were used a posteriori and mapped onto the resulting cladograms. As data for these latter characters were not available for all species, they were not used in the parsi- mony analysis (Platnick et al. 1991; Maddison 1993). Data analysis. — Maximum parsimony analyses were conducted using PAUP* (Ver- sion 4.0b 10) (Swofford 2002) with 1000 ran- dom starting point heuristic searches using Stepwise-addition option and 1000 random taxon addition sequences and tree-bisection- reconnection (TBR) branch swapping. The re- sults from other swapping algorithims were compared to TBR branch swapping. All char- acters were unordered and several weighting schemes were examined. The trees were root- ed by setting the three non-congeners as out- groups. Characters were mapped using MacClade Version 3.0 (Maddison & Maddi- son 1992). The degree of internal support for the resulting clades was estimated using boot- strap analysis (using 100 random addition se- quence replicates and 100 bootstrap repli- cates). Weighting options. — The data set was first analyzed with all characters weighted equally. However, preliminary analyses suggested there was little phylogenetic signal in the so- matic characters. Subsequent analysis inves- tigated several different weighting options in- cluding reweighting with the rescaled consistency index as well as weighting the genitalic characters more heavily. As most modern students of wolf spiders use genitalic characters extensively in revisions and de- scriptions (Dondale & Render 1978, 1990; Brady 1962, 1979, 1986; Brady & McKinley 1994) and because wolf spiders appear to be very conservative in their somatic morpholo- gy, I favor the trees produced by weighting the genitalic characters more heavily. In the final analysis, I used the following weighting scheme: somatic characters weighted as “2”; genitalic characters weighted as “3” and sec- ondary sexual characters weighted as “1.” Fi- nally, as I was interested in the evolution of secondary sexual characters, I compared trees produced by excluding secondary sexual char- acters with trees including those characters. Description of characters. — Somatic char- acters: 1. Body size: 0 = medium; 1 — large; 2 — small. Measurements of the carapace were made as the best representation of body size, as unlike the spider’s abdomen, it does not vary with a recent meal. Carapace length and width were measured dorsally using an ocular micrometer. The ratio of the carapace length to width was similar across all species used in this study (Table 1). Carapace length Figures 1 1-14, — Left pedipalps of males from selected species of Schizocosa. Localities are indicated in Appendix 2: 11, S. avida; 12. S. retrorsa; 13. S. saltatrix; 14, S. duplex. ELL = ear-like lobe, IPE = intromittent portion of embolis, MA = median apophysis, TA = terminal apophysis. Scale lines = 500 jxm. 354 THE JOURNAL OF ARACHNOLOGY Table 1 . — Body size (carapace length, carapace width and ratio of length/width) of male and female j Schizocosa and outgroups used in this study. All measures are in mm. Measures of female 5'. aulonia, \ and S. maxima are from Dondale & Redner (1978), measures from S. salsa are from Barnes (1952). I Species Male Female Length Width LAV Length Width LAV S. aulonia 4.0 2.8 1.4 5.5 4.0 1.4 S. avida 4.8 3.6 1.3 6.8 5.0 1.4 S. bilineata 2.9 2.1 1.4 3.5 2.5 1.4 S. ce spit urn 3.6 2.6 1.4 4.2 3.0 1.4 S. chiricahua 4.2 3.0 1.4 3.8 2.5 1.5 S. communis 4.6 3.4 1.4 4.0 3.0 1.3 S. crassipalpata 3.1 2.3 1.3 3.1 2.2 1.4 S. crassipes (FL) 3.2 2.6 1.2 3.8 2.7 1.4 S. crassipes (MS) 3.2 2.4 1.3 3.4 2.6 1.3 S. duplex 3.2 2.5 1.3 3.2 2.4 1.3 S. fioridana 2.6 2.0 1.3 3.1 2.5 1.2 S. humilis 3.1 2.3 1.3 3.6 2.6 1.4 S. maxima 9.0 7.7 1.2 12.1 9.0 1.3 S. mccooki 3.7 2.8 1.3 4.2 3.2 1.3 S. mimula 4.0 2.9 1.4 4.2 2.8 1.5 S. minnesotensis 4.6 3.2 1.4 4.9 3.6 1.4 S. ocreata (OH) 3.8 2.8 1.4 4.1 3.2 1.3 S. ocreata (MS) 3.7 2.7 1.4 4.5 3.4 1.3 S. ocreata (FL) 4.0 3.0 1.3 4.2 3.3 1.3 S. retrorsa 3.7 2.7 1.4 3.4 2.7 1.3 S. rovneri 3.4 2.6 1.3 3.7 3.0 1.2 S. salsa 4.1 3.0 1.4 4.1 2.9 1.4 S. saltatrix 3.7 3.0 1.2 3.6 3.0 1.2 S. S. sp. nr. saltatrix 4.0 3.2 1.3 3.6 2.8 1.3 S. segregata 3.0 2.2 1.4 3.2 2.2 1.5 S. stridulans 3.2 2.4 1.3 3.0 2.3 1.3 S. uetzi 3.6 2.8 1.3 3.7 3.0 1.2 S. ocr X rov hybrids 3.8 3.1 1.2 4.3 3.5 1.2 A. georgicola 8.0 6.9 1.2 10.0 7.5 1.3 G. pulchra 6.0 4.5 1.3 6.4 5.0 1.3 R. rabida 7.8 5.9 1.3 9.8 7.4 1.3 as states: carapace length > 4.5 mm for ‘Targe”, carapace length between 3.5 and 4.5 mm for “medium” and carapace length <3.5 mm for “small”. 2. Median band (MB) width: 0 = narrow; 1 = medium; 2 = broad. The median band is a light band on the dorsal surface of the car- apace. It may be a thin line (as in A. georgi- cola. Fig. 1, which is distinctly thinner than the posterior median eyes) or it may be “broad,” or as broad as the posterior median eyes (as in G. pulchra, Fig. 2, and members of Schizocosa, Figs. 3-6). A medium-sized median band is present in Rabidosa spp. (Bra- dy and McKinley 1994, p. 140, figs. 1-5). 3. Median band edges: 0 = straight; 1 = not straight. The edges of the median band can be straight as in Fig. 1, 3, 5, 6), or can have some constriction as is seen in G. pulchra or : S. retrorsa Banks 1911 (Figs. 2, 4). : 4. Submarginal band (SMB), edges: 0 = ab- sent; 1 = smooth; 2 = wavy or spots. The submarginal band is a light band near the edg- j es of the carapace. It can be present as either a relatively smooth band (as in S. avida and S. rovneri. Figs. 3, 5) or diffuse or lacking as in A. georgicola (Fig. 1). In S. retrorsa (Fig. 4) the SMB consists of spots at the edges of the carapace. 5. Submarginal band, size: 0 = absent; 1 = narrow; 2 = broad. In most species of Schb zocosa, the SMB is narrow or absent. In Ra- bidosa spp. (Brady & McKinley 1994, figs. 1-5, p. 140) and in S. salsa, the SMB is wider STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 355 Figures 15-18. — Pedipalps of males from selected populations of Schizocosa ocreata and S. crassipes. Localities as indicated in Appendix 2; 15. S. ocreata (FL); 16. S. ocreata (MS); 17. S. crassipes (FL); 18. S. crassipes (MS). Scale lines = 500 |jLm. 356 THE JOURNAL OF ARACHNOLOGY than the median band and scored as broad (Barnes 1953, fig. 16). 6. Heart mark (HM): 0 = absent or faint; 1 = strong. In some Lycosinae, the heart mark is a darkened region on the dorsum of the ab- domen where the heart is located (e.g. A. georgicola & S. avida. Figs. 1, 3). 7. Light bands on abdomen: 0 = absent; 1 = present. In some cases, the HM is appar- ently accentuated by the presence of white lines on either side of the HM (as in S. avida. Fig. 3). 8. Sternum color: 0 = yellow or light brown; 1 == orange, dark brown or black, 9. Bands on sternum: 0 = absent; 1 = pre- sent. Sternum bands, as used for this study, are longitudinal bands that extend from the anterior to the posterior end of the sternum. 10. Shield on venter: 0 = absent; 1 = pre- sent. When present, this is a light patch on black background on the venter of the animal as in S. avida (Fig. 10). The shape of the light patch varies between individuals from the same geographic area (Stratton, unpublished data). 1 1. Ventral color of abdomen: 0 = light or mostly light; 1 = black or mostly black; 2 = light shield on black. Light or mostly light is as in S. crassipes or S. bilineata (Emerton 1885) (Figs. 7, 9); black or mostly black, as in S. retrorsa (Fig. 8); and light shield on black as in S. avida (Fig. 10). 12 venter: 0 = absent; 1 == present. The “V” is formed by dark pigmented spots near the lateral edges of the venter {S. bili- neata, Fig. 12). The inner “V” on S. bilineata is not pigmented and originates in points of muscle attachment. 13. Spots on venter: 0 = absent; 1 = pre- sent. Similar to character 12 but in some cases and in S. crassipes, the spots on the venter are scattered. 14. Color of coxae: 0 = light to light brown; 2 = dark brown. 15. Color of coxae relative to the femur: 0 same as femora; 1 = lighter than femur; 2 = darker than femur. 16. Dark lines on chelicerae: 0 = absent; 1 = present. When present, these are vertical lines running the length of the chelicerae when viewed from the front. 17. Relative tibial length of males: 0 = short to medium; 1 = long. Examination of the specimens suggested that some of the spe- cies had particularly stout legs, others were relatively long-legged, while many fell be- ! tween these extremes (also noted by Krones- | tedt 1990). Since there was a wide range of 1 body sizes in the examined species, I took a ; relative measure for leg length. Tibia 1 (mea- ' sured with an ocular micrometer along the dorsal edge) divided by the carapace length | was taken as the relative measure of leg length * that could be potentially informative. Values I for this ratio ranged from 0.5 for S. cespitum Dondale & Redner 1978 to 0.92 for S. salsa (Table 2). Gaps were identified by examina- [ tion of the distribution of values. These data I were first collapsed into a binary coding (0 = I short to medium, tl/cl < 0,71 and 1 = long ' tl/cl > 0.71). The data were also scored as a ; multistate character with any gap > 0.03 used ' to define states. With this coding there were ; seven character states. During character ex- i ploration, analyses were run both ways; this character was judged to have a large amount i of homoplasy and in the final analysis, I used ■ the binary coding. j 18. Femur I annulations (male): 0 = absent; J 1 = present. 1 19. Femur II-IV annulations (male): 0 = ab- ' sent; 1 = present. ; Male palpal characters: Palpal structures ) have been successfully employed in spider i taxonomy and have been central in wolf spi- i der revisions. Eberhard (1994) suggested that the complexity of genitalic characters has been shaped by sexual selection. The com- f plexity of the pedipalps has made determina- 1 tion of homologies of structures between fam- \ ilies difficult. Even within a single family, j such as the Lycosidae, terms have been used i in a variety of ways. I here follow the termi- I nology suggested by Dondale & Redner i (1978) with comparisons to Brady (1986) and '■ have focused primarily on structures that aid I in the determination of species. ; The cymbium of the tarsus (terminal seg- ment) forms the body of the tarsus of the ped- , ipalp and holds the sclerites (genital bulb) in- volved in copulation. At the tip of the cymbium there may be macrosetae (MS, Fig. : 12), which function to hold the pedipalp in j place while the spider stridulates [e.g., S. sab tatrix and R, rabida (Walckenaer 1837), Rov- ner 1975]. The genital bulb is an inter-con- \ nected assemblage of sclerites and distensible sacs (the hematodocha). During copulation, STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 357 Figures 19-22.— Epigyna of selected species. Localities as indicated in Appendix 2: 19. S. avida, 20. S. retrorsa, 21. S. crassipes (MS); 22. S, duplex. MS = median septum, TP = transverse piece. Scale lines = 250 pm. the hematodocha becomes visible as an ex= panding sac of haemolymph. The cymbium is attached to the palpal tibia (Tib, Fig. 14). Located on the genital bulb is the terminal apophysis (TA, Fig. 11) which in Schizocosa is a small, free sclerite near the base of the iritromitteet portion of the embolus (IPE) (Figs. 11-18; see also Dondale & Redner 358 THE JOURNAL OF ARACHNOLOGY Table 2. — Tibial length and relative tibial length (tibial L/carapace L) of males of Schizocosa and outgroups. All measures are in mm. Species Male Cara length Tibia length Tibial L/ Cara L S. aulonia 4.0 2.5 0.61 S. avida 4.8 3.6 0.75 S. bilineata 2.9 2.0 0.67 S. cespitum 3.6 2.0 0.50 S. chiricahua 4.2 2.5 0.60 5, communis 4.6 2.5 0.68 S. crassipalpata 3.1 1.8 0.57 S. crassipes (FL) 3.2 2.8 0.85 S. crassipes (MS) 3.2 2.5 0.84 S. duplex 3.4 2 0.66 S. floridana 2.6 1.5 0.58 S. humilis 3.1 1.8 0.58 S. maxima 9.0 6.4 0.71 S. mccooki 3.7 2.7 0.74 S. mimula 4.0 2.7 0.68 S. minnesotensis 4.6 2.4 0.51 S. ocreata (OH) 3.8 3 0.78 S. ocreata (MS) 3.7 2.6 0.71 S. ocreata (FL) 4.0 2.6 0.65 S. retrorsa 3.7 2.6 0.71 S. rovneri 3.4 2.6 0.76 S. salsa 4.1 3.8 0.92 S. saltatrix 3.7 2.5 0.67 S. S. sp. nr. saltatrix 4.0 2.5 0.62 S. segregate 3.0 1.8 0.59 S. stridulans 3.2 2.4 0.75 S. uetzi 3.6 3.0 0.83 S. ocr X rov hybrids 3.8 2.6 0.68 A. georgicola 8.0 5.5 0.68 G. pulchra 6.0 4.8 0.83 R. rabida 7.8 7.0 0.89 1978). In Gladicosa and Hogna, the TA is an elongate structure that parallels the embolus and may assist in the proper placement of the IPE in the female epigynuro, (see Brady 1986, p. 314, fig. 41). Also clearly visible is the me- dian apophysis (MA, Figs. 11-18) which is directed retrolaterally and has a sinuous chan- nel on the dorsal surface, which are defining characters of the subfamily Lycosinae. A spur of the median apophysis engages the hood of the female epigynum during copulation. The palea is a partly sclerotized pad at the distal end of the genital bulb that sometimes bears processes or extensions (PPR) (Figs. 14-18). An ear-like lobe (ELL, Fig. 1 1) is also present in Schizocosa and Gladicosa (called the con- ductor by Dondale & Redner 1978). The base of the cymbium holds the scraper portion of the stridulatory structure, not visible in the ventral view shown. The tip of the tibia has the file of the stridulatory organ, found only in mature males. 20. Palpal tibia: 0 = length > width; 1 = length < width. Measured ventrally using an ocular micrometer. For some species (i.e., A. georgicola), the tibia of the pedipalp is long and thin (nearly twice as long as wide) (Table 3). For others, as in S. ocreata (FL), it is very stout and is wider than long (Table 3). A short, stout tibial palp is probably related to muscle mass needed for stridulation during courtship (Rovner 1975). Although all members of the Schizocosa and outgroups included here pos- sess a stridulatory organ, for some, palpal drumming is an important component of the courtship behavior (Table 5; see discussion below). 21. Terminal apophysis (TA): 0 = present; 1 = small or absent. 22. Terminal apophysis, size: 0 = elongate; 1 = small or absent. The TA in Schizocosa is generally present but relatively small com- pared to that in the outgroups, A. georgicola and G. pulchra (Figs. 1 1-18; compare to, Bra- dy 1986, p. 314, fig. 41), where the TA is an elongate structure. 23. Terminal apophysis with slight arch: 0 = absent; 1 = present. The TA may form a slight arch, as seen in Figs. 15 and 16. 24. Terminal apophysis, with tear-drop shape: 0 = absent; 1 = present. See fig. 13 in Dondale & Redner (1978). 25. Terminal apophysis as inverted triangle: 0 = absent; 1 = present. 26. Embolus, basic shape: 0 = hair-like; 1 = sword-like. All pedipalps in Figs. 11-18 show a hair-like embolus. The sword-like em- bolus is found in Gladicosa, whose name re- fers to a sword (Brady 1986, p. 314, fig. 41). 27. Width of intromittent portion of embo- lus (IPE, Fig. 11): 0 == thin; 1 = stout. 28. Palea of pedipalp with triangular pro- cess: 0 = absent; 1 = present. In a few spe- cies, notably S. duplex, S. saltarix and an un- described species, the pedipalp has a triangular shaped process (Fig. 13, 14). 29. Palea of pedipalp with finger-like pro- cess: 0 = absent; 1 = present. A finger-like process (PPR) is shown in Fig. 15-18. During copulation this process pushes against the cu- STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 359 Figures 23-31. — Shape and size of tibial bristles in all taxa with the tibial bristles in the Nearctic Schizocosa: 23, S. aulonia; 24. S. salsa; 25. S. bilineata; 26. S. segregata; 27. S. crassipes (MS); 28. S. stridulans; 29. S. ocreata (FL); 30. S. ocreata (OH) ; 31. S. ocreata (MS). tide to the side of the female epigynum (Strata ton, unpublished data). 30. Number of macrosetae: 0 = few (4-9); 1 = many (10 or more). The macrosetae can be found on the tip of the cymbium. Often there is a dense cluster, but sometimes they are much fewer and easily counted. 31. Long hamlike setae between macrose- tae and genital bulb: 0 = absent; 1 = present. 32. Relative size of macrosetae: 0 = thin; 1 = stout or very stout. 33. Cymbium (tarsus) of pedipalp, length to width: 0 = short to regular; 1 = long and thin. Measured ventrally with an ocular micrometer (Table 3). As with other quantitative charac- ters, the data were examined for gaps. During character analysis, this character was first scored as a multistate character but for the fi- nal analysis, it was treated as a binary char- acter. The species that showed palpal drum- ming during courtship had relatively long, thin pedipalps (e.g. S. avida, S. retrorsa, S. mccooki, and S. communis; Tables 3 & 5). Female epigynal characters: The promi- nent features of the female epigynum include the well sclerotized median septum (MS, Figs. 19-22), the posterior transverse piece (TP, Fig. 19). The edges of the median septum can be flared at the posterior end, parallel, or can be widest at the center. The anterior end of the median septum has a funnel-like hood (Hood, Fig. 19) that generally is double but is single in A bilineata and S. crassipalpata Roewer 1951. The median apophysis of the male “catches” on the hood of the female epigyn- um during the first stages of copulation; this serves to brace the pedipalp and immediately following the engagement of the MA with the hood, one can see the expansion of the he- matodocha of the male pedipalp (Dondale & Redner 1978; Stratton, unpublished data) The depth of the hood varies between species and 360 THE JOURNAL OF ARACHNOLOGY Table 3. — Length, width and length/width of tibia and tarsus of pedipalp for Schizocosa and outgroups. All measures in mm. Species Tibia of palp Tarsus of palp Length Width LAV Length Width LAV S. aulonia 0.2 0.2 1.2 LI 0.5 2.0 S. avida 0.7 0.5 1.5 0.9 0.4 2.4 S. bilineata 0.4 0.4 1.0 0.9 0.5 1.8 S. cespitum 0.5 0.4 1.3 1.2 0.6 2.1 S. chiricahua 0.5 0.4 1.4 1.2 0.6 2.1 S. communis 0.6 0.4 1.4 1.3 0.6 2.1 S. crassipalpata 0.4 0.3 1.1 0.9 0.5 1.7 S. crassipes (FL) 0.3 0.3 1.0 0.7 0.4 1.9 S. crassipes (MS) 0.6 0.5 1.1 1.1 0.6 1.9 S. duplex 0.4 0.5 0.9 1.0 0.5 1.9 S. floridana 0.5 0.4 1.3 0.9 0.5 1.9 S. humilis 0.5 0.5 1.1 1.0 0.6 1.7 S. maxima 0.4 0.3 1.3 1.0 0.6 1.6 S. mccooki 0.6 0.4 1.4 1.2 0.6 2.1 S. mimula 0.3 0.2 1.5 1.1 0.6 2.1 S. minnesotensis 0.7 0.5 1.4 1.5 0.7 2.2 S. ocreata (OH) 0.6 0.6 1.0 1.4 0.8 1.9 5. ocreata (FL) 0.5 0.6 0.8 1.3 0.8 1.7 S. ocreata (MS) 0.6 0.6 1.0 1.3 0.7 1.8 S. retrorsa 0.6 0.4 1.3 1.3 0.6 2.1 S. rovneri 0.6 0.6 0.9 1.3 0.7 1.8 S. salsa 0.7 0.5 1.3 1.3 0.6 2.3 S. saltatrix 0.7 0.6 1.1 1.1 0.6 1.8 S. sp. nr saltatrix 0.7 0.6 1.1 1.2 0.7 1.8 S. segregata 0.4 0.4 1.1 0.9 0.5 1.9 S. stridulans 0.5 0.5 1.0 1.1 0.6 1.8 S. uetzi 0.6 0.6 1.1 1.3 0.7 1.9 S. ocr. X rov hybrids 0.6 0.6 1.0 1.1 0.6 1.8 A. georgicola 1.6 0.8 1.9 2.7 1.1 2.4 G. pulchra 1.0 0.6 1.6 2.1 1.1 1.9 R. rabida 1.7 0.8 2.1 2.9 1.1 2.6 is much deeper in G. pulchra and A. georgi- cola than most of the Schizocosa species (Ta- ble 4; compare Figs. 19-22). The transverse piece (TP) in Schizocosa is either truncate as in S. avida and S. retrorsa (Figs. 19 & 20) or has excavations (Figs. 21 & 22). The excavations can be located at the lateral edges of the transverse piece as seen in S. duplex (Fig. 22) and S. saltatrix, or the ex- cavations can be almost touching near the cen- ter of the transverse piece as seen in S. uetzi (see Stratton 1997a). More often, the location of the excavations are somewhat between these extremes. In lycosids, the copulatory opening is near lateral edges of the transverse piece and the genital opening (where the eggs leave the gen- ital tract) is in the epigastric furrow. When the female genitalia are dissected and the dorsal side is examined, the spermathecae are the dominant structures (see Stratton 1997, fig. 5). 34. Edges of median septum (MS): 0 = MS widest at center; 1 = edges of MS parallel; 2 = MS widest at base. For MS widest at center, see Figs. 19 and 20 (S. avida and S, retrorsa). Schizocosa duplex provides an example of the parallel edges (Fig. 22) while S. crassipes shows a MS widest at its base (Fig. 21). 35. Excavations on transverse piece of epi- gynum: 0 = absent; 1 = present; see Figs 19- 22. 36. Separations of excavations: 0 = absent; 1 = widely separated with excavations near lateral edge; 2 = intermediate in placement; 3 = narrowly separated. The excavations were considered widely separated if a hypothetical STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 361 S. ocreata Clade 56 Clade A ► 89 Clade B Clade C S. crassipes FL S. fioridana S. uetzi S. stridulans ocr OH X rov hybrid S. rovneri S. ocreata OH S. crassipes MS S. ocreata FL S. ocreata MS S. duplex S. saltatrix S. sp. nr saltatrix S. segregate S= biiineata L— S. crassipalpata — S. humilis 52r— S. avida — S. mccooki — S. aulonia S. communis — S. maxima - — — S. cespitum — S. retrorsa 56 r — S. chiricahua L — - S. salsa — S. mimula — S. minnesotensis — G. pulchra — H. georgicola — R. rabida Figure 32. — Single most parsimonious tree from heuristic searches including all taxa, excluding secondary sexual characters and applying preferred weighting (somatic characters = “2,” genitalic characters = “3.” The tree shows clades A, B, & C as well as the S. ocreata clade. Bootstrap values above 50% are shown. 362 THE JOURNAL OF ARACHNOLOGY additional excavation could fit between the excavations. 37. Ratio of median septum to transverse piece: 0 = MS/TP < 0.72; 1 = MS/TP < 1.179; 2 - MS/TP < 1.357; 3 = MS/TP < 1.52; 4 = MS/TP = 1.583. As with other quantitative characters, the ratio of MS to TP was sorted, graphed and examined for gaps. The ratio of the overall length to width was also examined but for this latter ratio, there were no clear gaps. 38. Epigynal hood, single or double: 0 = double; 1= single. Figs 19-22 show a double hood; a single hood seen in S. bilineata and S. crassipalpata (Dondale & Redner 1978; figs. 47 & 49). 39. Depth of epigynal hood: 0 = deep; 1 = shallow. See Table 4. 40. Spermathecae, shape: 0 = rounded; 1 = elongate; 2 pointed. 41. Spermathecae texture: 0 = smooth; 1 = bumpy. The “bumpy” texture was visible with light microscopy; the function of these bumps is unknown. 42. Copulatory tube: 0 = simple; 1 = com- plex. The copulatory ducts for Schizocosa (Dondale & Redner 1978; e.g. figs. 28, 32 & 37) have a single elbow and were scored as simple. The copulatory duct in R. rabida (Bra- dy & McKinley 1994; hg. 13) is convoluted and was scored as complex. 43. Pigment around the epigynum: 0 = ab- sent; 1 = present. In some species, there is a distinct “box” of dark pigment surrounding the epigynum. Male secondary sexual characters (orna- mentation).- Male secondary sexual characters include both pigmentation on the first legs and bristles or brushes, primarily on the tibia of legs 1. Work by Stratton & Uetz (1986) showed that through hybridization of S. ocrea- ta and S. rovneri these characters are inherited independently. 44. Femur I with dark stripe: 0 = absent; 1 = present. 45. Femur I with dark pigment: 0 = absent; 1 = present. 46. Femur II-IV with longitudinal stripe: 0 = absent; 1 = present. 47. Presence of bristles on tibia: 0 = absent; 1 — present (see Figs. 23-31 for examples of species with bristles). 48. Dark pigment covering tibia I: 0 = ab- sent; 1 = present. 49. Metatarsus I bristles: 0 = absent; 1 = present (e.g., S. ocreata, FL, Fig. 29). RESULTS Sexual ornamentation. — Sexual ornamen- tation in the Nearctic genus Schizocosa con- sists of pigmentation on all or part of the legs I of mature males and/or bristles which are generally limited to all or part of the tibia of legs L Dark pigmentation can be limited to the femur of leg I (as in S. retrorsa. Fig. 4), or to part of the femur and extending to the tibia (as in S. stridulans) (Fig. 28), or may be limited to the tibia (as in S. uetzi. Fig. 35), or extend from the patella to the metatarsus (as in S. floridana Bryant 1934, Fig. 35). In some cases, such as S. humilis and S, retrorsa, the dark pigmentation of the femur contrasts sharply with light hairs found on the tibia. Dark pigmentation can occur without bristles, as is seen in S. retrorsa, S. uetzi and in the outgroup, Rabidosa rabida. When present, the bristles are always as- sociated with the tibia of legs I of mature males but may be found only on the distal end of the tibia (as in S. salsa. Fig. 24), may ex- tend to much of the metatarsus, as is seen in S. ocreata from Florida (Fig. 29), or may be limited to the tibia as is seen in S. bilineata (Fig. 25). In S. aulonia Dondale 1969 and S. segregata, the bristles are longest along the ventral side of the tibia (Figs. 23, 26), while in S. crassipes and S. ocreata (Figs. 27, 29, 30, 31), the bristles extend both dorsally and ventrally, providing a large rectangular ap- pearance when viewed from the side. In gen- eral, the bristles are largest when viewed from the side, at eye-level with the spider. A dis- cussion of the phylogenetic distribution of the bristles follows the presentation of the pre- ferred phylogeny. Comparison of courtship behavior* — Courtship behavior for 20 of the 31 taxa rep- resented in this study has been documented (summarized in Table 5), including all mem- bers of the S. ocreata clade, as well as the hybrids between S. ocreata (OH) and S. rov- neri. All species studied to date show chemo- exploration by males in the presence of fe- males or female silk (indicating the presence of chemical signals) and males of all species produce seismic signals either by stridulation by the male palp (first described by Rovner 1975), palpal drumming (Stratton & Lowrie STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 363 Femur with dark pigment (male) □ absent ■i present M equivocal Dark pigment covering tibia (males) □ absent ■i present S. crassipes FL S. floridana S. uetzi S. stridulans S. fO¥neri ocr OH X rov hybrid S. ocreata OH S. crassipes MS S. ocreata FL S. ocreata MS S. duplex S. saltatrix S. nr. saltatrix S. segregate S. bilineata S. crassipalpata S. humilis S. Bvida S. mccooki S. aulonia S. communis S. maxima S. cespitum S. retrorsa S. chiricahua S. saisa S. mimula S. minnesotensis G. pulchra A. georgicola R. rabida S. crassipes FL S. floridana S. uetzi S. stridulans S. rovneri ocr OH X rov hybrid S. ocreata OH S. crassipes MS S. ocreata FL S. ocreata MS S. duplex S. saltatrix S. nr. saltatrix S. segregate S. bilineata S. crassipalpata S. humilis S. avida S. mccooki S. aulonia S. communis S. maxima S. cespitum S. retrorsa S. chiricahua S. salsa S. mimula S. minnesotensis G. pulchra A. georgicola R. rabida Figure 33. — Mapping of the ornamentation characters “pigment on femur” and “pigment on tibia, seen in mature males on the preferred phylogeny. 364 THE JOURNAL OF ARACHNOLOGY 1984; Hebets et al. 1996), body vibration or some combination of these (Table 5). Many species (but not all) also have some movement that appears to be a visual signal produced by males during courtship. Several species incor- porate an arch of legs 1 (scored as a “ + ” in Fig. 36). For example, this is seen in S. avida, and S. saltatrix as well as S. ocreata (OH) (Table 5). Other species have movements that appear to be more intense visual signals. For example, S. retrorsa males have a vigorous leg 1 wave that is associated with courtship (Hebets et al. 1996), and S. ocreata (FL) has a bilateral double arch of the first pair of legs. These more overt visual signals are scored as “ + + ” in Fig. 36. One species that apparently lacks visual signals is S. duplex. All popula- tions of S. ocreata and S. crassipes have leg arching plus either leg waving or tapping. When the different populations of S. crassipes were compared, each population showed sim- ilar elements of courtship but differed in the proportion of time spent doing each behavior and in the sequences of behavior (e.g.. Miller et al. 1998; Germano et ah, unpublished data). Kaston (1936) included S. bilineata in his comparative courtship study but he did not see any behaviors preceding copulation in this species. Finally, the behavior of several spe- cies remains unknown and it is hoped that this study may stimulate interest in the behavior and ecology of these species. Results of phylogenetic analysis. — Parsi- mony analysis with all characters weighted equally and trees not rooted resulted in very little resolution of clades. The agreement sub- tree for these trees did not include the three outgroups, suggesting their position in the tree was not stable. When the preferred weighting option was applied and the trees rooted with the three outgroups, there was more resolu- tion. A single most parsimonious was tree re- sulted from the analysis with the secondary sexual characters excluded and is the preferred tree presented here (Fig. 32; bootstrap values above 50% shown on figures; 570 steps, Con- sistency Index = 0.271, Retention Index = 0.567, Homoplasy Index = 0.729). When sec- ondary sexual characters were included, a consensus of seven trees had 578 steps (Con- sistency Index = 0.265, Homoplasy Index = 0.735 and Retention Index = 0.549) and when compared to the preferred tree, differed only Table 4, — Epigynal measures including total length of epigynum, and width of epigynum, the ratio of epigynal length/width, and depth of the epi- gynal hood. All measures in mm. Values for three species {S. salsa, S. segregata, S. aulonia) were not available and for these species, ratio of length to width was calculated from published figures in Dondale & Redner (1978). Species Length Width LAV Hood S. aulonia 1.4 S. avida 0.9 0.7 1.3 0.3 S. bilineata 0.5 0.6 0.9 0.1 S. cespitum 0.7 0.7 1.0 0.1 S. chiricahua 0.7 0.6 1.2 0.1 S. communis 0.8 0.7 1.1 0.1 S. crassipalpata 0.5 0.5 0.9 0.1 S. crassipes (FL) 0.7 0.6 1.2 0.1 S. crassipes (MS) 0.7 0.6 1.3 0.1 S. duplex 0.6 0.6 1.0 0.1 S. floridana 0.6 0.5 1.2 0.1 S. humilis 0.8 0.6 1.3 0.1 S. maxima 0.8 0.7 1.2 0.1 S. mccooki 0.7 0.7 1.0 0.1 S. mi mu la 0.8 0.6 1.3 0.1 S. minnesotensis 0.9 0.6 1.7 0.1 S. ocreata (OH) 0.8 0.7 1.1 0.1 S. ocreata (FL) 0.9 0.7 1.3 0.1 S. ocreata (MS) 0.8 0.6 1.3 0.1 S. retrorsa 0.8 0.6 1.3 0.1 S. rovneri 0.8 0.7 1.1 0.1 S. salsa 1.1 S. saltatrix 0.7 0.7 1.1 0.1 S. sp. nr. saltatrix 0.6 0.7 1.0 0.1 S. segregata 1.2 S. stridulans 0.6 0.6 1.1 0.1 S. uetzi 0.8 0.7 1.2 0.1 S. ocr X rov hybrids 0.7 0.6 1.2 0.1 A. georgicola 1.4 1.0 1.4 0.2 G. pulchra 0.9 0.8 1.1 0.2 R. rabida 1.3 1.0 1.3 1.8 in the positions of S. aulonia and S. maxima Dondale & Redner 1978. The analysis resulted in three main clades, with S. minnesotensis basal to these clades. The largest clade (Fig. 32, Clade A) is com- prised of species found in the eastern half of North America; the range of some of the spe- cies is limited to the Northeast, some to the Southeast and some ranges extend to the Mid- west (see range maps, Dondale & Redner 1978; Stratton 1991, 1997a). Clade A includes 9 of the 11 taxa with tibial bristles (Fig. 34). Nested within Clade A is a clade that includes S. duplex, S. saltatrix and S. nr. saltatrix plus STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 365 tibia! bristles unordered □ absent HI present equivocal S. crassipes FL S. floridana S. uetzi S. striduians S. rovneri ocr OH X rov hybrid S. ocreata OH S. crassipes MS S. ocreata FL S. ocreata MS ® S. duplex S. saltatrix S. nr. saltatrix S. segregata S. biiineata S. crassipaipata S. humilis S. avid a S. mccooki S. auionia S. communis S. maxima S. cespitum S. retrorsa S. chiricahua S. saisa 30 S. mimula S. minnesotensis G. pulchra A. georgicola - S. aufonia ^ R. rabida Figure 34. — Mapping of representative tibial bristles, seen in mature males, on the preferred phylogeny. the S. ocreata clade, all defined by a process on the palea of the pedipalp of the male (either a triangular process, Fig. 14, or a finger-like process, Fig. 15) and all found in deciduous woods, or a mix of deciduous and pine forests. Also within Clade A is a clade that includes S. biiineata + S. crassipaipata, one of the few clades with significant support from the boot- strap analysis and defined by six characters shared in common. 366 THE JOURNAL OF ARACHNOLOGY Also nested within Clade A is the S. ocrea- ta clade, which consists of S. ocreata (multi- ple populations), S. crassipes (multiple pop- ulations), S. rovneri, S, stridulans, S. uetzi and S. floridana. The clade is unified by the very distinctive finger-like process on the palea of the male pedipalp (Character 29), sternum col- or (Character 8), long setae on the ventral sur- face of the pedipalp (Character 31), and the relative size of the macrosetae (Character 32). Of these characters, only Character 29 is not lost or reversed. Members of this clade have been the most intensively studied with respect to courtship and mating behavior. Two of the taxa in this clade are paraphyletic {S. ocreata and 5. crassipes), suggesting that there are possibly multiple (yet unrecognized) cryptic species. Clade B includes the widespread eastern S. avida, the western S. mccooki, as well as S, maxima, S. aulonia, S. cespitum, S. communis and S, retrorsa. The third clade, Clade C include S. chira- cahua, S. mimula and S. salsa. The former two are western, while S. salsa is found on the Gulf Coast and Atlantic coast. Mapping ornamentation and behavior on preferred phylogenies. — An examination of the distribution of pigmentation on the first legs of males and bristles on the first pair of legs in males across all taxa in this study sug- gests that pigmentation may have evolved in- dependently of the bristles, and that both traits can be gained and lost (Figs. 33, 34). For ex- ample, pigmentation without bristles is seen in S. retrorsa, S. uetzi and R. rabida, while bristles without pigmentation is seen in S. bib ineata. When the tibial bristles are mapped onto the preferred phytogeny (Fig, 34) it ap- pears that it evolved independently five or six times with two or three losses in the S. ocrea- ta clade. A more detailed comparison of or- namentation on legs of males in the S. ocreata clade is presented in Fig. 35. It is clear that in this clade, the full range of pigmentation and bristles occur. Mapping the major patterns of seismic communication (i.e., drumming vs. stridula- tion) onto the preferred cladogram shows that the palpal drumming is concentrated in Clade B, while stridulation is prevalent in Clade A (Fig. 36). The S. ocreata clade is of particular interest with respect to courtship behavior as within the clade there is the full range of sec- ondary sexual characters (full tibial bristles, pigment on tibia and or femur, or complete lack of pigmentation or bristles). Within the S. ocreata clade, there is a correlation between overt visual signals and the presence of or- namentation. Several species within the clade have reduced or absent pigmentation {S. rov- neri and S, uetzi)', these also have reduced vi- sual signals. This study suggests that S. rov- neri, S. uetzi and S. stridulans evolved from ancestors that had overt visual signals and these species have subsequently lost visual signals in their courtship. However, as court- ship behavior is still unknown for several spe- cies with tibial bristles outside of the S. ocrea- ta clade (e.g. S. bilineata, S. salsa, S. segregata and S. aulonia) strong conclusions concerning the correlation of ornamentation and behaviors throughout the genus is not yet possible. Comparison of genital morphology. — Ex- amination of the pedipalps of these species in light of the hypothesized phylogeny sheds some light on potential homologies of geni- talic structures. The most distinctive feature in the male pedipalp of some members of Schi- zocosa is the finger like paleal process (Fig. 16), which is solely found in the S. ocreata group. However, the presence of a triangular structure in the same location in the species immediately basal to the S. ocreata clade, e.g., S. saltatrix (Fig. 13) and S. duplex (Fig. 14), suggests that both forms of the paleal process are homologous. DISCUSSION This phylogenetic study confirms that members of Lycosidae are conservative in morphology with large amounts of homoplasy in many characters and low bootstrap support for several branches. The preferred phylogeny presented here (Fig. 32) is a hypothesis of re- lationships within Schizocosa based on a weighting scheme that weights genitalia more heavily than somatic characters and excludes secondary sexual characters. The main clades that are proposed in this study are grouped either by geography (i.e. Clade A) or by a suite of morphological characters (i.e. S. ocreata clade, Clade B and Clade C) . This genus is of particular interest for evo- lutionary biologists because of the relatively large number of species with ornamentation (Figs. 33 & 34). Mapping the tibial bristles on STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 367 tibia I bristles unordered r~~l absent ■H present equivocal II S. crassipes FL a S. fioridana a S. uetzi II S. stridulans ^ S. rovneri II ocr OH X rov ^ S. ocreata OH S. crassipes MS „ S. ocreata FL S. ocreata MS S. duplex S. saltatrix S. sp. nr saltatrix Figure 35. — Cladogram of S. ocreata clade + basal species showing presence and extent of bristles on leg I. the preferred phytogeny shows the distribution (and variation) of the ornamentation. For the tibial bristles, there appears to be five or six independent gains of this character across all of the species in this study, with losses oc- curring in S. rovneri, S, uetzi and S. fioridana. In Clade A, the character possibly evolved two or three separate times. While it is challenging to trace the evolu- tion of specific courtship elements, there is clearly a grouping of taxa that show palpal drumming in courtship and a grouping of those that show stridulation in courtship. If the phytogeny presented in this study is correct, it appears that the species immediately basal to the S. ocreata clade (e.g., S. duplex and S. saltatrix) rely primarily on seismic signals in their courtship. The brush-legged taxa within the S. ocreata clade (e.g. populations of both S. ocreata and S. crassipes) have courtship that involves both seismic signals and visual signals. And, several species within the S. ocreata clade have apparently subsequently lost some of the more visual aspects of court- ship (e.g., S. uetzi, S. stridulans and S. rov- neri). 368 THE JOURNAL OF ARACHNOLOGY Table 5. — Comparison of elements of courtship behavior (both seismic and visual) in species of Schi- zocosa, Gladicosa and Allocosa. Species This study Previous studies Not known S. aulonia S. avida rapid papal drumming, Behavior not known leg 1 arch (Grey & Stratton 1998) S. bilineata Behavior not known S. cespitum Behavior not known S. chiricahua S. communis Palpal drumming (Don- Behavior not known dale & Redner 1978) S. crassipalpata S. crassipes (FL) Stridulation Stridulation, cheliceral strike, leg 1 extend and wave (Miller et al. 1998) S. crassipes (MS) Stridulation, cheliceral strike, leg 1 extend and wave (Miller et al. 1998) S. duplex Stridulation Stridulation (Hebets & Uetz 2000) S. floridana S. humilis Stridulation, leg 1 tap Behavior not known S. maxima S. mccooki Papal drumming (Strat- Behavior not known ton & Lowrie 1984) S. mimula Behavior not known S. minnesotensis S. ocreata (OH) “Jerky walk” “double Behavior not known tap” (Stratton & Uetz 1983, 1986) S. ocreata (FL) Cheliceral strike double (bilateral arch with legs 1) Stratton, Miller & Miller unpublished data) S. ocreata (MS) Cheliceral strike walk and tap legs 1 (Stratton, Miller & Miller unpub- lished data) S. retrorsa Palpal drumming, leg 1 wave (Hebets et al. 1996) S. rovneri Series of body bounces (Uetz & Denterlein 1979; Stratton & Uetz 1983, 1986) S. salsa S. saltatrix Stridulation, leg 1 arch Behavior not known S. sp. nr. saltatrix Stridulation or body vibra- tion S. segregata S. stridulans Stridulation, quick tap Behavior not known of leg 1 (Stratton 1991, 1997a 1997b) STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 369 Table 5. — Continued. Species This study Previous studies Not known S. uetzi S. ocr X rov hybrids A. georgicoia G. pulchra Stridulation, leg 1 extend and vibrate Stridulation Stridulation, leg 1 arch (Stratton 1997b; He- bets 2003) Stridulation, body bounce, double tap, jerky walk (Stratton & Uetz 1986) R. rabida Stridulation Papal scraping, stridula- tioe leg wave (Rov- ner 1968) Early work on S. ocreata from Ohio and S. rovneri suggested that these two species were sibling species (Uetz & Denterlein 1979; Stratton & Uetz 1983, 1986). The subsequent discovery of additional species in the clade (S. stridulans and S. uetzi) raised questions about the relationship between S. ocreata and S. rov- neri. However, this phylogenetic analysis con- firms that these two taxa are each other’s clos- est relatives. Some currently recognized species within this genus may not reflect phylogenetic line- ages as several of the taxa used in this study are paraphyletic (e.g., S. ocreata, S. crassipes and S. saltatrix). The separation of S. ocreata (OH) from S. ocreata (FL and MS) suggests that these populations are diverging and could be considered separate taxa if a phylogenetic species concept is applied. Miller et al. (1998) showed reduced breeding between crosses composed of S. ocreata (OH) and S. ocreata (MS, population from Washington Count, Ler- oy Percy State Park, indicated LP in that study). Likewise, the clear separation of S. crassipes (MS) from S. crassipes (FL) also suggests divergence. An unpublished study showed reduced interbreeding but a similarity of courtship behavior between both popula- tions of S. crassipes (Germano et al. unpub-' lished data). It is intriguing that the S. ocreata from Mississippi is basal in the S. ocreata clade. The Mississippi River valley has been suggested to be a refuge of deciduous woods during the last ice ages (Delcourt et al. 1980; Delcourt & Delcourt 1987); it is tempting to speculate that perhaps the S. ocreata clade di- verged from populations along the Mississippi River valley in the time since the last glaciers. By this scenario, as glaciers retreated in the north, and deciduous woods expanded in the southeast, S. ocreata spread first across the southeast and, as populations became isolated they speciated to S. ocreata (FL), then S. cras- sipes and the other species within the clade eventually spreading north. The 5. ocreata clade is now found throughout the eastern U.S.A. with most species found in the south- east. Molecular data could provide an inde- pendent test of this hypothesis. McClintock & Uetz (1996) showed that fe- males of S. rovneri preferred S. rovneri males that were artificially given tibial bristles. Their preliminary phylogenetic study suggested that S. rovneri was basal to S. ocreata, thus poten- tially providing an example of the sensory bias hypothesis or an example where the fe- male preference for a trait (in this case, or- namentation in the form of tibial bristles) pre- ceded the evolution of the trait and provided selection for the trait in subsequent species (in this case S. ocreata). However, based on the analyses presented here, S. rovneri is derived relative to S. ocreata and thus the S. rovneri female preference for males with bristles (re- ported by McClintock & Uetz 1996) is a re- tained characteristic. The fieger-like process on the palea of the pedipalp seen in the S. ocreata clade is unique in wolf spiders. The somewhat similar process seen in Sosippus does not appear to be ho- mologous to this structure (Sierwald 2000, fig. 7). High magnification video from the ventral view during copulation confirms that the me- dian apophysis of the male engages the epi- 370 THE JOURNAL OF ARACHNOLOGY Courtship behavior unordered □□ stridulation ■ palpal drumming ^ polymorphic ron equivocal rrn TH m m mr ^s: =30 S + Lc IDS + □ s ++ ns 3D S ++ 3DS ++ 3DS ++ ]D S ++ 3D S 3D S + 3DS + 6 IDS □ D D mn iri pm L IDS ID S++ S. crassipes FL S. floridana S. uetzi S. stridulans ocr OH X rov hybrid S. rovneri S. ocreata OH S. crassipes MS S. ocreata FL S. ocreata MS S. duplex S. saltatrix S. nr. saltatrix S. segregate S. bilineata S. crassipalpata S. humilis S. avida S. mccooki S. aulonia S. communis S. maxima S. cespitum S. retrorsa S. chiricahua S. salsa S. mimula S. minnesotensis G. pulchra A. georgicola D ++ R. rabida STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 371 gynal hood of the female, and the paleal pro= cess of the male pinches against the side of the epigynum of the female (Stratton, Miller & Miller, unpub. data), Eberhard (1994) sug- gested that any time there is physical contact between structures during copulation, the po- tential exists for that trait to be influenced through female choice during copulation. Thus, this character may give females an ad- ditional means to evaluate mates and poten- tially exercise female choice of gametes. After the examination of thousands of male speci- mens I have never seen the paleal process bro- ken, It is curious that the clade with a con- centration of species with conspicuous secondary sexual characters also has the unique trait of the paleal process that could also be shaped by female choice. It is sug- gested here that the morphology of primary and secondary genitalic characters in males in this clade may be largely shaped by sexual selection by female choice. As may be expected due to a lack of infor- mative characters in Lycosinae, the preferred phylogeny did not have strong bootstrap sup- port and a combined morphological and mo- lecular analysis may show better supported re- sults in regard to the phylogenetic relationships of Schizocosa. Indeed, further clarification of this genus and its relatives will provide the phylogenetic context to best interpret behavioral, ecological and evolutionary questions. ACKNOWLEDGMENTS Thanks are extended to many individuals for their assistance in this study. For the re- vision that provided the groundwork for much of this study and for many good conversa- tions, I thank C. Doedale and J. Redner. N. Platnick (AMNH), J. Coddington and S. Larcher (NMNH), L. Leibeesberger (MCZ), G. B. Edwards (ESC A), R. Brown (MEM) and H. Guarisco were all generous when loaning specimens. R Miller helped extensively with field collections and identifications. H. Guar- isco contributed live specimens for behavioral studies. The SEMs were done by G. Baker. A. Douglas presented an excellent class in phy- logenetics and provided feedback at numerous stages of this study. This study was informed by many conversations over the years. I ex- tend my thanks to R Miller, G, Uetz, E, He- bets, C. Dondale, A. Brady and W. Maddison. Several individuals read earlier drafts of this paper and their suggestions and corrections were much appreciated. For this I extend my thanks to V. Framenau, C. Vink, E. Hebets, A. Douglas and W. Maddison. Finally, thanks are extended to C. Dondale, J. Redner & A. Brady for their many revisions of North American lycosids. LITERATURE CITED Barnes, R.D. 1953. Report on a collection of spi- ders from the coast of North Carolina. American Musem Novitates 1632:1-21. Barthel, J. & O. von Helversen, 1990. Pardosa wagleri (Hahn 1822) and Pardosa saturatior Si- mon 1937, a pair of sibling species (Araneae, Lycosidae). Bulletin for the Society of European Arachnology 1:17-23. Brady, A.R. 1962. The spider genus Sosippm in North America, Mexico, and Central America (Araneae, Lycosidae). Psyche 69:129-164. Brady, A.R, 1979. Nearctic species of the wolf spi- der genus Trochosa (Araneae: Lycosidae). Psy- che 86:167-212. Brady, A.R. 1986. Nearctic species of the wolf spi- der genus Gladicosa (Araneae: Lycosidae). Psy- che 93:285-319. Brady, A.R. & K.S. McKinley. 1994. Nearctic spe- cies of the wolf spider genus Rabidosa (Araneae: Lycosidae). Journal of Arachnology 22:138-160. Chamberlin, R.V. & W. Ivie. 1944. Spiders of the Georgia region of North America. Bulletin of the University of Utah 35:143-231. Delcourt, P.A & H.R. Delcourt. 1987. Long-term Forest Dynamics of the Temperate Zone. A case study of late-Quatereary forests in Eastern North America. Springer- Verlag, New York. Delcourt, P.A., H.R. Delcourt, R.C. Brister & L.E. Lackey. 1980. Quaternary vegetation history of the Mississippi embayment. Quaternary Re- search 13:111-132. Dondale, C.D. 1986. The subfamilies of wolf spi- ders (Araneae: Lycosidae). Actas X Coegreso In- <- Figure 36. — Distribution of major courtship elements on the preferred phylogeny. The symbol “D” refers to species in which the males are known to show palpal drumming during courtship, “S” refers to species in which males are known to show palpal stridulation. The “-h” symbol shows there is arching of leg 1, + ” shows species which have extensive visual signals. 372 THE JOURNAL OF ARACHNOLOGY ternacional de Aracnologia Jaca/Espana, 1986. I: 327-332. Dondale, C.D. & J.H. Redner, 1978. Revision of the Nearctic wolf spider genus Schizocosa (Aranei- da: Lycosidae). Canadian Entomologist 1 10:143- 181. Dondale, C.D. & J.H. Redner. 1984. Revision of the milvina group of the wolf spider genus Pardosa (Araneae: Lycosidae) Psyche 91:67-117. Dondale, C.D. & J.H. Redner. 1990. The Insects and Arachnids of Canada, Part 17. The wolf spi- ders, nurseryweb spiders, and lynx spiders of Canada and Alaska. Biosystematics Research Centre, Ottawa. Eberhard, W.G. 1994. Evidence for widespread courtship during copulation in 131 species of in- sects and spider, and implication for cryptic fe- male choice. Evolution 48:711-733. Grey, E. & G.E. Stratton. 1998. The courtship be- havior of Schizocosa avida (Walckenaer) (Ara- neae; Lycosidae). American Arachnology 56:7. Griswold, C.E. 1987. A revision of the jumping spi- der genus Habronattus F.O.P.-Cambridge (Ara- neae; Salticidae) with phenetic and cladistic anal- yses. University of California Publications in Entomology 1 07 : 1 -345 . Hebets, E.A. 2003. Subadult experience influences adult mate choice in an arthropod: exposed fe- male wolf spider prefer males of a familiar phe- notype. Proceedings of the National Academy of Sciences 100:13390-13395. Hebets, E.A. 2005. Attention-altering signal inter- actions among signals in the multimodal court- ship display of the wolf spider Schizocosa uetzi. Behavioral Ecology. 16:75-82. Hebets, E.A. & G.W. Uetz. 1999. Female responses to isolated signals from multimodal male court- ship displays in the wolf spider genus Schizocosa (Araneae: Lycosidae). Animal Behaviour 57; 865-872. Hebets, E.A. & G.W. Uetz. 2000. Leg ornamenta- tion and the efficacy of courtship display in four species of wolf spider (Araneae: Lycosidae). Be- havior Ecology and Sociobiology 47:280-286. Hebets, E.A., G.E. Stratton & G.L. Miller. 1996. Habitat and courtship behavior of the wolf spider Schizocosa retrorsa (Banks) (Araneae, Lycosi- dae). Journal of Arachnology 24:141-147. Homann, H. 1971. Die Augen der Araneae. Ana- tomic, Ontogenie und Bedeutung fiir die Syste- matik (Chelicerata, Arachnida). Zietschrift fiir Morphologic der Tiere 69:201-272. Kaston B.J. 1936. The senses involved in the court- ship of some vagabond spiders. Entomologica Americana 16:97-167. Kronestedt, T. 1990. Separation of two species stand- ing as Alopecosa aculeata (Clerck) by morpho- logical, behavioural and ecological characters, with remarks on related species in the pulverulen- ta group (Araneae, Lycodidae). Zoologica Scripta 19:203-225. Maddison, W.R 1993. Missing data versus missing characters in phylogenetic analysis. Systematic ; Biology 42:576-581. Maddison, W.P. & M. Hedin. 2003. Phylogeny of Habronattus jumping spider (Araneae: Saltici- ^ dae), with consideration of genital and courtship evolution. Systematic Entomology 28:1-21. Maddison, W.P. & D.R. Maddison. 1992. Mac- Clade: analysis of phylogeny and character evo- lution. Version 3. 06. Sinauer, Sunderland, MA. McClintock, W.J. & G.W. Uetz. 1996. Female choice and preexisting bias: visual cues during courtship in two Schizocosa wolf spiders (Ara- neae: Lycosidae). Animal Behaviour 52:167- 181. Miller, G.L., G.E. Stratton, P.R. Miller, & E.A. He- bets. 1998. Geographical variation in male court- ship behaviour and sexual isolation in wolf spi- ders of the genus Schizocosa (Araneae: Lycosidae). Animal Behaviour 56:937-951. Peckham, G.W. & E.G. Peckham. 1889. Observa- tions on sexual selection in spiders of the family Attidae. Occasional Papers of the Wisconsin Nat- ural History Society 1:3-60. Peckham, G.W. & E.G. Peckham. 1890. Additional observations on sexual selection in spiders of the family Attidae, with some remarks on Mr. Wal- lace’s theory of sexual ornamentation. Occasion- al Papers of the Wisconsin Natural History So- ciety 1:117-151. Platnick, N.I. 2005. The World Spider Catalog ver- sion 5.5. http://research.amnh.org/entomology/ spiders/catalogS 1 -87/index. html Platnick, N.L, Griswod, C.E. & J. A. Coddington. 1991. On missing entries in cladistic analysis, i Cladistics 7:337-343. | Rovner, J.S. 1975. Sound production by nearctic j wolf spiders: A substratum-coupled stridulatory i mechanism. Science 190:1309-1310. Rovner, J.S. 1968. An analysis of display in the ' lycosid spider Lycosa rabida Walckenaer. Ani- i mal Behaviour 16:358-368. Ryan, M.J. & A.S. Rand. 1993. Sexual selection and signal evolution: the ghost of biases past. Philosophical Transactions Royal Society of j London Series B 340:187-196. Scheffer, S.J., G.W. Uetz & G.E. Stratton. 1996. ’ Sexual selection, male morphology, and the ef- ficacy of courtship signaling in two wolf spiders : (Araneae: Lycosidae). Behavior Ecology and So- ciobiology 38:17-23. Sierwald, P. 2000. Description of the male Sosippus placidus with notes on the subfamily Sosippinae. Journal of Arachnology 28:133-140. Stratton, G.E. 1991. A new species of wolf spider, Schizocosa stridulans (Araneae, Lycosidae). Journal of Arachnology 19:29-39. ' STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 373 Stratton, G.E, 1997a. A new species of Schizocosa from the southeastern USA (Araneae, Lycosi- dae). Journal of Arachnology 25:84-92. Stratton, G.E. 1997b. Investigation of species di- vergence and reproductive isolation of Schizo- cosa stridulans (Araneae: Lycosidae) from Illi- nois. Bulletin of the British Arachnological Society 10:313-321. Stratton, G.E. & D.C. Lowrie. 1984. Courtship be- havior and life cycle of the wolf spider Schizo- cosa mccooki (Araneae, Lycosidae). Journal of Arachnology 12:223-228. Stratton, G.E. & G.W. Uetz. 1983. Communication via substratum-coupled stridulation and repro- ductive isolation in wolf spiders (Araneae: Ly- cosidae). Animal Behaviour 31:164-172. Stratton, G.E. & G.W. Uetz. 1986. The inheritance of courtship behavior and its role as a reproduc- tive isolation mechanism in two species of Schi- zocosa wolf spiders (Araneae; Lycosidae). Evo- lution 40:129-141. Stratton, G.E., RR. Miller, G.L. Miller and E.A. He- bets. 1996. Copulatory pattern and duration in wolf spiders (Araneae; Lycosidae). Journal of Arachnology. 24: 1 86-20. Swofford, D.L. 2002. PAUP*. Phylogenetic Anal- ysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Mas- sachusetts. Uetz, G.W, & G, Denterlein. 1979. Courtship be- havior, habitat and reproductive isolation in Schi- zocosa rovneri Uetz and Dondale (Araneae, Ly- cosidae). Journal of Arachnology 7:121-128. Uetz, G.W, & C.D, Dondale, 1979. A new wolf spider in the genus Schizocosa (Araneae, Lycos- idae) from Illinois. Journal of Arachnology 7:86- 87. Uetz, G.W. & J. A. Roberts. 2002. Multisensory cues and multimodal communication in spiders: Insights from video/audio playback studies. Brain Behavior and Evolution 59:222-230. Uetz, G.W. & E. 1. Smith. 1999. Asymmetry in a visual signaling character and sexual selection in a wolf spider. Behavior Ecology and Sociobiol- ogy 45:87-93. Uetz, G.W, W. J. McClintock, D. Miller, E.I. Smith & K.K. Cook. 1996. Limb regeneration and sub- sequent asymmetry in a male secondary sexual character influences sexual selection in wolf spi- ders. Behavior Ecology and Sociobiology. 38: 253-257. Vink, C.J. 2002. Lycosidae (Arachnida: Araneae). Manaaki Whenua Press, Lincoln. Wallace, H.K. 1942a. A revision of the burrowing spiders of the genus Geolycosa (Araneae, Lycos- idae). American Midland Naturalist 27:1-62. Wallace, H.K. 1942b. A study of the lenta group of the genus Lycosa with descriptions of new spe- cies (Araneae, Lycosidae). American Museum Novitates 1185:1-21. Manuscript received 22 September 2004, revised 1 April 2005. 374 THE JOURNAL OF ARACHNOLOGY Appendix 1 . — Data matrix of character states for the phylogenetic analysis of Schizocosa wolf spiders and outgroups. 1111111111222222222233333333334444444444 Taxon/Node 1234567890123456789012345678901234567890123456789 S. aulonia S. avidci S. bilineata S. cespitum S. chiricahiici S. communis S. crassipcilpata S. crassipes (FL) S. crassipes (MS) S. duplex S. floridana S. hiimilis S. maxima S. mccooki S. mimida S. minnesotensis S. ocreata (OH) S. ocreata (FL) S. ocreata (MS) ocr (OH) X rov hybrid S. retrorsa S. rovneri S. salsa S. saltatrix S. sp. nr saltatrix S. segregata S. stridulaus S. uetzi G. pulchra R. rabida A. georgicola 121211101000000100000100000001111000301000?101110 0202111101200101111001000000011010002010000010000 2201111010011001000001001000011101110100100000211 1212111100100101000011000010011111003021000111000 1202111010000000000001001010001011001030000000000 1210010100100101001001011000001110003020000010000 2201110110011001000001011000010101101110100000000 1202100000001011011101100000111001110031001010210 2200000010001001100001100000111002113010000010210 2200010100001011001101100001001101112030000010000 2212200000001001011001100000111001112021001000010 2200000100000100000001000000010101001020000010010 0211111101100100000001011000010100003020000000010 0201111101100001011001010000010110002010000000000 1201111000000000000011010110000110002020000000010 0201100100100100000011000000000111001010000010010 1200000100000101101101100000100102121021000010210 1200000000001011001101100000101001123010000010212 1201100100000011001101100000101002123020001010210 1200000100000000001101101000101101123011001110110 1212100111100000001011010000010110001020001010000 1201100100000010101101100000100102124011001000000 1101211000000001100011001010001011001031001000111 1200000100001001011001100001000001113021000000010 12 0120000000100101100110000101000711102 0000000010 2201100110001011011001100001000100001020007000111 2202100000001001101101100000100002121031000010110 1202100000001001111001100000100002120031001000010 0212100100100000111000000110001101002011010000010 01012 110000 0000 710000 0000000 0111110010 71010111010 0000000100001110000000000100001110001002010000000 Appendix 2 Specimens examined for this study and literature references for figures. Deposition of specimens are as indicated; GES (collection of G. Stratton), PRM (collection of R Miller), AMNH (American Muse- um of Natural History), FSCA (Florida State Col- lection of Arthropods), and MCZ (Museum of Comparative Zoology). Coordinates are provided for as many collecting localities as possible, either by transcribing data included on collecting labels, or searching for localities on the USGS National Mapping Information website (http://geonames. usgs.gov/) search under “U.S. and territories que- ry.” Coordinates not included on original labels are in brackets. Schizocosa aulonia Dondale 1969. — U.S. A.: Il- linois: Madison County, Poag, SE 1/4 NW 1/4SE SE 1/4 Sec. 13 T4N, R9W, 138°48'N, 9()°()2'W], pitfall, 12-19 July 1993, David Landes, 2 6 (GES). Schizocosa avida (Walckenaer 1837). — U.S. A.: Mississippi: Pontotoc County, 3 miles NE of Pon- totoc off Highway 9, field near Mubby Creek, R3E T95 Sect 15 [34°18'N, 88°56'W], 13 June 1992, G. Stratton, P. Miller, W. Miller, 1 (3(GES); Missisippi: Marshall County, 3 miles S. of Waterford [34°36'N, 89°29'W], 9-12 June 1994, G. Stratton, 1 ? (GES). Published figures: Dondale & Redner (1978, figs. 10-12, 51-54). Schizocosa bilineata (Emerton 1885). — U.S. A.: Michigan: Eaton County, T2N R6W Sect 6 [42°36'N, 85°02'W], on porch of farm, 29 June 1989, G. Stratton, 1 6 (GES); Mississippi, Panola County, 3 miles NE. of Como, west side of Nelson Creek, on Compress Road, T6S R7W Sec 23 & 24 [34°32'N, 89°54'W], 15 May 1998, G. Stratton, P Miller, 1 9 (GES). Published figures: Dondale & Redner (1978, figs. 8, 47, 48). Schizocosa cespitum Dondale & Redner 1978. — CANADA: Saskatchewan: Matador, 40 miles N. of Swift Current, short grass prairie, P.W. STRATTON— EVOLUTION IN SCHIZOCOSA WOLF SPIDERS 375 Riegert, 1 cJ, 3 ? paratypes (AMNH). Published figures: Dondale & Redner (1978, figs. 20, 69, 70). Schizocosa chiricahua Dondale & Redner 1978. — U.S.A.: Arizona: Chochise County, Chiri- cahua Mts, Southwest Research Station [31°53'N, 109°12'W], 5400', in swimming pool, 15 July 1967 & June 7 1968, V. Roth, 2 6 paratypes (AMNH); Arizona: Graham County, near Safford [32°50'N, 109°42'W], 14 July 1956, W. Gertsch & V. Roth, 2 d paratypes, 1 9 paratype (AMNH). Published fig- ures: Dondale & Redner (1978, figs. 18, 19, 66- 68). Schizocosa communis (Emerton 1885). — CAN- ADA: Ontario: 9 miles NW. of Bloomfield, 44°04'N, 77°18'W, 17 July 1965, Jean & Wilton Me, 1 d (AMNH); U.S.A.: Pennsylvania: Berks County, Lenhartsville [40°34'N, 75°53'W], July 1964, Vaurie, 1 9 (AMNH). Published figures: Dondale & Redner (1978, figs. 14, 55-58). Schizocosa crassipalpata Roewer 1951. — CAN- ADA: Ontario: Rednersville [44°07'N, 77°27'W], field pitfall, 3-19 June 1964, C.D. Dondale, 2 d, 2 9 (ESC A). Published figures: Dondale & Redner (1978, figs. 9, 49, 50). Schizocosa crassipes (Walckenaer 1837). — U.S.A.: Florida: Alachua County, 0.25 miles E. of River Styx on Highv/ay 346 [29°3ri7"N, 82H5'47"W], 10 May 1994, G.B. Edwards & R Cushing, 3 d, 4 9 (GES). Published figures: Don- dale & Redner (1978, figs. 2, 27-30); Stratton (1991, fig. 4). Schizocosa crassipes (Walckenaer 1837). — U.S.A. : Mississippi: Grenada County, T21N R2E sect. 12, 13N &R3E Sec 7S, 18N [33°44'N, 90°00'W], deciduous woods on hillside, 16 May 1993, G. Stratton, P. Miller, W. Miller, B. Grantham, 1 d, 1 9 (GES). Schizocosa duplex Chamberlin 1925. — U.S.A.: Mississippi: Lafayette County, 8 miles SE. of Ox- ford, TIOS R3W Sec 35, 34°36'N, 89°29'W, in pine litter, 18 May 1993, G. Stratton, E. Hebets, 1 d, 2 9 (GES). Published figures: Dondale & Redner (1978, figs. 6, 42-44). Schizocosa floridana Bryant 1934. — U.S.A.: Florida: Marion County, Hopkins Prairie, Ocala National Forest [29°10'N, 8r47'W], 25 March 1983. G. Stratton, litter, 4 d, 2 9 (GES). Published figures: Dondale & Redner (1978, figs. 3, 31-34, 35). Schizocosa humilis (Banks 1892). — U.S.A.: Pennsylvania: Bucks County, east of Jamieson, Horseshoe Bend, Neshaminy Creek, May 1954, W. Ivie (GES). Published figures: Dondale & Redner (1978, figs. 7, 45, 46). Schizocosa maxima Dondale & Redner 1978. — U.S.A.: California: Solano County, Fair- field [38°15'N, 122°02'W], April-August 1955, K.W. Haller, 1 d paratype (AMNH); California: Tu- olumne County, S. of Highway 108, 5 miles E of Sonora, eiev. 2000', drowned in swimming pool, 14 October 1973, W. Icenogle, 1 9 paratype (AMNH). Published figures: Dondale & Redner (1978, figs. 16, 17, 63-65). Schizocosa mccooki (Montgomery 1904), — U.S.A.: New Mexico: Santa Fe County, Santa Fe [35°3rN, 105°56'W], in pinyon scrub, 17 June 1979, D. Lowrie, 3 d, 1 9 (GES). Published fig- ures: Dondale & Redner (1978, figs. 13, 15, 59- 62). Schizocosa mimula (Gertsch 1934). — U.S.A.: Colorado: Otero County, Highway 109, 5 June 1967, 1 d (AMNH). Schizocosa minnesotensis (Gertsch 1934), — U.S.A.: Wyoming: Lincoln County, Kemmerer [4r48'N, 110°32'W], 24 August 1983, R. Parmen- ter, 2 d, 2 9 (AMNH). Schizocosa ocreata (Hentz 1844). — U.S.A,: Ohio: Clermont County, Cincinnati Nature Center [39°07'N, 84°15'W], 6 May 1980, G. Stratton, 1 d, 1 9 (GES). Published figures: Stratton (1991, figs. 3, 9). Schizocosa ocreata (Hentz 1844). — U.S.A.: Mississippi: Washington County, Leroy Percy State Park, N. of Highway 12 near entrance of park, T15N R7W, 90°50'W, 33°10'N, 9 April 1993, on knoll by flooded bottomlands, G. Stratton, P. Miller, 1 d, 1 9 (GES). Schizocosa ocreata (Hentz 1844). — U.S.A.: Florida: Alachua County, 0.25 miles E. of River Styx on Highway 346 [29°3r 17"N, 82°15'47"W], 1 March 1993, G.B. Edwards, P. Cushing, 9 d, 3 9 (GES). Schizocosa retrorsa (Banks 1911). — U.S.A.: Mississippi: Lafayette County, 8 miles SE. of Ox- ford, TIOS R3W Sec 35, 34°36'N, 89°29'W, pitfall in pine woods, 28 June-5 July 1993, 1 d,l 9, G. Stratton (GES). Published figures: Dondale & Red- ner (1978, figs. 21, 75-78). Schizocosa rovneri Uetz & Dondale 1979. — U.S.A.: Kentucky: Boone County, 5 miles W. of Taylorsport, floodplain of Ohio River, “Sandy Run” [39°05'N, 84°41'W], 3 May 1996, K. Dela- ney, 1 d, 1 9 (GES). Published figures: Stratton (1991, figs. 2, 7). Schizocosa salsa Barnes 1953. — U.S.A.: Missis- sippi: Hancock County, mouth of Jordan River [30°16'N, 89°37'W], on Juncus marsh island, 25 June 1993, 1 d (PRM). Published figures: Dondale & Redner (1978, figs. 24, 79, 80). Schizocosa saltatrix (Hentz 1844). — U.S.A.: Mississippi: Lafayette County, 8 miles SE. of Ox- ford, TIOS R3W Sec 35, 34°36'N, 89°29'W, “Lone- some 80,” pine deciduous woods, 16 March 1996, G. Stratton & P. Miller, 1 d, 1 9 (GES). Published figures: Dondale & Redner (1978, figs. 4, 39-41). Schizocosa sp. nr saltatrix. — U.S.A.: Mississip- pi: Wilkinson County, 5 miles E of Doloroso on Smith Rd, S. of Homochito River, 3r20'N, 376 THE JOURNAL OF ARACHNOLOGY 92°45'W, uplands deciduous forest, 10 April 1993, G. E. Stratton, RR. Miller, 1 d, 1 9 (GES). Schizocosa segregata Gertsch & Wallace 1937. — U.S.A.: Florida: Levy County, 28 April 1934, # 298, H.K. Wallace, 1 d, 1 ?, paratypes (poor condition); Texas: Edinburg [26°18'N, 98°10'W], 1934, S. Mulaik, 1 d (poor condition) (AMNH). Published figures: Dondale & Redner (1978, figs. 23, 81, 82). Schizocosa stridulans Stratton 1984. — U.S.A.: Illinois: Mason County, Sand Ridge State Forest [40°24'N, 89°52'W], 7 June 1985, G. Stratton, L. Hartz, 1 (3, 1 ? (GES). Published figures: Stratton (1991, figs. 1, 5, 6, 13). Schizocosa uetzi Stratton 1997. — U.S.A.: Mis- sissippi: Lafayette County, 8 miles SE of Oxford, TIOS R3W Sec 35, 34°36'N, 89°29'W, “Lonesome 80,” mixed pine and hardwoods, 4 July 1992, G. Stratton, 13,19 (GES). Published figures: Strat- ton (1997, figs. 1-5). Schizocosa ocreata x Schizocosa rovneri hy- brids.— Cross between S. ocreata 9 from Ohio and S. rovneri 3 from Kentucky, 1980-1982, 1 3 (GES). Cross between S. rovneri 3 from Kentucky and S. ocreata 9 from Ohio, 1980-1982, 1 9 (GES). Allocosa georgicola (Walckenaer 1837). — U.S.A.: Mississippi: Lafayette County, 8 miles SE. of Oxford, TIOS R3W Sec 35, 34°36'N, 89°29'W, “Lonesome 80,” pitfall in mixed pine and hardwoods, 16-24 September 1992, G. Stratton, P. Miller, 1 3,1 9 (GES). Published figures: Cham- berlin & Ivie (1944, fig. 57). Pedipalps and epigyna of the closely related H. helluo (Walckenaer 1837) are figured in Kaston (1948, figs. 1065, 1066, 1090, 1 105) and Dondale & Redner (1990, figs. 43-47). Gladicosa pulchra (Keyserling 1877). — U.S.A.: Tennessee: Hardeman County, Chickasaw State Park [35°22'N, 88°50'W], pine deciduous woods, 1 November 1992, G. Stratton, 13,19 (GES). Pub- lished figures: Brady (1986, figs. 3, 10-14, 39-42). Rabidosa rabida (Walckenaer 1837). — U.S.A.: Mississippi: Lafayette County, 8 miles SE. of Ox- ford, TIOS R3W Sec 35, 34°36'N, 89°29'W, pine deciduous mixed, 5 August 1991, G. Stratton, R Miller, G. Miller, 1 3, 1 9 (GES). Published fig- ures: Brady & McKinley (1994, figs. 1, 6, 11-14) and Kaston (1948, figs. 1077, 1079, 2006). 2005. The Journal of Arachnology 33:377-383 FACTORS AFFECTING CANNIBALISM AMONG NEWLY HATCHED WOLF SPIDERS (LYCOSIDAE, PARDOSA AMENTATA) Aino Hvam', David Mayntz^ 2,3 Rikke Kruse Nielsen*: “Department of Ecology and Genetics, University of Aarhus, Bldg. 540, DK-8000 Arhus C, Denmark; ^Department of Zoology, University of Oxford, South Parks Road, Oxford 0X1 3PS, UK ABSTRACT. Cannibalism is a common phenomenon among young wolf spiders (Lycosidae). The pur- pose of this study was to investigate how various factors influence cannibalistic tendencies in hatchlings of Pardosa amentata (Clerk 1757). The basic experimental approach was to place pairs of unfed hatchlings of similar body mass in small containers without prey and to measure if and when cannibalism happened. From the data, we identified three different cannibalistic strategies. One large group of hatchlings never cannibalized and thus died from starvation. Another group cannibalized shortly before the time at which they were predicted to die from starvation. In these spiders, there was a strong positive relationship between average body mass of the contestants and their latency to cannibalize. A third group cannibalized quickly and the latency to cannibalize in these spiders was independent of body mass. We also tested if cannibalistic tendencies were higher among unrelated pairs than among pairs of siblings, but we did not find any support for this hypothesis. In another experiment we tested if maternal effects influenced can- nibalism, i.e. if siblings from certain mothers were more cannibalistic than siblings from others. We did not find any evidence that maternal effects influenced whether or not cannibalism occurred. However, when cannibalism did occur, the latency to cannibalize varied significantly among siblings from different mothers beyond what would have been predicted solely from hatchling body mass. Keywords: Lycosidae, intraspecific predation, spiderlings, kin recognition, maternal effects Cannibalism among wolf spiders is often observed in the field. One example is Pardosa lugubris (Walckenaer 1802) which seems to be the most important predator of its own spe- cies (Edgar 1969), and it was estimated that juveniles of this species included conspecifics as 29% of their total diet (Hallander 1970). Other examples are Schizocosa ocreata (Hentz 1844) and Pardosa milvina (Hentz 1844) in which cannibalism is assumed to be an important regulating factor on population density (Wagner & Wise 1996; Buddie et al. 2003). A variety of factors has been suggested as potential selective forces, promoting or inhib- iting cannibalistic behavior. The most obvious advantage connected with cannibalism is that the cannibal gains a meal in addition to the normal diet and cannibals often show higher growth and survival rates than their non-can- nibalistic conspecifics (Polis 1981). As can- ^ Corresponding author. nibals are facing prey with similar predatory abilities, an obvious cost of cannibalism is the risk of retaliation. Another intriguing cost of cannibalism is the potential loss of inclusive fitness when a cannibal kills a genetically re- lated individual (Elgar & Crespi 1992; Pfen- nig & Sherman 1995). If this cost is large, we would predict spiders to be able to distinguish between kin and non-kin and to treat kin and non-kin differently (Pfennig & Sherman 1995). Kin recognition has been shown to oc- cur in many cannibalistic animals (see refer- ences in Pfennig 1997) and also some spiders seem to be able to identify and subsequently avoid eating a close relative (Evans 1999; Bil- de & Lubin 2001; Anthony 2003; Roberts et al. 2003). Wolf spider females carry their young on the abdomen for about a week. Thus, hatchlings have a good opportunity to learn chemical or visual cues, which could lat- er be used to recognize siblings from non-sib- lings. Adult wolf spiders can survive starvation 377 378 THE JOURNAL OF ARACHNOLOGY for several months (Anderson 1974). Newly hatched spiderlings on the other hand can only survive a few days or weeks before their nu- trient reserves are depleted (Wagner & Wise 1996; Toft & Wise 1999). This means that the first meal is of utmost importance for spider- lings and cannibalism can therefore be impor- tant for juvenile survival. In the present experiments we investigated cannibalistic tendencies among equally sized pairs of unfed hatchlings and provided them no choice other than to cannibalize or to die from starvation. Using this approach we eval- uated different hypotheses about what influ- ences cannibalistic tendencies in the hatch- lings. In the first experiment, we paired sibling and non-sibling hatchlings in order to test if cannibalism was dependent on kinship. From these results we also describe three apparently different strategies among hatchlings. Poten- tially, a mother of a brood can affect the con- dition of her spiderlings and thus also their cannibalistic propensities, for example through her nutritional status before reproduc- tion. The rates of cannibalism may also vary between closely related species or among and even within populations, due to genetic dif- ferences (Thibault 1974; Stevens & Mertz 1985; Tarpley et al. 1993). In a second exper- iment we therefore tested if there was varia- tion in cannibalistic tendencies among hatch- lings descending from different mothers. Average body mass (mg) 2 6.5 ■O o > 6.0 U) (/) Q 5.5 Starvation Cannibalism METHODS Cause of death The wolf spider. — The wolf spider Par- dosa amentata is abundant in Europe in many open, humid habitats, especially grasslands and agricultural fields with a well-developed litter layer (Alderweireldt & Maelfait 1988). Reproduction takes place in May-July. Fe- males carry the eggsac for 2-3 weeks and hatchlings spend about one week on their mother’s abdomen before they disperse (Rob- erts 1995). Experiment 1. — The purpose of this ex- periment was to test if kinship affected can- nibalism and to describe the cannibalistic ten- dency of hatchlings in general. Subadult male and female P. amentata spiders were collected in spring in a meadow at Stjaer, Denmark (56°07'N, 9°9UE), and brought to the labo- ratory. They were housed individually in plas- tic containers (diameter 35 mm, height 80 mm) with a plaster bottom, which was wetted Figures 1-2. — 1. Effects of body mass and cause of death on the survival time of wolf spider hatch- ' lings (Experiment 1). Survival time was measured from the time spiderlings were paired. Each point represents the time passed until one of the two spi- ders in a pair died from starvation (black triangles) or from cannibalism (open circles). Body mass is the average mass of the two hatchlings in a pair. Regression lines are based on spiders that died after day 3, i.e > 7days old (above horizontal dotted line); death from starvation = broken regression line, cannibalism = solid regression line. 2. Number ^ of days survived adjusted for average body mass f (least squares means, calculated on spiders dying after day 3 in the experiment, i.e. > 7 days old). frequently to maintain a permanent high hu- “ midity in the container. The spiders were fed wild type Drosophila melanogaster (Meigen) in excess until maturity. Fruit flies were raised HVAM ET AL.— CANNIBALISM AMONG WOLF SPIDER HATCHLINGS 379 Average body mass (mg) Figure 3. — Relationship between average body mass of hatchling pairs and their latency to canni- balize (Experiment 2). The two spiderlings always originated from the same eggsac. Each point rep- resents the time passed until cannibalism occurred. The regression line is based on spiders cannibaliz- ing after day 1 (above horizontal dotted line), i.e. > 7 days old. About 64% of the spiders did not cannibalize at all and are not shown in the figure. on instant Drosophila medium (formula 4-24 Plain, Carolina Biological Supply, Burlington NC), mixed with crushed dogfood (Techni-Cal ADULT®, Martin Pet Foods, Canada). Fully matured females were mated with a single male and different males were used for each female in order to avoid offspring from dif- ferent females being half-siblings. The young descending from eight eggsacs were chosen for the experiment. These eggsacs hatched within a period of four days and when hatch- lings were 4 ± 1 days old, they were weighed to the nearest pug. Pairs of hatchlings descend- ing from the same eggsac {n = 71) or from different eggsacs {n = 71) were then placed in the same plastic tube (diameter 20 mm, height 60 mm). The tubes contained a plaster bottom, which was wetted frequently to main- tain a permanent high humidity in the con- tainer. Body mass asymmetry was avoided by pairing spiders of almost equal body mass (mean body mass ± SE = 421 ±4 pg; mean weight difference ± SE = 2,6 ± 0,2pg; max. weight difference = 14pg). Spideiiing age (days since hatching) at the start of the ex- periment varied up to three days within a pair. Experimental conditions were set at 25 ±0.1 °C; 16L:8D. The spiderlings never received any food but had constant access to water from the plaster. Spiders were checked for deaths twice daily. Cannibalism left clear marks of partly or fully digested body parts and a stereomicroscope was used in case of doubt. An outcome of the experiment was re- corded when one of the two hatchlings was dead, due to starvation or cannibalism. Experiment 2. — The purpose of this ex- periment was to test if the tendency to can- nibalize varied among spiderlings from differ- ent eggsacs. Females of P. amentata with an eggsac were collected from the same location as in experiment 1, thus, we were only able to test for maternal effects on cannibalism and not for paternal effects. The spiders were tak- en to the laboratory and kept as in experiment 1. At 4 days of age, hatchlings from 19 egg- sacs were weighed. At day 6, hatchlings of approximately the same body mass were paired (mean body mass ± SE = 519 ± 3 pg; mean weight difference ± SE = 2.7 ± 0.2pg; max. weight difference = 13pg). In all pairs, the two hatchlings descended from the same eggsac {n = 205 pairs, 4-17 pairs from each eggsac). Experimental conditions and proce- dures were the same as in experiment 1. Data analysis. — Differences in the cause of death between kin and non-kin hatchling pairs were analyzed using the Pearson statistic. The latency to cannibalize between kin and non- kin were analyzed using Student’s t-test, after testing for equal variances (Bartlett’s Test, a > 0.05). Linear regression was used to test for correlation between mean body mass of pairs and the time spent before one of the two hatchlings died from either cannibalism or starvation. Regression lines were analyzed us- ing Analysis of Covariance (ANCOVA), with body mass as the covariate. First, we tested for equal slopes and if they were not signifi- cantly different, we tested if intercepts were equal and calculated means adjusted for the covariate (least squares means). We used lo- gistic regression to test if hatchlings from dif- ferent eggsacs differed in their probability to cannibalize or to die from starvation. All sta- tistical analyses were performed with IMP 5.0 for Windows (SAS Institute). 380 THE JOURNAL OF ARACHNOLOGY RESULTS Experiment 1. — The proportion of pairs resulting in a cannibalistic event was not af- fected by kinship, i.e. whether or not the two spiders in a pair originated from the same eggsac (Pearson = 0.26, P = 0.61; siblings 41 %, non-siblings 45 %, w = 142). Further- more, the time passing until a cannibalistic act occurred did not differ between sibling and non-sibling pairs (t-test, DF = 59, P > 0.80; siblings = 8.82 ± 0.40 days ± SE, non-sib- lings = 8.99 ± 0.49 days ± SE). Thus, we found no evidence that relatedness affected the cannibalistic tendency in hatchlings. In our description of general patterns of cannibalism below, we therefore pool data from siblings and non-siblings. Fig. 1 shows the relationship between av- erage body weight of hatchling pairs and the time until one of the two hatchlings died. The data indicate a presence of three different can- nibalistic strategies. One group of hatchlings never cannibalized (57%) and consequently died from other reasons than cannibalism. In this non-cannibalistic group, there was a pos- itive correlation between mean body mass and survival time of the first dying hatchling (lin- ear regression, t = 8.45, n = 79, P < 0.0001; R2 = 0.48). A similar type of positive corre- lation was found in pairs where cannibalism happened after 3 days of the experiment (33% of all pairs; linear regression, t = 6.95, n — 47, P < 0.0001; R2 = 0.52). The regression line of spiders cannibalizing after day 3 of the experiment (i.e. > 7 days old) and the regres- sion line of spiders dying from other reasons than cannibalism did not have significantly different slopes (ANCOVA, SS = 0.04, F = 0.03, P = 0.85, Fig. 1), but the intercepts of the two regression lines differed significantly (ANCOVA, SS = 4.53, F = 4.20, P = 0.04). The least squares means of survival days ad- justed for body mass showed that spiders can- nibalizing after day 3 (>7 days old), did so on average 0.4 days (i.e. less than 10 h) before equal sized spiders would die from other rea- sons than cannibalism (Fig. 2). Besides the two strategies where spiders either died or cannibalized in a size dependent way, 10% of the pairs cannibalized early, within the first 3 days of the experiment. Among these pairs, there was no correlation between mean body mass and the time passing until cannibalism occurred (linear regression, t = 0.93, w = 14, P = 0.37, R2 = 0.07). Experiment 2. — We found the same three cannibalistic patterns in this experiment as de- scribed from experiment 1 (Fig. 3). Either spi- ders did not cannibalize at all (64.4%); they cannibalized in a body mass dependent way (25.4%, linear regression, n = 52, t = 4.57, P < 0.0001, R2 = 0.29); or they cannibalized within the first day of the experiment (i.e. be- fore being 7 days old) regardless of body mass (10.2%, linear regression, n = 21, t = 0.18, P = 0.86, R2 - 0.002). Mother identity did not affect whether or not cannibalism occurred within a pair of hatchlings (logistic regression, Wald ~ 23.05, DF = 18, R = 0.19). However, when cannibalism did occur, the latency to do so varied significantly among hatchlings from different eggsacs, after correcting for the ef- fect of body mass (ANCOVA on the latency to cannibalize with mean body mass as cov- ariate, DF = 13, SS = 186.5, F = 5.20, P < 0.0001, Fig. 4); five eggsacs in which fewer than three pairs cannibalized were omitted from this analysis, thus, 14 eggsacs were in- cluded with a total of 65 pairs. DISCUSSION The results of this study indicate that three different cannibalistic strategies exist in the wolf spider hatchlings. Either we observed no cannibalism, late and size dependent canni- balism, or early and size-independent canni- balism. This pattern appeared in two separate experiments, which suggests that it is a gen- eral pattern of this wolf spider species. More than half of the spiderlings belonged to the group that never cannibalized and con- sequently died from other causes than canni- balism. As all spiderlings were deprived of food we expect that the main part of these non-cannibalizing spiders died from starva- tion. The body mass of an animal probably correlates positively with the amount of nu- trient reserves that are stored in the body. Fur- thermore, light animals are often found to have proportionally higher specific metabolic rate than heavier animals (Edwards 1946; Phillipson 1963). Together, these two factors may explain the observed pattern of lighter spiders dying from starvation sooner than heavier spiders. The spiders that did cannibalize could be HVAM ET AL.— CANNIBALISM AMONG WOLF SPIDER HATCHLINGS 381 81 Eggsacs Figure 4. — Effect of eggsac origin on the latency to cannibalize in pairs of equal sized siblings (Ex- periment 2). Points show the time passing until cannibalism occurred adjusted for the effect of body mass {n = 3--9 pairs of siblings per eggsac, five eggsacs were not included because less than 3 pairs of spi- derlings from these eggsacs cannibalized). divided in two groups: a group where the on- set of cannibalism was dependent on body mass, and a group, which cannibalized early. In the body mass dependent cannibalism ligh- ter pairs cannibalized earlier than heavier pairs, which suggests that the latency to can- nibalize depended on their level of nutrient reserves. In fact, this group of spiderlings gen- erally waited to cannibalize almost until the time when they were predicted to die from, starvation, which suggests that they chose to cannibalize as a very last option. In a rela- tively small proportion of spider pairs (ca. 10%) cannibalism appeared early in the ex- periment regardless of their body mass. This group of spiders did not seem to be under se- vere food stress when the cannibalism oc- curred, suggesting that these spiders had a higher keenness to cannibalize. Different can- nibalistic strategies among individuals within a species have also been observed in other animals. In salamanders (Laenoo et al. 1989) and spadefoot toad tadpoles (Pfennig et al. 1993) individuals can be divided into canni- balistic and non-canriibalistic forms and can- nibalistic individuals are often characterized by actual morphological and physiological differences that enhance this feeding strategy. Field studies have shown that conspecifics comprise a large part of the diet in juvenile and adult wolf spiders (Edgar 1969; Hallander 1970). However, in this experiment spider- lings were rather reluctant to cannibalize, even though they were kept in the same container with no escape possibilities. Why did the ma- jority of hatchlings refuse to cannibalize when the consequence of such a decision is death from starvation? Our experimental setup does not provide a clear answer to that question. One likely explanation is that they fear the cost of retaliation. The risk associated with attacking decreases as the asymmetry in body mass/size increases. Samu et al. (1999) found that the body mass 382 THE JOURNAL OF ARACHNOLOGY ratio between two juvenile spiders was the most important factor influencing cannibal- ism, and cannibalism was not observed within 24 hours if the body mass ratio was less than 2:1 (predatoriprey). Here we paired spider- lings of equal body mass, which in principle have similar predatory abilities and therefore provide roughly 50/50 chance of dying, unless there are different risks associated with being an attacker or a defender. The fact that a large proportion of the spiders postponed cannibal- ism almost until they died from starvation in- dicates that risk of retaliation or other factors inhibit cannibalism. It is possible that canni- balism occurred when the risk of dying from starvation had outweighed these risks. We cannot exclude the possibility that some of the cannibalistic events happened after one of the spiders was dead or almost dead from star- vation. If so, then cannibalistic acts should only confer little or no risk of retaliation. An- other potential cost of cannibalism is the risk of receiving pathogens from conspecific prey (Pfennig et al. 1998). If this is a real cost in the field, it would explain the general reluc- tance to cannibalize in the majority of the spi- ders. However, we are not aware of any path- ogens that might cause such a risk in wolf spiders, especially not among young hatch- lings. A general inhibition of cannibalism can be an indirect method to avoid eating relatives. Where such an inhibition has been demon- strated, it is often expressed in certain life stages. Filial cannibalism, for example, is in- hibited in reproductively active females of the wolf spider Scizocosa ocreata (Wagner 1995). Moreover, cannibalism was less frequent in the 2nd instar of the wolf spider Hogna helluo (Walckenaer 1837), compared to 3rd instar spiderlings (Roberts et al. 2003). Avoidance of related prey can also be direct through kin recognition where relatives are recognized and disregarded as prey (Pfennig 1997). There is one study that supports kin discrimination among young spiderlings in a wolf spider (Roberts et al. 2003). In this species a higher frequency of cannibalism was observed in pairs of non-siblings compared to pairs of sib- lings. In the present experiment, we did not find any evidence supporting the hypothesis that siblings cannibalized each other less fre- quently than non-siblings. Thus, either Par- dosa amentata hatchlings cannot recognize a sibling from a non-sibling, or they do not care and cannibalize nevertheless. These results are also in contrast to data on social (Diaea er- gandros Evans 1995) and sub-social (Stego- dyphus lineatus Latreille 1817) spiders (Evans 1999; Bilde & Lubin 2001), in which the stud- ies showed kin recognition and kin discrimi- nating cannibalistic behavior. Compared to solitary spiders, social spiders and sub-social spiderlings spend long periods of time close to relatives and it is possible that such fre- quent encounters with relatives are a require- ment for the evolution of kin recognition (Bil- de & Eubin 2001). Spiderlings of a clutch do not leave their mother’s abdomen at the same time but dispersal is distributed over several days and over a relatively large area (D. Mayntz, pers. obs.). Thus, the only time wolf spiders have a high chance of meeting siblings is when the spiderlings are gathered on their mother’s abdomen. Avoiding cannibalism of kin may possibly be accomplished during oth- er routes than actual kin recognition. For ex- ample, intra-brood cannibalism in Pardosa pseudocmnidata (Bosenberg & Strand, 1906) rarely occurred due to the small size differ- ence within the brood (lida 2003). Moreover, P. pseiidoannulata did not seem to cannibalize siblings less frequently than non-siblings (i.e. no evidence for kin recognition). When we tested for variation in cannibal- istic tendencies among hatchlings from differ- ent eggsacs, we did not find any evidence that maternal effects influenced whether or not cannibalism happened. However, when can- nibalism did occur, hatchlings from different eggsacs showed variable latencies to do so (Fig. 4). We collected the eggsacs in the field. This made it impossible for us to assess the genetic influence from the fathers, and pre- vented us from separating genetic effects from other maternal effects that might have affected the hatchlings’ tendency to cannibalize. Be- yond pure genetic factors, possible maternal factors affecting cannibalistic tendency may include the nutritional history of the mother, the age of mother, or size of the brood. Her- itability of cannibalistic behavior has been shown in fish, flour beetles, corn borers and ladybird beetles (Thibault 1974; Stevens & Mertz 1985; Tarpley et al. 1993; Wagner et al. 1999) but so far not in spiders. Half-sib ex- periments or actual selection experiments are needed before we can clarify how much ge- hvam et al=— cannibalism among wolf spider hatchlings 383 netic effects contribute to the observed vari= ation in the latency to cannibalize. ACKNOWLEDGMENTS We are indebted to S0ren Toft, Gitte Skov- iund and Jeroee Vandee Borre for valuable comments on the manuscript and to Else Bomholt Rasmussen for assistance in the lab- oratory. LITERATURE CITED Alderweireldt, M. & J.-R Maelfait. 1988. Life cy- cle, habitat choice and distribution of Pardosa amentata (Clerck, 1757) in Belgium (Araneae, Lycosidae). Bulletin de la Societe Scientifique de Bretagne 59:7--15. Anderson, J.E 1974. Responses to starvation in the spiders Lycosa lenta Hentz and Filistata hiber- nalis (Hentz). Ecology 55:576-585. Anthony, C.D. 2003. Kinship influences cannibal- ism in the wolf spider, Pardosa milvina. Journal of Insect Behavior 16:23-36. Bilde, T. & Y. Lubin. 2001. Kin recognition and cannibalism in a subsocial spider. Journal of Evolutionary Biology 14:959-966. Buddie, C.M., S.E. Walker, & A.L. Rypstra. 2003. Cannibalism and density-dependent mortality in the wolf spider Pardosa milvina (Araneae: Ly- cosidae). Canadian Journal of Zoology 81:1293- 1297. Edgar, W.D. 1969. Prey and predators of the wolf spider Lycosa lugubris. Journal of Zoology 159: 405-411. Edwards, G.A. 1946. The influence of temperature upon the oxygen consumption of several arthro- pods. Journal of Cellular and Comparative Phys- iology 27:53-64. Elgar, M.A, & B.J. Crespi. 1992. Ecology and evo- lution of cannibalism. Pp 1 — 12. In Cannibalism: Ecology and Evolution Among Diverse Taxa. (M.A. Elgar & B.J. Crespi, eds.). Oxford Uni- versity Press, New York. Evans, T.A. 1999. Kin recognition in a social spider. Proceedings of the Royal Society of London, Se- ries B"Biological Science 266:287-292. Hallander, H. 1970. Prey, cannibalism and micro- habitat selection in the wolf spiders Pardosa che- lata O. E Miiller and P. pullata Clerck. Oikos 21:337-340. lida, H. 2003. Small within-clutch variance in spi- derling body size as a mechanism for avoiding sibling cannibalism in the wolf spider Pardosa pseudoannulata (Araneae: Lycosidae). Popula- tion Ecology 45:1-6. Laneoo, M.J., L. Lowcock & J.P. Bogart, 1989. Sib- ling cannibalism in eoncaenibal morph Ambys- toma tigrinum larvae and its correlation with high growth rates and early metamorphosis. Ca- nadian Journal of Zoology 67:1911-1914. Pfennig, D.W. 1997. Kinship and cannibalism. Bio- science 47:667-675. Pfennig, D.W., S.G. Ho & E.A. Hoffman. 1998. Pathogen transmission as a selective force against cannibalism. Animal Behaviour 55: 1255-1261. Pfennig, D.W, H.K. Reeve & RW. Sherman. 1993. Kin recognition and cannibalism in spadefoot toad tadpoles. Animal Behaviour 46:87-94. Pfennig, D.W. & RW. Sherman. 1995. Kin recog- nition. Scientific American 272:68-73. Phillipson, J. 1963. The use of respiratory data in estimating annual respiratory metabolism, with particular reference to Leiobunum rotundum (Latr.) (Phalangiida). Oikos 14:212-223. Poiis, G.A. 1981. The evolution and dynamics of ietraspecific predation. Annual Review of Ecol- ogy and Systematics 12:225-251. Roberts, M.J, 1995. Spiders of Britain and Northern Europe. Harper Collins Publishers, London. Bath, England. Roberts, J.A., RW. Taylor & G.W. Uetz. 2003. Kin- ship and food availability influence cannibalism tendency in early-instar wolf spiders (Araneae: Lycosidae). Behavioral Ecology and Sociobiol- ogy 54:416-422. Samu, E, S. Toft & B. Kiss, 1999. Factors influ- encing cannibalism in the wolf spider Pardosa agrestis (Araneae, Lycosidae). Behavioral Ecol- ogy and Sociobiology 45:349-354, Stevens, L. & D.B. Mertz. 1985. Genetic stability of cannibalism in Tribolium confusum. Behavior Genetics 15:549-559, Tarpley, M.D., F. Breden & G.M. Chippendale. 1993. Genetic control of geographic variation for cannibalism in the southwestern corn borer, Dia- traea grandiosella. Entomologia Experimentalis et Applicata 66:145-152. Thibault, R.E. 1974. Genetics of cannibalism in a viviparous fish and its relationship to population density. Nature 251:138-140. Toft, S. & D.H. Wise. 1999. Growth, development, and survival of a generalist predator fed single- and mixed-species diets of different quality. Oec- ologia i ] 9: 191-197. Wagner, J.D. 1995. Egg sac inhibits filial cannibal- ism in the wolf spider, Schizocosa ocreata. An- imal Behaviour 50:555-557. Wagner J.D. & D.H. Wise. 1996. Cannibalism reg- ulates densities of young wolf spiders: Evidence from field and laboratory experiments. Ecology 77:639-652. Wagner, J.D., M.D. Glover, J.B. Moseley & A.J. Moore. 1999. Heritability and fitness conse- quences of cannibalism in Harmonia axyridis. Evolutionary Ecology Research 1:375-388. Manuscript received 22 November 2004, revised 17 June 2005. 2005. The Journal of Arachnology 33:384-389 DATA ON THE BIOLOGY OF ALOPECOSA PSAMMOPHILA BUCHAR 2001 (ARANEAE, LYCOSIDAE) Csaba Szinetar and Janos Eichardt: Department of Zoology, Berzsenyi College, Karolyi Caspar ter 4. Szombathely, H-9700 Hungary. E-mail: szcsaba@bdtf.hu Roland Horvath: Department of Evolutionary Zoology and Human Biology, University of Debrecen, FOB. 3, Debrecen, H-4010 Hungary ABSTRACT. This paper presents electron micrographs of the genitalia of Alopecosa psammophila, describes the morphological characteristics of the species and also gives information on its habitat pref- erence, the co-occurring ground-dwelling spiders, and the phenological characteristics of the species. Bar- ber pitfall trappings have been carried out since 2000 in dry sandy grasslands in three regions of Hungary: the Kiskunsag area (Kiskunsag National Park); the Nyirseg area (Hortobagy National Park); and since 2004 the Kisalfold area (Ferto-Hansag National Park). Specimens of the species, hitherto unknown in Hungary, have been collected from 17 localities in all three areas. We collected specimens in calciferous open sand steppes and in acidic open sand steppes. In the females, two activity periods were apparent (from April to end July and in October). A few males were collected in April and in October-November they had an extreme activity peak. We assume that the species has adult specimens throughout the winter. Alopecosa psammophila is most similar to Xysticus ninni Thorell, 1872 and Zelotes longipes (L. Koch 1866) in terms of its environmental needs. Keywords: Wolf spider, sandy grasslands, palpal organ, phenology, habitat preference The species Alopecosa psammophila Bu- char 2001 is known only from warm and dry sandy habitats of southern Moravia and from southern Slovakia (Buchar 2001). On the basis of the habitat characteristics of the holotype, and because they were found so close to Hun- gary, it was highly likely that the species would occur in Hungary, as dry sandy grass- lands occur in large areas in the Carpathian Basin. The ultimate goal of the investigations into Hungarian sandy grasslands was to explore the biology of the species in precise details. We wished to focus primarily on the pheno- logical characteristics, the habitat preference and the co-occurring spiders. In addition we also wished to publish pictures of the genitalia of the Hungarian specimens taken using a scanning electron microscope, as only draw- ings of the species have hitherto been known- in the international literature (Buchar 2001). METHODS Barber pitfall trappings have been carried out in nine sandy grasslands in the Kiskunsag area (Kiskunsag National Park — coordinates of the central site of the study area (Fiilopha- za) Lat. 19°24' N, Long. 46°52' E) and eight sandy grasslands in the Nyirseg area (Horto- bagy National Park — coordinates of the cen- tral site of the study area (Batorliget) Lat. 47°42' N, Long. 17°47' E) since 2000 as part of the project “Monitoring grasslands,” itself part of the national program Biodiversity Monitoring. The appropriate processing of the specimens collected in 2000 in the 170 ground traps, operated throughout the entire vegeta- tion season with 10 traps in each of the 17 habitats, provided us with an ample opportu- nity to examine the occurrence of Alopecosa psammophila. In addition to the two large regions under investigation since 2000, we commenced sim- ilar investigations in the sandy grasslands of the Small Hungarian Plain (Kisalfold, part of the Ferto-Hansag National Park — coordinates of the central site of the study area (Gonyu) Lat. 47°40' N, Long. 20°14' E), which lies in the northwestern part of the country. Figure 1 of Central Europe shows the type locality of the species, its habitat in Slovakia (Buchar 2001), as well as the sampling sites in Hun- gary. 384 SZINETAR ET AL.— BIOLOGY OF ALOPECOSA PSAMMOPHILA 385 Figure 1. — The occurrence of the Alopecosa psammophila in Central Europe. ■ = the type locality; A = sampling site in Slovakia; • == sampling sites in Hungary. !• = Kiskunsag National Park: 9 different study sites; 2 • = Nyirseg region: 8 different study sites; 3* = Small Hungarian Plain (Kisalfdld). Figures 2-5. — Genitalia of Alopecosa psammo- phila. 2. Left male palp, ventral view. In the Kiskunsag area we also paid close attention to the characteristics of the habitats. This way we had the opportunity to explore the relationship between the abundance of A. psammophila and the environmental variables including size of the habitat investigated, av- erage open sand surface, average coverage of lichens and mosses, average coverage of the sand and mosses, average coverage of the leaf-litter, average coverage of vegetation, av- erage vegetation height. We used multiple re- gression models to evaluate the effects of the seven variables on the number of individuals (Barta et al. 2000). For investigating which species have the closest habitat association with Alopecosa psammophila we used hier- archical cluster analysis (Tothmeresz 1993). The electron micrographs were made in the Hungarian Natural History Museum, Buda- pest (dr. Krisztina Buczko) with a HITACHI SN 2600 scanning electron microscope. The voucher specimens are deposited in the col- lection of HNHM. RESULTS Morphology. — Figs. 2-5 show electron micrographs of the genitalia of the Hungarian specimens. In both male and female genitalia 386 THE JOURNAL OF ARACHNOLOGY Figure 3. — Tegular apophysis and apical division. the three-dimensional structure of the organs carries the specific information that allows the recognition of the species. In the male palp the spear shape of the tegular apophysis can be understood by mentally combining the ven- tral (Fig. 2-3.) and retrolateral (Fig. 4) views. Occurrence in Hungary, — This species, hitherto unknown to Hungary, has been col- lected from 17 different habitats (sampling sites) in the three different areas. The 17 hab- itats examined yielded specimens of the spe- cies from 12 and 1 1 locations in 2001 and 2002, respectively. The species occurred in 16 out of the 17 Hungarian sandy grasslands in- vestigated. Add to this the sandy grasslands in the Kisalfold area which also yielded speci- mens of the species. Phenology. — In the collection period be- tween the beginning of April and the begin- ning of November 2001 (the longest collec- tion period within one calendar year), 46 females and 54 males were collected. In the case of the females, two activity periods were evident. The first period lasted from April to the end of July, culminating in the second half of May (Fig. 6). The peak coincided with the collection dates published by Buchar (2001). We also collected males in April, which also coincided with Buchar’s (2001) observations. In summer and early autumn periods there were no males in the samples, but in the sec- ond half of October and in November there was an extreme activity peak of males. We assume that the species has adult specimens throughout the winter. The November (2004) trappings also yielded males at Gonyu (Small Hungarian Plain). Habitat preference. — We collected the specimens in the calciferous open sand steppes (Festucetum vaginatae danubiale) in the area between the rivers Danube and Tisza and on the Small Hungarian Plain, and in the acidic open sand steppes of the Nyirseg area (Festuco vaginatae-Corynephoretum). We can conclude that the species is generally wide- spread in the ground-dwelling fauna of any dry sandy grassland in the Carpathian Basin. However, we were unable to find any signifi- cant relationships between its presence/ab- sence or its relative abundance and the mea- sured characteristics of the flora of the grasslands investigated (non-significant ef- fects of all seven variables), apart from it beeing a very strong indicator of sand. SZINETAR ET AL.— BIOLOGY OF ALOPECOSA PSAMMOPHILA 387 Figure 4. — Same as Fig. 3, retrolateral view. Spider communities* — Sandy grasslands seem to have rather similar spider communi- ties all over Hungary. These communities were characterised by specialist psammophiL ous species and were basically unaffected by wide regional separation and/or sand type. The dominant species at the study sites in- cluded Alopecosa cuneata (Clerck, 1757); Al~ opecosa cursor (Hahn, 1831); Alopecosa sulz- eri (Pavesi, 1873); Berlandina cinerea (Menge, 1872); Callilepis nocturna (Linnae- us, 1758); Gnaphosa mongolica Simon, 1895; Thanatus arenarius Thorell, 1872; Thanatus pictus L. Koch, 1881; Zelotes longipes (L. Koch, 1866). A. psammophila was the domi- nant species at one sampling site and ranked second, third, fourth, fifth or lower at other sites. The average relative frequency of A. psammophila was considerably higher in the Kiskunsag National Park (0.12 ± 0.14 (mean ± SD) than in the Nyirseg area (0.04 ± 0.05), but its dominance status was very variable 388 THE JOURNAL OF ARACHNOLOGY Figure 5. — Epigynum, ventral view. even within one region. For the cohabiting species that were present at least at 50% of all the sampling sites (5 locations), we carried out an association test. The results suggest that out of the cohabiting species A. psammophila shows the closest relationship with Xysticus ninnii Thorell 1872 and Zelotes longipes (L. Koch 1866) as far as their environmental needs are concerned (Fig. 7). Sampling dates Figure 6. — Phenology of Alopecosa psammophi- la based on pitfall samples collected at Fiilophaza in 2001. DISCUSSION Morphologically, Alopecosa psammophila can be well distinguished from the other Al- opecosa species in Central Europe by the highly specific three-dimensional shape of its genitalia, Phenologically, it is noted that the species shows the greatest activity in October and November. The co-occurring species of the genus (A. cursor, A. cuneata and A. sulz- eri) have their maturity season in the summer. An exception to this is Alopecosa accentuata (Latreille, 1817) which similarly to A. psam- mophila overwinters as adults; thus it has adult specimens in autumn, in spring and in early summer (Nentwig et al. 2003). On the basis of the investigations carried out so far, we conclude that A. psammophila is a species generally and frequently occurring in the sandy grasslands of Hungarian plains in the Carpathian Basin, and that it can even be the dominant species of the ground-dwelling spi- der communities in these habitats. In the case of sandy grasslands in the plains we found great differences in the coverage of vegeta- tion, coverage of lichens and mosses, cover- szinetAr et al.— biology of alopecosa psammophila 389 Zelotes decUmns Gnaphosa mongolica Berlandia cinerea Zelotes longipes Xysticus ninnii — i Alopecosa psammophila — ' Thanatus arenarius Xysticus kochi Alopecosa cursor Alopecosa cuneata o'o 01 02 03 04 Dissimilarity Figure 7. — The species in the closest association with Alopecosa psammophila in the Great Hungarian Plain (the Matusita index of similarity and the Ward-Orloci fusion method were used). age of the sand and mosses, and in vegetation height, but these differences seemed not to af- fect the abundance of A. psammophila. We as- sume that the species is widespread in the dry sandy grasslands of Central Europe, ACKNOWLEDGMENTS The authors are grateful to Dr. J. Buchar for checking the A. psammophila specimens and wish to thank Dr. Viktor Marko, the organizer of the national program, Biodiversity Moni- toring. We are grateful to Dr. Krisztina Bucko and Dr. Tamas Szuts for their help in creation of the scanning micrographs; to Dr. Tibor Ma- gura, Viktor Kodobocz and to Ferenc Pal Sza- bo for their participation in the fieldwork; to Eva Kovacs and Balazs Deak for providing us with botanical information. We also wish to thank the Organizing Committee of XVII ICA for supporting our participation on the XVII ICA. The authors thank, with gratitude, the Directorates of the Kiskunsag, the Hortobagy, and the Ferto-Hansag National Parks for their permission to carry out the surveys. Csaba Szinetar was supported by a Bolyai Fellow- ship of the Hungarian Academy of Sciences. LITERATURE CITED Barta, Z., I. Kassai & T Szekely. 2000. Alapveto kutatastervezesi, statisztikai es projektertekelesi modszerek a szupraindividualis bioldgiaban. Kossuth Egyetemi Kiado, Debrecen. 163 pp. Buchar, J. 2001. Two new species of the genus Al- opecosa (Araneae: Lycosidae) from south-east- ern Europe. Acta Universitas Carolinae Bioloica 45:257-266. Nentwig, W., A. Hanggi, C. Kropf & T Blick. 2003. Central European Spiders — Determination Key. http//www. araneae . unibe . ch/index . html version 8.01. 2003. Tothmeresz, B. 1993. NuCoSA 1.0: Number for Cruncher for Community Studies and other Eco- logical Applications. Abstracta Botanica 17:283- 287. Manuscript received 4 January 2005, revised 1 September 2005. 2005. The Journal of Arachnology 33:390-397 SIZE DEPENDENT INTRAGUILD PREDATION AND CANNIBALISM IN COEXISTING WOLF SPIDERS (ARANEAE, LYCOSIDAE) Ann L. Rypstra: Department of Zoology, Miami University, 1601 Peck Blvd. Hamilton, Ohio 45011 USA. E-mail: RypstraL@muohio.edu Ferenc Samu: Department of Zoology, Plant Protection Institute, Hungarian Academy of Sciences, PO Box 102, Budapest, H-1525 Hungary ABSTRACT. Two species of wolf spider, Hogna helluo (Walckenaer 1837) and Pardosa milvina Hentz 1 844 dominate the predatory community on the soil surface of agroecosystems in eastern North America. Although as adults they are very different in size, differences in phenology ensure that they overlap in size at various times during the year. In a laboratory experiment, we explored the propensity of each species to attack and kill the other wolf spider species (intraguild predation), conspecifics (cannibalism) or crickets (ordinary predation). Both spiders attacked and killed a broader size range of crickets more quickly than they approached other spiders. We found no differences in Hogna foraging on conspecifics or Pardosa, but Pardosa attacked and killed Hogna more readily than conspecifics. Because Hogna was so slow in attacking other spiders, their impact as an intraguild predator may be quite small, especially if their approach to crickets is an indication of their predatory tendencies with insects. On the other hand, Pardosa attacked and killed small Hogna as readily as crickets, which suggests they may have an influence on Hogna populations if Hogna young emerge coincident with large juvenile or adult Pardosa. Keywords: Cannibalism, intraguild predation, agrobiont spiders, predator-prey Cannibalism and intraguild predation (IGP) are important to spider communities and have the potential to affect population sizes and/or species diversity of spiders as well as that of potential insect prey (Wagner & Wise 1996; Hodge 1999; Samu et al. 1999; Finke & Den- no 2002; Matasumura et al. 2004; Denno et al. 2004). Predation is a dynamic process, the outcome of which depends on the relative siz- es of the predator and prey, their physiological state, attack strategy and inherent aggressive- ness (Walker et al. 1999; Persons et al. 2001; Buddie 2002; Balfour et al. 2003; Buddie et al. 2003; Mayntz et al. 2005). Many of these factors will shift over time both with age and recent experience and thus the relative impor- tance of cannibalism and/or IGP to foraging individuals, population structure and commu- nity composition will shift as well (Wagner & Wise 1996; Balfour et al. 2003; Buddie et al. 2003). For spiders that coexist, an understand- ing of the situations under which cannibalism and IGP occur is critical to understanding how and when they can persist in the same habitat. In the present study we explore the preda- tory tendencies of two species of wolf spider (Araneae, Lycosidae) that coexist on the soil surface in agricultural fields across the eastern portion of North America. Because the species differ in size, activity, and phenology, we wanted to characterize the circumstances un- der which these spiders engaged in cannibal- ism or intraguild predation and compare those predatory interactions to attacks on insect prey. Under controlled laboratory conditions, we paired a wide size range of individuals with conspecifics, the other species of spider, or crickets and documented the outcome and timing of predation. In this way, we hoped to gain a better understanding of the specific predatory strategy of each of the spider spe- cies and the relative influence that these spe- cies have on their insect prey, which would help us to gain insight into the nature of their co-existence. METHODS Study species. — Hogna helluo (Walckenaer 1837) and Pardosa milvina Hentz 1844 co- exist on the soil surface in disturbed riparian habitats and agroecosystems throughout the 390 RYPSTRA & SAMU— PREDATION & CANNIBALISM IN WOLF SPIDERS 391 Figure 1. — Mean prey to predator mass ratio (PPR ± S.E.) for captured prey vs. the time (min ± S.E.) it took the prey to be captured. Trials where Hogna was predator are indicated by solid squares and those where Pardosa was predator are indicated by open squares. Specific prey types are listed with an arrow pointing to the data for that treatment. eastern portion of North America (Dondale & Redner 1990; Marshall & Rypstra 1999; Mar- shal! et al. 2002). Pardosa is small (20 mg), active, and can be found at high densities (10- 15 per m^) whereas Hogna is large (800 mg), less active, and found at relatively low den- sities (1-2 per wd) in soybean fields in the midwestern section of the United States (Mar- shall & Rypstra 1999; Walker et al. 1999; Marshall et al. 2002). Pardosa is an annual species with a mid-July population peak. Ex- cept for a relatively short period during which the adults and spiderlings co-occur, the size distribution of Pardosa individuals active in the fields at any given time is fairly narrow (Marshall et ak 2002). On the other hand, Hogna seems to have a two-year life cycle with more stages occurring in the fields at the same time (Marshall et al. 2002). Although Hogna are usually larger than Pardosa, be- cause of the variability in their life cycle, it is possible for large subadult or adult Pardosa to coexist with early stages of Hogna. Previ- ous studies have revealed that each species readily consumes smaller individuals of the other in the laboratory (Persons et al. 2001; Balfour et al. 2003). Here we explore those predatory interactions systematically across a broad range of size ratios. Both species of spiders were collected from com and soybean fields at the Miami Univer- sity Ecology Research Center (Oxford, Butler County, Ohio, USA) and held in the labora- tory or reared from animals collected at that site. When not involved in experimentation, spiders were housed individually in translu- cent plastic cylindrical containers 8 cm in di- ameter with 5 cm walls with 1—2 cm of damp peat moss covering the bottom. Spiders were watered and fed once or twice weekly on a diet of crickets (Acheta domesticus), fruit flies {Drosophila spp.) and or meal worms (Tene- brio spp.). Containers with spiders were held in an environmental chamber between 23 — 25 °C on a 12:12 L:D cycle at 60-75% humidity. Experimental protocol. — Spiders were randomly selected from the laboratory popu- lation and brought to standard hunger levels by feeding them ad libitum with Drosophila melanogaster for 2 days. Spiders were then held for 7 days before testing to ensure that they were similarly hungry. Spiders were ran- domly assigned to be paired with conspecifics (to monitor cannibalism), heterospecifics (to monitor intraguild predation) or crickets (to monitor ordinary predation). Those assigned to be paired with conspecifics were marked with a drop of acrylic paint on the abdomen or cephalothorax so that we could identify in- dividuals. All spiders and crickets were weighed and then introduced into a testing arena simultaneously. The arenas consisted of 14 cm diameter Petri dishes with a base of dampened plaster of Paris (as in Samu et al. 1999). Animals were allowed to interact in the arena for 24 h during which time we recorded if and when predation occurred. Experiments were run in groups that included representa- tives of all treatments between July 1998 and July 2001. Statistical analysis. — In order to determine how similar the spiders and insects used in each treatment were, we compared the mass of predators and prey across all treatments in ANOVAs. In addition, we calculated prey/ predator mass ratio (PPR) by dividing the mass of the prey by the mass of the predator. In cases where there was no predation, we randomly assigned one of the spiders as prey and the other as predator using a coin toss algorithm. In order to ensure pairings were similar across treatments, PPRs were also compared in an ANOVA. The effects of pred- ator species, prey type, and PPR on the fre- quency of predation were compared using a 392 THE JOURNAL OF ARACHNOLOGY Q. CL Prey type Figure 2. — Fifty percent lethal mass ratio (LR50) (± 95% confidence interval) for Hogna (on the left) and Pardosa (on the right). Prey type is on the X-axis. Where treatments are indicated with the same small letter in the middle of each figure, the overlap of the 95% confidence ranges suggests that there was no significant difference. logistic regression analysis. From the logistic regression, we determined the PPR at which there was a 50% likelihood of a predatory event (LR50). Differences in LR50s across treatments were evaluated by comparison of the 95% conhdence intervals. The effects of the same factors (predator species, prey type, and PPR) on the time until predation were evaluated using a parametric survival analysis using the Proportional Hazards model. In this case pairwise comparisons were made using the Bonferroni test with an overall P-value of 0.05. Both the logistic regression and survival models were run initially with all interactions included. The non-significant interactions were removed after the first run and the mod- els were run again. We also wanted to determine if any of the observed preferences were due to size or if they had to be attributed to some other quality of the prey (e.g. nutrition, taste). We hypoth- esized that, if the preference was size related, then there would be a size ratio (PPR^.,.iticai) be- low which predators would not discriminate between prey types. We defined PPR^riticai the maximum PPR value where the differenc- es between predation on two prey types were no longer significantly different {P = 0.05). In order to find the PPR^,,.,ticab we started by removing the sample with the highest PPR and rerunning the statistical test, if it was still significant, we removed the sample with the next highest PPR, and ran the test again. We continued this process until the P-value as- sociated with any difference was equal to 0.05. RESULTS Overall there were no differences in the mass of the Hogna, Pardosa, or crickets used in our treatments (Predator mass F = 1.81, df - 5, 401, P = O.Il; Prey mass F = 1.02, df = 5, 401, P = 0.40) (Table 1). In addition, animals were paired so that the PPR values were similar across treatments (F = 1.71, df = 5, 401, P = 0.13) (Table 1). Even though not significantly different, the PPR for Hogna on crickets was somewhat higher than the oth- RYPSTRA & SAMU— PREDATION & CANNIBALISM IN WOLF SPIDERS 393 HOGNA AS PREDATOR TIME {MIU) PARDOSA AS PREDATOR TIME (MIN) Figure 3. — Capture success over time of Hogna and Pardosa on the three different prey types. Those indicated with the same letter (at the right of the line) were not significantly different using Bon- ferroni comparisons with a critical value of 0.05. ers. To some degree, this variation was inten- tional as we attempted to observe interactions across the complete prey size range that each spider would take. Note that even though the PPR is close to one, meaning that the prey were the same size as the predator, the capture rate is still very high (94.3%) as compared to other treatments (Table 1). Predator species, prey species and the PPR all affected the oc- currence and the timing of predation in com- plex ways (Table 2). Predation on crickets vs. spider prey.- — Both Hogna and Pardosa had higher capture success on crickets than on spiders {Pardosa: X^95 = 10.19, P = 0.001; Hogna: x^209 = 54.54, P < 0.0001; Table 1). Both species killed larger crickets than they killed other spiders (higher PPR) (Fig. 1). However at PPRs less than 0.54 for Pardosa (PPRcriticai, X^97 = 3.82, P = 0.05) and 0,57 for Hogna (PPR^riticab xV ~ 3.46, P = 0.05) there were no differences between spider vs, cricket prey. The process by which we generated the PPR^iticai inevitably resulted in a loss of sam- ple size, and therefore power, however, we re- gard the remaining case numbers still large enough (Np^^dosa = 98, Nnogna = 75) to draw valid conclusions on PPRcnticai values. The PPR at which there was a 50% chance of a predatory event (LR50) was significantly higher for crickets than for spider prey (Fig. 2) . Likewise, overall crickets were killed more quickly than other spiders (Figs. 1, 3). Comparing the predation strategy of Hogna and Pardosa, — The two predators dif- fered in their responses to the prey types test- ed (Table 2). Hogna consistently took larger prey than Pardosa from every prey category (Figs. 1, 2). On the other hand, Pardosa was consistently faster than Hogna in taking every prey type (Figs. 1, 3). Pardosa was more like- ly to capture Hogna than conspecifics — 5.53, P = 0.018, Table 1) but there was no difference in the capture rate of Hogna on ei- ther spider species (x^oi = 0.27, NS; Table 1). Likewise, the LR50 was larger for Pardosa preying on Hogna than it was for Pardosa cannibalism (Fig. 2) but there was no differ- ence in the LR50 for Hogna preying on het- erospecifics or conspecifics (Fig. 2). Similarly Pardosa captured Hogna more quickly than it captured other Pardosa but there were no dif- ferences in the Hogna'^ predatory speed on conspecifics or heterospecific spiders (Figs. 1, 3) . DISCUSSION Clearly these two spider species, Hogna helluo and Pardosa miivina, differ in their for- aging behavior across the various sizes of the different prey types tested here. Hogna is gen- erally slower to attack and kill a potential prey but generally take prey in larger size classes than Pardosa. Although Hogna differentiated between crickets and spiders, they did not seem to differentiate between conspecifics and a common coexisting intraguild predator, Par- dosa. On the other hand, Pardosa reacted dif- ferently to all three prey types; killing larger crickets faster than they killed Hogna and kill- 394 THE JOURNAL OF ARACHNOLOGY Table 1. — Summary of the sample size, capture frequency, predator mass (±S.E.), prey mass (±S.E.), and prey to predator mass ratio (PPR ± S.E.) for each treatment. Number Predator Treatments n captured (%) mass (mg) Prey mass (mg) PPR Hogna on crickets 70 66 (94.3%) 14.4 ± 1.2 14.6 ± 1.9 0.97 ±0.11 Hogna on Hogna 78 34 (43.6%) 19.2 ± 1.2 1 1.7 ± 1.7 0.84 ± 0.07 Hogna on Pardosa 54 28 (48.1%) 19.5 ± 1.4 9.9 ± 2.1 0.63 ± 0.05 Pardosa on crickets 64 50 (78.1%) 19.6 ± 1.3 11.1 ± 1.8 0.75 ± 0.07 Pardosa on Hogna 74 44 (59.5%) 17.6 ± 1.2 9.2 ± 1.8 0.64 ± 0.06 Pardosa on Pardosa 66 38 (42.4%) 19.1 ± 1.3 12.2 ± 1.9 0.68 ± 0.06 All Groups 406 260 (64.1%) 18.4 ± 1.6 1 1.4 ± 2.1 0.76 ± 0.03 ing larger Hogna faster than they killed con- specifics. Predation on crickets vs. spider prey. — Both cannibalism and IGP have been exten- sively documented in wolf spiders (Wagner & Wise 1996; Samu et al. 1999; Balfour et al. 2003; Buddie et al. 2003). Because these in- teractions carry with them an increased risk of injury and/or reciprocal predation, we expect- ed a different, and perhaps more cautious, predatory approach to other spiders when compared to crickets. We reasoned that the relative size of the prey to its predator (PPR) would be one measure of risk and we found that the PPR at which there was a 50% chance of a predatory event was much higher for crickets than spider prey (Fig. 2). However further exploration of the data reveals that, for both spider species, there was a PPR^.^iti^-a, be- low which there were no differences in the rate of predation on spiders as compared to crickets. Thus, the significant differences in predatory strategy that we observed were due to behavioral shifts that occurred when the prey were large relative to the predator. These results confirm that relative size was more im- portant for spider on spider contests than for attacks of crickets and suggests that both spi- der species were sensitive to the risk that a large predatory prey item might pose. This connection may be particularly true for Hog- na, which easily subdued large crickets but were much slower to take smaller individuals of either spider species (Fig. 2). Even though risk may be important to the observed differences in predation frequency, there may be other reasons for spiders to pre- fer insect over spider prey. It has been argued that organisms feeding on the same trophic level, and especially conspecifics, provide nu- trients in proportions that are more closely aligned with the predator’s nutritional needs (Polls 1981; Wildy et al. 1998; Fagan et al. 2002), however several studies have demon- strated that growth and survival of wolf spi- ders is lower when maintained on spider diets than when provided with insect prey (Toft & Wise 1999; Oelbermann & Scheu 2002; Mat- sLimura et al. 2004). Another reason not to eat closely related species is that they may carry pathogens that can invade more easily when consumed by a conspecific or phylogenetical- ly close host (Pfennig, et al. 1998; Pfennig 2000; MacNeil et al. 2003). Thus, selection may favor preferences for non-spider prey. Of course wolf spiders behave differently from crickets, which may have reduced their susceptibility to capture. We attempted to con- trol the circumstances of the interaction so that the predator had access to the same kind of sensory information in a confined space, which should minimize the small differences in capture and escape tactics. Nevertheless, it is impossible to totally uncouple the prey pref- erences and ease of capture from the specific signals by which the predator detects and | identifies prey items (Uetz 2000; Uetz & Rob- erts 2002). Thus a further exploration of the role of specific sensory modalities in the pred- ator interactions of these species is warranted. Comparing the predation strategy of Hogna and Pardosa. — A variety of differenc- es between the foraging strategies of Pardosa j and Hogna have been documented (Walker et S al. 1999; Walker & Rypstra 2002) and this j study clarifies some additional aspects of those differences. In particular, although Hog- j na was the most effective predator on crickets, j Pardosa distinguished between the three prey types in the proportion (Table 1), size (Figs. \ RYPSTRA & SAMU— PREDATION & CANNIBALISM IN WOLF SPIDERS 395 Table 2. — Results of logistic regression to predict prey capture and the results of the proportional haz- ards survival model to predict the time it took the spiders to capture prey. Both models used predator species, prey species and prey to predator mass ra- tio (PPR) as predictors. Source df Chi squared P Logistic regression model for outcome Whole model 7 299.782 <0,0001 Predator species 1 9,333 0,0023 Prey species 2 66.809 <0,0001 PPR 1 82,225 <0,0001 PPR * predator 1 20.895 <0.0001 PPR * prey 2 21,807 <0.0001 Survival model for time until capture Whole model 9 401.4444 <0.0001 Predator species 1 8.534 0.0035 Prey species 2 189,060 <0.0001 PPR 1 302.410 <0.0001 PPR * predator 1 36.730 <0.0001 PPR * prey 2 85.380 <0.0001 Predator * prey 2 14.433 0.0007 1, 2) and timing (Figs. 1, 3) of predation. Par- dosa is a much more active species than Hog- na (Walker et ah 1999; Walker & Rypstra 2002), which might have caused us to predict that they would be more susceptible to pre- dation by other wolf spiders which use motion to detect prey (Persons & Uetz 1997). How- ever, there is no evidence here to demonstrate that activity made Pardosa any more suscep- tible to the sit and wait predator, Hogna. In fact, it appears here that activity translated into effective search behavior that increased Pardosa\ ability to detect and attack more sluggish arthropod prey such as Hogna. Although not significantly different, the PPRs for Hogna paired with crickets were somewhat higher than the other pairings be- cause of our desire to cover the full size range of prey that each spider v/ould attack. Thus we considered whether the longer capture times observed for Hogna on crickets (Fig. 1) might be due primarily to the fact that they were tested with larger prey items. However, if we compare the mean capture time for crickets larger than the Hogna (PPR > 1,0; n = 19) with the capture times for those prey smaller than the Hogna (PPR < LO; « = 51), there was no difference {t = 2.0, P ~ 0.24). This fact furthers the characterization of Hog- na as a slow selective predator that, in the context of the options offered here, prefers large harmless prey. On the other hand, Par- dosa was generally faster to attack and dis- criminated more finely between the three prey options we included in this study. Implications for species co-existence in the field. — These results may be especially important for agrobiont spiders, such as Hog- na and Pardosa, as they may be important agents of biological control. The fact that crickets were more susceptible to predation across a much larger size range than spider prey suggests that the influence that two wolf spider species have on one another may not be exceedingly strong when alternative insect prey are abundant. To fully assess the field importance of these interactions, our findings need to be interpreted in the context of the life history of natural populations. Unfortunately, in spite of existing field surveys (Marshall & Rypstra 1999; Marshall et al. 2002), the life histories of the two species seem to be highly variable and, as a result, are not well enough understood to be predictable. Nevertheless, the available data suggest that Hogna typical- ly has a two or three-year life cycle with sev- eral juvenile stages coexisting and Pardosa is an annual species with a narrower size dis- tribution at any given time in the season. Thus, all life stages of Pardosa have the po- tential of coexisting with larger Hogna where- as only when Hogna spiderlings emerge at times of the year when large juveniles and adult Pardosa are around, do Hogna face pre- dation risk from Pardosa. Although the trials suggest that Hogna exert modest predatory pressure on both conspecifics and Pardosa, their attacks were very slow (Figs. 1, 3). As a consequence, Hogna may not exert much predation pressure on a quick wolf spider like Pardosa that in an open field situation could run away. On the other hand, Pardosa appear to prey on small Hogna as quickly as on crick- ets, so Hogna that emerge and attempt to go through the first few iestars during the early summer, when Pardosa are adults, maybe se- verely impacted by Pardosa predation. Clear- ly further explorations of the interactions be- tween these two species in more natural situations are required to fully quantify their influence on one another, the nature of their coexistence and their potential role in the eco- system. 396 THE JOURNAL OF ARACHNOLOGY ACKNOWLEDGMENTS We would like to thank M. Brueseke, E. Henley, C. Burkett and C.M. Buddie for aS“ sistance with execution of experiments. We are grateful to S. Marshall, M. Persons, E Channell, R. Balfour, S.E. Walker, B. Reif, C. Weig, and A. Tolin for ensuring that we had spiders to work with when they were needed. S. Wilder, J. Riem, J. Schmidt and other mem- bers of the Miami University spider lab pro- vided useful suggestions on early drafts of this manuscript. E Samu was Bolyai Fellow of the Hungarian Academy of Sciences whilst this work was conducted. Funding was provided by the following sources: OTKA (Hungary) Grants No. T048434 & F030264, NKFP (Hungary) project No. 6/00013/2005; NSF (USA) grants DEB 9527710, DBI 0216776 & DBI 0216947; Miami University’s Philip and Elaina Hampton Fund for Faculty Internation- al Initiatives & International Visiting Scholar Fund. LITERATURE CITED Balfour, R. A,, C. M, Buddie, A. L. Rypstra, S, E. Walker & S. D. Marshall. 2003. Ontogenetic shifts in competitive interactions and intra-guild predation between two wolf spider species. Eco- logical Entomology 28:25-30. Buddie, C.M. 2002. Interactions among young stag- es of the wolf spiders Pardosa moesta and P. mackenziana (Araneae, Lycosidae). Oikos 96: 130-136. Buddie, C.M., S.E. Walker & A.L. Rypstra. 2003. Cannibalism and density-dependent mortality in the wolf spider Pardosa milvina (Araneae: Ly- cosidae). Canadian Journal of Zoology 81:1293- 1297. Denno, R.E, M.S. Mitter, G.A. Langellotto, C. Grat- ton & 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. Dondale, C. D. & J. H, Redner. 1990. The Insects and Arachnids of Canada and Alaska. Part 17. The Wolf Spiders, Nurseryweb spiders, and Lynx spiders of Canada and Alaska (Araneae: Lycos- idae, Pisauridae and Oxyopidae). Biosystematics Research Centre, Ottawa, Canada. Fagan, W.F., E. Seimaen, C. Mitter, R.E Denno, A.F. Hubery, A.H. Woods & J.J. Elser. 2002. Ni- trogen in insects: Implications for trophic com- plexity and species diversification. American Naturalist 160:784-802. Finke, D.L. & R.F. Denno. 2002. Intraguild preda- tion diminished in complex-structured vegeta- tion: implications for prey suppression. Ecology 83:643-652. Hodge, M.A. 1999. The implications of intraguild |, predation for the role of spiders in biological I control. Journal of Arachnology 27:351-362. [ Langelotto, G.A. & R.F. Denno. 2004, Response of invertebrate natural enemies to complex-struc- tured habitats: a meta-analytical synthesis. Oec- ologia 139:1-10. MacNeil, C., J.T.A. Dick, M.J. Hatcher, N.J. Field- ing, K.D. Hume & A.M, Dunn. 2003. Parasite transmission and cannibalism in an amphipod (Crustacea). International Journal for Parasitolo- gy 33:795-798. Marshall, S. D., D. M. Pavuk & A. L, Rypstra. 2002. A comparative study of phenology and daily activity patterns in the wolf spiders Par- dosa milvina and Hogna helluo in soybean agroecosystems in southwestern Ohio (Araneae, Lycosidae). Journal of Arachnology 30:503-510. Marshall, S. D. & A. L, Rypstra. 1999. Patterns in the distribution of two wolf spiders (Araneae: Lycosidae) in two soybean agroecosy terns. En- vironmental Entomology 28:1052-1059. Matsumura, M., G.M. Trafelet-Smith, C, Gratton, i D.L. Finke, W.F. Fagan & R.F. Denno. 2004. Does intraguild predation enhance predator per- ; formance? a stoichiometric perspective. Ecology 85:2601-2615. Mayntz, D., D. Raubenheimer, M. Salomon, S. Toft, & S.J. Simpson. 2005, Nutrient-specific foraging in invertebrate predators. Science 207:111-113. I Oelbermann, K. & S. Scheu. 2002. Effects of prey type and mixed diets on survival, growth and development of a generalist predator, Pardosa lu- gubris (Araneae: Lycosidae). Basic and Applied Ecology 3:285-291. Persons, M.H. & G.W. Uetz. 1997. The effect of prey movement on attack behavior and patch res- idence decision rules of wolf spiders (Araneae: Lycosidae). Journal of Insect Behavior 10:737- : 752. 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-as- sociated predator cues (Araneae: Lycosidae). An- imal Behaviour 61:43-51. Pfennig, D.W. 2000. Effect of predator-prey phy- i logenetic similarity on the fitness consequences of predation: A trade-off between nutrition and disease? American Naturalist 115:335-345. Pfennig, D.W., S.G. Ho & E.A. Hoffman. 1998. Pathogen transmission as a selective force i against cannibalism. Animal Behaviour 55: 1255-1261. Polis, G.A. 1981. The evolution and dynamics of intraspecific predation. Annual Review of Ecol- ogy and Systematics 12:225-251. Samu, E, S. Toft & B. Kiss. 1999. Factors influ- ' RYPSTRA & SAMU— PREDATION & CANNIBALISM IN WOLF SPIDERS 397 encing cannibalism in the wolf spider Pardosa agrestis (Araneae, Lycosidae). Behavioral Ecol- ogy and Sociobiology 45:349-354. Toft, S. & D.H. Wise. 1999. Growth, development, and survival of a generalist predator fed single- and mixed-species diets of different quality. Oec- ologia 119:191-197. Uetz, G.W. 2000. Signals and multi-modal signal- ing in spider communication. Pp. 387-405. In Animal Signals: Signaling and Signal Design in Animal Communication (Y. Espmark, T Amund- sen & G. Rosenquvist, eds.). Tapir Publishers, Trondheim, Norway. Uetz, G.W. & J.A. Roberts. 2002. Multisensory cues and multimodal communication in spiders: Insights from video/audio playback studies. Brain, Behavior and Evolution 59:222-230. Wagner, J. D. & D. H. Wise. 1996. Cannibalism regulates densities of young wolf spiders: Evi- dence from field and laboratory experiments. Ecology 77:639-652. Walker, S. E., S. D. Marshall, A. L. Rypstra & D. H. Taylor. 1999. The effects of hunger on loco- motory behaviour in two species of wolf spider (Araneae, Lycosidae). Animal Behaviour 58: 515-520. Walker, S.E. & A.L, Rypstra. 2002. Sexual dimor- phism in trophic morphology and feeding behav- ior of wolf spiders (Araneae: Lycosidae) as a re- sult of differences in reproductive roles. Canadian Journal of Zoology 80:679-688. Wildy, E.L., D.P. Chivers, J.M. Kiesecker & A.R. Blaustein. 1998. Cannibalism enhances growth in larval long-towed salamanders {Ambystoma macrodactylum). Journal of Herpetology 32: 286-289. Manuscript received 12 January 2005, revised 15 August 2005. 2005. The Journal of Arachnology 33:398-407 REVIEW OF THE ORIENTAL WOLF SPIDER GENUS PASSIENA (LYCOSIDAE, PARDOSINAE) Pekka T« Lehtinen: Centre for Biodiversity, University of Turku, 20014 Turku, Finland. E-mail: pekleh@utu.fi ABSTRACT, The pardosine genus Passiena Thorell 1890 is redefined and relimited. Passiena has ex- cellent diagnostic characters, in particular the male pedipalp that carries a unique group of soft spicules on the distal part of the palea. The female of the type species, Passiena spinicrus Thorell 1890 from Malaysia, is illustrated for the first time. A new species, P. torbjoerni, is described from Thailand. All specimens of Passiena were collected from the ground layer of or nearby dense jungle or bush, an exceptional habitat for Oriental Pardosinae. Males of P. torbjoerni carry modified setae on the ventral side of the abdomen, similar to Hygrolycosa rubrofasciata (Ohlert 1865) and Pardosa sphagnicola (Dahl 1908), where they play an important role in the courtship behavior of males. Five African species currently listed in Passiena do not conform to the generic diagnosis as provided here. Three of these show clear affinities with Pardosa C.L. Koch 1847 and are consequently transferred from Passiena: Pardosa praepes (Simon 1885); Pardosa elegantula (Roewer 1959) new combination; and Pardosa upembensis (Roewer 1959) new combination. Passiena auberti (Simon 1898) and Passiena albipalpis Roewer 1959 are con- sidered incertae sedis pending a generic revision of African Lycosidae as they cannot be placed with certainty into any other lycosid genus. Keywords: Taxonomy, systematics, Pardosinae, ventral abdominal setae, acoustic communication Since its original description, the systematic position of the wolf spider genus Passiena Thorell 1890 has remained problematic. The knowledge of characters of the male pedipalp is crucial for an inteipretation of lycosid re- lationships (e.g. Dondale 1986; Zyuzin 1993) but only the female of the type species was known. In his revisionary work on worldwide lycosids, Roewer (1955, 1959, 1960) neglect- ed the importance of characters of the copu- latory organs at generic level, and a strong emphasis on minor details in variable somatic characters (especially eye pattern, cheliceral dentition and spination of legs) led to numer- ous, ill-founded taxonomic changes to the ly- cosid classification. For example, purely the presence of more than three pairs of ventral spines on the tibiae led him to the suggestion that Passiena from Oriental region should be synonymized with the subarctic-alpine Pa- laearctic genus Acantholycosa Dahl 1908 (Roewer 1959). Simon (1898) listed Passiena as synonym of Pardosa C.L. Koch 1847, noting the sim- ilarity of the type species, P. spinicrus Thorell 1890, with Pardosa bifasciata (C.L. Koch 1834) and Pardosa auberti Simon 1898. Sub- sequently, Roewer (1955) wrongly attributed to Simon (1898) the inclusion of the latter two species in Passiena (see also Tongiorgi 1966) and added Pardosa schenkeli Lessert 1904, a close relative of R. bifasciata, to Passiena. Bonnet (1958) synonymized Passiena with Pardosa, probably without personal study of the type species. Roewer (1959) did not ac- cept this synonymy, listed P. auberti and Par- dosa praepes Simon 1885 in Passiena and de- scribed three new species from Africa, P. albipalpis Roewer 1959, P. elegantula Roew- er 1959 and P. upembensis Roewer 1959. Tongiorgi (1966) was the first to note that the genital organs of the group around P. bisfas- data were different from the African species of Passiena sensu Roewer (1959) when he in- cluded P. bifasciata and P. schenkeli in his revision of Italian Pardosa. Tanaka (1993) listed Passiena as a junior synonym of Par- dosa without justification. This was not ac- cepted by Platnick (2005) who, prior to this study, included six species in Passiena: P. spinicrus, P. auberti, P. praepes and the three species described by Roewer (1959). The aim of this study is to provide a mod- ern diagnosis for Passiena based on genital and somatic characters of the Oriental type species of which SEM photographs of diag- 398 LEHTINEN— REVIEW OF PASSIENA 399 Figures 1-4. — Diagnostic features of Passiena, SEM photographs. 1-4. Passiena torbjoerni paratypes, Nam Nao National Park, Phetchabun Province, Thailand: 1. Epigynum, female; 2. Tip of male palea with soft pointed spicules; 3-4. Modified setae on venter of male (cf. Hygrolycosa nibrofasciata in Kronestedt 1996: figs. 17, 18). nostic characters are presented. Three of the African species incorrectly placed in Passiena are transferred to Pardosa, whereas the re- maining two are considered incertae sedis pending revisional studies of the main African groups of the Pardosinae. METHODS All specimens of Passiena were examined with an OLYMPUS SZH stereomicroscope. Scanning Electron Micrographs photographs of male and female genitalia and the specialized ventral abdominal setae of the male of P. torb- joerni were taken with a JEOL JSM-5200 and digitized using the software package SemAfore (JEOL Ltd., Tokyo). The digital photographs were taken with an Olympus digital camera and enhanced using the “Helicon Locus” soft- ware. A critical evaluation of the African spe- cies Passiena praepes (Simon 1885) and P. upembensis Roewer 1959 was possible by comparing the descriptions with material of re- lated African species of Pardosinae and Wadi- cosinae available for comparison. Paratypes will be deposited in Stockholm (NHRS), Washington (NMNH) and Paris (MNHN). Terminology. — The terminology of the structures of the copulatory organs is prob- lematic in spiders as presumed homology and similar function and topography of structures have led to deviating nomenclatures. Zyuzin (1993) used functional (embolus, conductor) but also topographical terms (terminal apoph- ysis, tegular apohysis) for the male pedipalp structures. Vogel’s (2004) terminology was based on previous concepts of Dondale & Redner (1990) and seemed to be a mixture of topography, function and homology, with the exception of embolus, palea and median apophysis, all of which are strictly based on homology. Their median apophysis should not be confused with various similarly named pedipalp structures in other families of differ- ent main lineages of spiders. To avoid any confusion, the term tegular apophysis is used here instead of median apophysis. The term terminal apophysis is here used for all sepa- rate sclerites, which topographically corre- spond to the terminal apophysis of the Par- dosinae sensu Dondale & Redner (1990). Additional sclerites between the tegulum and palea are not named here, as Zyuzin’s (1993) term synembolus and several “lamellae” were especially created for Lycosinae, and their ap- plication to Pardosinae might be misleading. Abbreviations. — Collections: MZT, Zoo- logical Museum, University of Turku, Turku, Finland; NHRS, Naturhistoriska Riksmuseet, Stockholm, Sweden; PTL, personal collection of the author. Morphology. AML, ALE, an- terior median and lateral eyes; PME, PLE, posterior median and lateral eyes. SYSTEMATICS Subfamily Pardosinae Simon 1898 Passiena Thorell 1890 Passiena Thorell 1890: 140. Thorell 1892: 186; Si- mon 1898: 355; Roewer 1955: 198; Roewer 1959: 182; Bonnet 1958: 3439 (as synonym of Pardosa). Pardosa C.L. Koch 1847. Bonnet 1958: 3423; Ta- naka 1993: 262. 400 Types species. — Passiena spinicrus Tho- rell 1890 by original designation and mono- typy. Diagnosis. — Passiena is mainly character- ized by a combination of genitalic and somatic characters of males. It can be distinguished from all other lycosid genera by the presence of a group of soft spicules on the distal part of the palea of the male pedipalp (Figs. 2, 18). The ventral side of the abdomen in males (Figs. 3, 4) carries unique modified setae that differ considerably from similar structures in Pardosa sphagnicola (Dahl 1908) and Hygro- lycosa rubrofasciata (Ohlert 1865), although their function could be similar. The base of female epigynum has variable sclerotizations of the lateral plates, although the basic pattern is typically that of Pardosinae. Description. — Small to medium spiders. Color pattern of both carapace and abdomen with a wide, light longitudinal median band (Figs. 5, 7, 11, 13); fovea on carapace very distinct and dark in color. Anterior eye row slightly procurved, AME and ALE subequal in size (Figs. 9, 14); PME row narrower than that of PEE (as in most Pardosinae) (Figs. 5, 13). Femora with oblique or irregular annu- lations in females (Figs. 6, 8, 15), males with different color pattern on femora I (Fig. 12), distal segments of legs uniform in color, sometimes a marmorous pattern on tibiae 1. Tibiae of leg I and II with 4-6 pairs of ventral spines, and metatarsi of legs I-II with usually 4 exceptionally long pairs of ventral spines. Dorsal and lateral spines conspicuous in all legs, relatively shorter in males than females. Abdomen of males ventrally with short mod- ified setae which carry secondary hair-like structures (Figs. 3, 4). Leg spination of male weaker than that in female. Subtegulum small, distally and laterally surrounded by tegular base (Fig. 17); tegular apophysis very short and distinctly separated from tegulum by a deep furrow on all sides; terminal apophysis bipartite, embolus distally curved and partly flattened. Ecology. — The habitat for all samples of Passiena spp. collected by myself is very dark jungle, representing a very unusual habitat for tropical Pardosinae and even for most Lycos- idae except Venoniinae (pers. obs.). Remarks. — Passiena is retained in the Par- dosinae because the male pedipalp morphol- ogy agrees at least partly with the synapo- THE JOURNAL OF ARACHNOLOGY | morphies listed by Dondale (1986) ('conductor shaftlike, lying transversely along the basal margin of the palea’). Two other wolf spider species have modi- j, fied setae on the ventral side of the male ab- domen, H. rubrofasciata and P. sphagnicola. . In H. rubrofasciata, these setae play an im- portant role during the courtship of males, i which is characterized by continuously drum- ming the abdomen on the ground (e.g. Kro- ' nestedt 1984, 1996; Kotiaho et al. 1996; Ko- ^ tiaho 1997). Similar drumming behavior has been reported in P. sphagnicola (Kronestedt ; 1996), but not studied in as much detail. Acoustic communication with other abdomi- nal structures and legs plays an important role in the reproductive biology of spiders (Uetz & Stratton 1982). Unfortunately, no observa- tions on the courtship behavior are available for Passiena spp. Although the function of the modified ventral setae appears to be similar for all above species, the ultrastmcture of these ; modified setae is very different (for H. rubro- [ fasciata see Kronestedt 1996: figs 14-18). The possibility of synapomorphy is exclud- ed for the ventrally spiny male abdomen, and the ultrastructure of these modified setae among normal setae is not even similar (cf. , Figs. 3, 4 with Kronestedt 1996: figs 14-18). Passiena torbjoerni and Pardosa sphagnicola both belong to Pardosinae, but the males of I the closest relatives of the latter (P. pullata I group) have fewer significant modifications in their ventral setae, consisting of uneven length, insignificant thickening and differenc- es in coloration (Holm & Kronestedt 1970; pers. obs.). Hygrolycosa rubrofasciata is not ’ regarded as a member of Pardosinae (Dondale 1986; Zyuzin 1993), although the phylogenet- ic relationships of this genus have not been clarified. The corresponding structures and be- havior of the East Asian Hygrolycosa umidb cola Tanaka 1978 are not known to me while all other species now assigned to Hygrolycosa ■ are known either as female or juveniles only and their generic placement is dubious (cf. Kronestedt 1996). The ultrastructural modifi- cation of the dorsal abdominal setae of female lycosids for attachment of the newly hatched spiderlings was documented by Rovner et al. (1973: figs. 3 a-c). All these results seem to | prove that the abdominal setae of Lycosidae are easily modified for variable adaptations. LEHTINEN— REVIEW OF PASSIENA 401 Figures 5-10. — Digital photographs of female Passiena spinicrus from Malaysia, Pinang; 5. dorsal view of carapace and abdomen; 6. Leg I; 6-10. Female of P. sp. from Sabah, Tawau; 6. Dorsal view of carapace and abdomen; 7. Dorso-lateral view of carapace and abdomen; 8. Lateral view of carapace and leg I; 9. Frontal view of carapace and chelicerae; 10. epigynum, ventral view. 402 THE JOURNAL OF ARACHNOLOGY Figures 1 1-16. — Passiena torbjoerni new species from Nam Nao National Park., Thailand, digital pho- tographs; 11. male dorsally; 12 leg 1 of male; 13. female dorsally; 14 frontal view and chelicerae of female; 15. leg I of female; 16. epigynum. Passiena spinicrus Thorell 1890 Figs. 5-10 Passiena spinicrus Thorell 1890: 140. Thorell 1892: 186; Simon 1898: 355; Roewer 1955: 199; Roewer 1959: 162. Pardosa spinicrus (Thorell). Bonnet 1958: 3423, 3439; Tanaka 1993: 262. Type material examined. — Holotype fe- male from Piilau Pinang, Malaysia [5°25'N, 100°20'E], O. Beccari and E. D' Albertis (NHRS) [erroneously reported lost by Roewer (1959)]. Other material examined. — MALAYSIA: 1 $ with cocoon, 1 juvenile, Pulau Pinang, Batu Ferringgi, 5°28'N, 100°15'E, 29 Novem- ber 1976, PT. Lehtinen, fern thicket (MZT). A female specimen from Malaysia, Sabah, Ta- wau district, Bal Estate, 3°46'N, 100°59'E, 3 November 1979, rubber plantation, PT Leh- tinen (MZT AA7373) with an exactly similar color pattern of the carapace, but slightly de- viating spination of legs may belong to this species, although the large, well-collected gap between these localities may suggest a new taxon for the specimen from Sabah. Diagnosis. — It is not possible to diagnose males of the two Passiena species as males of P. spinicrus are not known. Female Passiena spinicrus are distinctly smaller than P. torb- joerni and the central epigynal septum is more distinct in its posterior part, while the basal integument under the lateral epigynal plates is partly sclerotized, contrasting to the complete- ly soft integument in P. torbjoerni. Description. — Female (Pulau Pinang, Ma- laysia): Medium to small-sized pardosine species. Color pattern of both carapace and abdomen with wide light longitudinal band (Fig. 5) and narrow light submarginal bands, LEHTINEN— REVIEW OF PASSIENA 403 fovea on carapace very distinct and also rec- ognizabie for its dark color; a pair of dark elongate spots within the narrower anterior part of the central band similar to the color pattern of this area in the specimen from Ta~ wau (Figs. 5,7): leg femora with irregular an= nulations (Fig. 6), the corresponding annuli in the Tawau specimen very distinct (Fig. 8); all more distal segments of legs of uniform color or sometimes with an obscure marmorous pat- tern of the front tibiae. Anterior eye row weakly procurved, AME and ALE subequal in size. Posterior eye row an anteriorly strong- ly narrowed trapezium (as in most Pardosi- nae). Front tibiae (I-II) with 6, metatarsi I-II with 4 exceptionally long pairs of ventral spines. Dorsal and lateral spines conspicuous in all legs. Female (Tawau, Sabah): Body total length 3.4 mm. Carapace 1.9 mm long, 1.0 mm wide. Leg I: femur 1.4, patella 0.60; tibia 1,25; metatarsus 1.25; tarsus 0.7 mm. Anterior eye row distinctly procurved, AME larger than ALE. Posterior eyes form an anteriorly strong- ly narrowed trapezium; PLE larger than PME. Labium much wider than long. Carapace brown, long median light stripe with a short triangle extending between the pO'Sterior eyes and a long light, gradually tapering triangle extending close to the posterior margin. Nar- row light submarginal stripes are present. Chelicerae mesally brown, laterally light, ster- num and coxae dirty white, gnathocoxae and labium uniform light brown. Remaining fem- ora (I and IV) with three dark very oblique U-shaped aenulations, patellae and tibiae brown with lighter marmorous pattern, meta- tarsi and tarsi (I and IV) uniformly pale yel- lowish brown. Lateral faces of abdomen rather dark brown with regularly placed minute pale spots, cen- tral light stripe wide with dark segmentally arranged lateral dentations, the anterior folium yellowish brown, slightly sclerotized. Ventral face of abdomen with wide central light band, its margins with numerous dark dentations. Two pairs of very distinct circular muscular apodemes behind the epigastric fold. Epigynal area distinctly darker than its surroundings. Spination of leg I: 5 ventral pairs of very long spines on tibia and 3 ventral pairs of very long spines on metatarsi, (both legs II and III missing) 2 shorter retrolateral spines are pre- sent both on tibiae and on metatarsi 1. Nu- merous short erect setae on ventral side of fe- mur I among 4 stronger and longer setae. Epigynum: shape of epigynal plates as in Fig. 10 (Tawau specimen: the epigynal mount of the MZT specimen from Pinang, compared with the topotypical holotype has not been found during this study). Epigynal septum an- teriorly rounded in ventral view, weakly scler- otized pair of inner arches (corresponding to margins of anterior pockets of most lycosid species). Vulva with two pairs of rounded re- ceptacula, connected to each other with a short constriction only. Epigynal median fur- row anteriorly rounded and slightly widened; posterior transverse bar well developed. Lat- eral epigynal plates with a distinct notch bor- dering the posterior bar. Male.- unknown. Egg-cocoon (Pulau Pinang, Malaysia).- light brown and with a distinct surface fold or seam as in all Pardosa s. lat. Distribution. — Passiena spinicrus is found on Pinang Island, Malaysia, and possibly in northeastern Borneo. The present study con- firms the presence of a population at the type locality Pulau Pinang still in 1976. I have also seen juvenile specimens from the Malayan Peninsula that may be referable to this spe- cies. The identification of juvenile specimens of Passiena at the generic level is possible in the field due to the unusual color pattern and leg spination, and the unusual habitat prefer- ence. In contrast to most other tropical wolf spider species, P. spinicrus occurs in excep- tionally dark habitats, in the ground layer of dense jungle or bush. Roewer (1955) and subsequently Platnick (2005) listed a much wider range (Tndia to Hong Kong, Sumatra, Sulawesi') which seems to be erroneous since the original description of the holotype female from Malaysia, Pulau Pinang, appears to be the only previous pub- lished record. In addition, I have not found P, spinicrus during extensive fieldwork in Su- matra, India and neighboring countries, or in Sulawesi and Hong Kong, although I have specifically searched for rainforest dwelling pardosines for many years. Passiena torbjoerni new species Figs. 1-4, 11-19 Material examined. — Holotype male, Nam Nao National Park, Phetchabun Province, Thailand, 16°43'N, 10r33'E, 19 November 404 THE JOURNAL OF ARACHNOLOGY Figures 17-19. — Passiena torbjoerni new species from Nam Nao National Park, Thailand, SEM-mi= crographs of male palp: 17. bulbus ventrally, arrow = subtegulum; 18 distal view of palp, arrow = spicules on palea; 19. cymbium, bulb removed, ventral view. 1976, RT Lehtinen, ground layer of savanna (NHRS). Paratypes: THAILAND: 1 c?, col- lected with holotype (NHRS); 3 S, 4 $, 2 juveniles, Nam Nao National Park, Phetcha- bun Province, 16°43'N, lOLSS'E, 19 Novem- ber 1976, RT Lehtinen, bamboo thicket (1 d, 1 9 PTL; remainder in NHRS); 1 c3, 1 juve- nile, same locality, 29 October- 19 November 1976, pitfall trap, J, Ruohomaki, E. Huitula, RT. Lehtinen, grassy margin of bamboo forest (NHRS); 1 subadult <3, same locality, 29 Oc- tober-19 November 1976, pitfall trap, J. Ruo- homaki, E. Huitula, RT. Lehtinen, grassy for- est (RTL); 1 juvenile, Doi Suthep, Kontathon waterfall, Chiangmai Province, 18°49'N, 98°54'E, 14 November 1976, RT. Lehtinen, in jungle (RTL). Etymology. — This species is dedicated to the Swedish lycosid specialist and my good friend Dr. Torbjorn Kronestedt, one of the few specialists who has seen a real Passiena, Diagnosis. — It is not possible to diagnose males of the two Passiena species as males of P, spinicrus are not known. Females of Pas- siena torbjoerni are distinctly larger than those of P, spinicrus and the central epigynal septum is less distinct posteriorly, while the basal integument under the lateral epigynal plates is desclerotized, contrasting to partly sclerotized integument in P. spinicrus. Description. — Male (from Nam Nao Na- tional Park, Thailand):.” Total body length 4,2 mm (including lengthened/outdrawn petiolar tube). Carapace 1.65 mm long, 1.58 mm wide. LEHTINEN— REVIEW OE PASSIENA 405 Abdomen 2.1 mm long. Leg I: femur 1.96; patella 0.56; tibia 1.47; metatarsus 1.61; tarsus O. 81 mm. Carapace with laterally steep ce- phalic region (Fig. 14); clypeus ca. three times the diameter of AME; dorsally with a cen- trally widened, very pale-brownish longitudi- nal median band, while the lateral parts are dark brown without further signs of pattern (Fig. 11) with a wide pale longitudinal band; median band continuing into the cephalic area after a dark constriction caused by the very dark brown surroundings of the posterior eyes, lateral parts uniform dark brown. Some spec- imens have faint narrow submarginal lighter stripes; posterior eyes on blackish wide tuber- cles; sternum, labium and gnathocoxae uni- form light brown. Fovea conspicuous, dark. Chelicerae with longitudinal dark stripes as in P, spinicrus. Abdomen (Fig. 11) with a reddish brown central longitudinal band, its margins uneven- ly serrate. Lateral parts brown, throughout with minute lighter spots. Ventral face of ab- domen pale brown, its central area throughout covered with short spines (Figs. 3, 4). Femora and tibiae of all legs with numerous dorsal and lateral spines, some dorsal spines on femora exceptionally long. Tibiae and metatarsi I-II with 3-4 pairs of ventral spines (Fig. 12), but these spines are much shorter than the corresponding spines on females of both known species of Passiena. Basal two thirds of femora I dark brown in dark speci- mens (Fig. 12), with oblique dark stripes on lighter specimens, other segments uniform pale brown. Male pedipalp (Figs. 17-19) dark brown, with conspicuous field of soft spicules in the distal part of palea (Figs. 2, 18). Cymbium distally screwed (Fig. 19); tegular apophysis short, terminal apophysis bilobate, embolus distally curved and flattened. The group of spiculae in the distal part of palea deviates from all modifications of palea in other Par- dosine groups, where all paleal modifications are well sclerotized. Female (paratypes from Nam Nao National Park).- Total length 4.2 mm. Carapace 2.21 mm long, 1,65 mm wide. Abdomen 2.2 mm long. Leg I: femur 1.75; patella 0.56; tibia 1.82; metatarsus 1.65; tarsus 0.74 mm. Cara- pace dorsally with a centrally widened, pale- brownish longitudinal median band bordered by a pair of dark brown longitudinal bands. The lateral parts have regularly additional pale submarginal bands as pale-colored males (Fig. 13); of PME and PLE; chelicerae pale with very distinct dark central stripes on the ante- rior face as in P. spinicrus. Row of anterior eyes slightly procurved, AME larger than ALE (Fig. 9). Abdomen dorsally with a very wide pale-brown central field, laterally with narrow stripes forming a dark brown reticu- lation; anterior margin encircled by a row of dark setae; venter and sternum uniform pale brown. Basal half of femora I dark brown; all other femora may have an indistinct dark mar- morous pattern; all other leg segments more or less uniform pale brown. Spination of legs: All femora with two strong dorsal spines and 2-3 short lateral spines on both sides; tibiae I-II with six long, strong pairs of ventral spines, one long retrolateral spine, and one subdistal dorsal spine; metatarsi I-II with 3 pairs of very long ventral spines in the basal half and few shorter ventral and lateral spines; patellae, tibiae and metatarsi III-IV irregularly covered with comparatively weak and short spines, very different from the very long and strong ventral spines of legs I-II. Epigynum: narrow, soft median septum be- tween lateral plates continued as a soft basal transverse bar; median septum inverted-T shaped; posterior basal transverse bar with slightly curved lateral arms. Remarks. — Passiena torbjoerni is found in Phetchabun and Chiangmai Provinces of Thai- land, where it inhabits the floor of rainforests. DISCUSSION The African Passiena, — All current Afri- can representatives of Passiena except P. au- berti have a dorsal abdominal pattern with an anterior folium, followed by an unpaired dark central area. In addition, their genitalia do not resemble in any way P. spinicrus or P. torb- joerni suggesting that they are misplaced in Passiena. Critical evaluation of the genital and somatic characters of three of these spe- cies allowed a tentative placement in other ly- cosid genera pending a generic revision of Af- rican Lycosidae. Passiena praepes (Simon 1885). — This species is only known from the female type specimen collected in Senegal (Simon 1885) and was originally described in the genus Par- dosa. Its transfer to Passiena by Roewer (1959) was primarily supported by the pres- 406 THE JOURNAL OF ARACHNOLOGY ence of four pairs of ventral spines on tibia L The drawing of the epigynum (Roewer 1959: 170, fig. 86a) strongly resembles that of P. micheli Simon 1901 (Roewer 1959:67, fig. 23a) and P. potteri Simon 1901 (Roewer 1959:70, fig. 27a) both of which are regarded as junior synonyms of P. naevia (C.L. Koch 1875), a typical representative of the Pardosa nebulosa-group (Alderweireldt & Jocque 1992). The abdominal pattern of P. praepes as illustrated by Roewer (1959) confirms its affinities with the P. nebulosa group. There- fore, P. praepes is here regarded as a repre- sentative of the Pardosa nebulosa group, and consequently returned to the genus Pardosa. Passiena auberti (Simon 1898). — This spe- cies from South Africa was originally de- scribed in Pardosa (Simon 1898). Due to a distinct and wide longitudinal pale band both on carapace and abdomen, combined with a strongly procurved anterior eye row it does not fit into any currently described group (ge- nus or species group) of the Pardosinae. Pend- ing a generic revision of African Lycosidae, I regard P. auberti as incerta sedis. Passiena albipalpis Roewer 1959. — This species from Cameroon has six pairs of ven- tral spines on tibiae I and II, an unusual type of tegular apophysis and a strongly sclerotized and widely arched palea on the male pedipalp. I am unable to place this species into any known genus within the Pardosinae. However, the somatic characters and male and female genitalia are very different to the two known species of Passiena, and therefore I regard this West-African species as incerta sedis. Passiena elegantula Roewer 1959. — This species from the Democratic Republic of Con- go is known from both sexes and Roewer’s (1959:236, fig. 118) illustrations including the male pedipalp with its enlarged palea region suggest a placement in Pardosa. In addition, it does not display the typical carapace and abdomen coloration of Passiena with the typ- ical wide light bands. Consequently, P. elegan- tula is here transferred to Pardosa: P. elegan- tula (Roewer 1959) NEW COMBINATION. Passiena upembensis Roewer 1959. — This species is known from a female collected in the Democratic Republic of Congo. It is cer- tainly related to Pardosa oncka Lawrence 1927 and Pardosa crassipalpis Purcell 1903 and may be a junior synonym of the latter. Similarly, Kronestedt’s (1987) revision showed that the widespread P. oncka was il- lustrated under six differently named species by Roewer (1959). However, I have not com- pared the type of P. upembensis with material of P. crassipalpis from Botswana available to me. ICronestedt (1987) suggested a potential placement of P. oncka in Wadicosa Zyuzin 1985. Here, I transfer P. upembensis to Par- dosa, P. upembensis (Roewer 1959) NEW COMBINATION, based on its similarity with P. oncka and P. crassipalpis pending a revi- sion of the African Pardosinae and Wadicos- inae. ACKNOWLEDGMENTS Dr. Torbjorn Kronestedt (NHRS) provided the opportunity to study the type material of P. spinicrus during my visits to Stockholm, and supplied loans of other Oriental pardosine species. Reijo Hakanen and Eirik Granqvist, both at the time at the University of Helsinki, put in substantial efforts to collect spiders in Botswana in 1973, resulting in rich material of African Pardosinae for comparative studies at our museum. Dr. Yuri Marusik (Magadan, Russia) assisted in the digital photography and numerous computer problems. Mr Veikko Rinne (University of Turku) also assisted with computing issues. Dr. Volker Eramenau (Western Australian Museum, Perth) succeed- ed to direct my arachnological activities again to Lycosidae and he and an unnamed referee also gave valuable suggestions and editorial advice on the manuscript. This article is a product of his invitation to the Lycosidae Symposium held during the International Congress in Gent in 2004. LITERATURE CITED Alderweireldt, M. & R. Jocque. 1992. A review of the nebulosa-gvou^ of Pardosa C.L. Koch, 1847 in Africa, a complex with some highly variable species (Araneae Lycosidae). Tropical Zoology 5:73-113. Bonnet, P. 1958. Bibliographia araneorum 2(4). Douladoure, Toulouse: 3027-4230. Dondale, C.D. & J.H. Redner 1990. The Insects and Arachnids of Canada, Part 17. The Wolf Spiders, Nurseryweb Spiders, and Lynx Spiders of Can- ada and Alaska, Araneae: Lycosidae, Pisauridae, and Oxyopidae. Research Branch, Agriculture Canada, Publication 1856. Kotiaho, J. 1997. Sexual selection and costs of sex- ual signaling in a wolf spider. Biological research reports from the University of Jyvaskyla 55:1- 25. LEHTINEN— REVIEW OF PASSIENA 407 Kotiaho, J., R.V Alatalo, J. Mappes & S. Parri. 1996. Sexual selection in a wolf spider: male drumming activity, body size and viability. Evo- lution 50:1977-1981. Kronestedt, T. 1984. Ljudalstring hos vargspindeln Hygrolycosa rubrofasciata. Fauna och Flora 79: 97-107. Kronestedt, T. 1987. On some African and Oriental wolf spiders (Araeeae, Lycosidae): redescription of Pardosa oncka Lawrence from Africa, with notes on its generic position. Journal of natural History 21:967-976. Kronestedt, T, 1996. Vibratory communication in the wolf spider Hygrolycosa rubrofasciata (Ar- aneae, Lycosidae). Revue Suisse de Zoologie, volume hors serie:341-354. Platnick, N.I. 2005. The World Spider Catalog, Ver- sion 5.5. American Museum of Natural History, online at http://research.amnh.org/entomology/ spiders/catalog/index .html . Roewer, C.F. 1955 [imprint date 1954]. Katalog der Araneen von 1758 bis 1940, bzw. 1954. VoL 2a. Institut Royal des Sciences Naturelles de Bel- gique, Bruxelles. Roewer, C.F. 1959 [imprint date 1958], Araneae Ly- cosaeformia II (Lycosidae). Exploration du Parc National de lUpemba — Mission G. F. de Witte 55:1-518. Roewer, C.F. 1960 [imprint date 1959]. Araneae Ly- cosaeformia II (Lycosidae) (Fortsetzung und Schluss). Exploration du Parc National de rupemba— Mission G. F. de Witte 55:519-1040. Rovner, J.S., G.A. Higashi, & R.F. Foelix. 1973. Maternal behavior in wolf spiders: the role of abdominal hairs. Science 182:1153-1155. Simon, E. 1898. Histoire Naturelle des Araignees. VoL 2:193-380. Roret, Paris. Tanaka, H. 1993. Lycosid spiders of Japan IX. The genus Pardosa C. L, Koch — amentata-group. Sonoda Women’s College Studies 27:261-318. Thorell, T. 1890. Diagnoses aranearum aliquot no- varum in Indo-Malesia inventarum. Annali del Museo Civico di Storia Naturale di Genova 30: 132-172. Thorell, T. 1892. Studi sui ragni Malesi e Papuani. IV,2. Annali del Museo civico di storia naturale di Genova 31:1-490. Toegiorgi, P. 1966. Italian wolf spiders of the genus Pardosa (Araneae: Lycosidae). Bulletin of the Museum of Comparative Zoology, Harvard Uni- versity, 134:275-334. Uetz, G.W. & G.E. Stratton. 1982. Acoustic com- munication and reproductive isolation in spiders. Pp. 123-159. In Spider communication. Mecha- nisms and Ecological Significance (RN. Witt & J.S. Rovner, eds). Princeton University Press, Princeton, New Jersey Vogel, B. 2004. A review of the spider genera Par- dosa and Acantholycosa (Araneae, Lycosidae) of the 48 contiguous United States. Journal of Ar- achnology 32:55-108. Zyuzin, A. A. 1993. Studies on the wolf spiders (Ar- aneae: Lycosidae). 1. A new genus and species from Kazakhstan, with comments on the Lycos- inae. Memoirs of the Queensland Museum 33: 693-700. Manuscript received 12 January 2005, revised 5 August 2005. 2005. The Journal of Arachnology 33:408-414 THE FUNCTION OF LONG COPULATION IN THE WOLF SPIDER PARDOSA AGRESTIS (ARANEAE, LYCOSIDAE) INVESTIGATED IN A CONTROLLED COPULATION DURATION EXPERIMENT Andras Sziranyi, Balazs Kiss, Ferenc Samu: Plant Protection Institute of the Hungarian Academy of Sciences, RO. Box 102. Budapest, H-1525 Hungary. E-mail: samu @ j ulia-nki.hu Wolfgang Harand: Bundesamt und Forschungszentrum fiir Landwirtschaft, Vienna, Austria. ABSTRACT. Copulation duration varies greatly in wolf spider species, ranging from a few seconds to several hours. In Pardosa agrestis (Araneae, Lycosidae), the most common ground dwelling spider in Central European fields, copulation typically takes more than two hours. Since long copulation is likely to entail certain costs, we address the question, “what is the function of long copulations?” We investigated the consequences of lengthy copulation in an experimental situation, where copulations either ended spon- taneously, or were interrupted after 1 0 min, 40 min or 90 min. There was no difference in the number of offspring per female when treatments were compared and we conclude that ten minutes of copulation was sufficient to fertilize all the eggs of a female. Long copulations should therefore have other functions than securing the necessary amount of sperm for fertilization. We also found that neither the time until egg production after copulation, nor offspring size was affected by copulation duration. This suggests the lack of transfer of ejaculatory substances that would either stimulate the egg sac formation or increase the size of the spiderlings. We propose that prolonged copulations gain meaning in multiple mating situations and should play a role in sperm competition or other forms of sexual selection. The extra time may be used for copulatory courtship, or for the transfer of surplus sperm or other substances to manipulate the female’s willingness to copulate with other males, or to use sperm from them. These hypotheses remain to be tested in multiple mating experiments. Keywords: Sperm transfer, copulation duration, copulation pattern, sexual selection, wolf spider Copulation time varies largely in spiders even between closely related species. In the family of wolf spiders many species copulate for just a few minutes, while others copulate for hours. In a survey 30 species of wolf spi- ders, Stratton et al. (1996) noted that Arctosa littoralis (Hentz 1844) copulated for the short- est time, 18 seconds, while Schizocosa salta- trix (Hentz 1844) represented the opposite ex- treme with more than 8 hours. Copulation duration may vary widely even among species within the same genus, e.g. a 20 fold differ- ence was found between the shortest and lon- gest copulating Hogna spp. (Stratton et al. 1996). Our own observations on Old World lycosid species showed similar variable pat- terns. In a pilot study, Pardosa hortensis (Thorell 1872) copulated for 20-30 minutes, while Pardosa agrestis (Westring 1861) took on average six times longer to copulate (A. Sziranyi unpubl. data). Given the examples for both short and long copulations, the question arises: what is the function of long copulation duration in wolf spiders, and in particular in the wolf spider, Pardosa agrestisl We chose to study P. agres- tis, because it typically exhibits long copula- tion, and as the most abundant agrobiont spi- der in Central European arable fields (Kiss & Samu 2000; Samu & Szinetar 2002), it is the primary model species of our research group. The species builds no retreat and it hunts ac- tively on the ground during the day (Samu et al. 2003). Pardosa agrestis has two reproduc- tive peaks, with mating periods mostly in May and in July-early August (Samu et al. 1998). So far, we could not establish whether females remate in nature, but they do so in the labo- 408 SZIRANYI ET AL.— LONG COPULATION IN PARDOS A AGRESTIS 409 ratory (Kiss 2003). Egg sacs are produced 2- 3 weeks after copulation, and they are carried by the female for an average of 3 weeks until hatching. The main function of copulation is sperm transfer, which may take a long or a short time. Long copulation duration might be nec- essary for complete fertilization, if sperm transfer is slow. Slow sperm transfer could be a common phylogenetic constraint, however in a taxonomic group in which copulation du- ration varies widely, this can be ruled out. On the other hand, long copulation has to be evo- lutionarily maintained, because it is likely to be costly. The time spent copulating is ener- getically demanding (Watson & Lighten 1994), and it entails a loss of opportunity to copulate with other partners or to forage. In species in which copulation takes place with- out hiding in a refuge, like in P. agrestis, an elevated predation risk can also be expected (Krupa & Sih 1998). Long copulation might also expose the spiders to parasite infection (Scheffer 1992). Thus, lacking a phylogenetic explanation, and considering possible costs, we should look for the adaptive value of pro- longed mating. To find the possible adaptive value of pro- longed copulation in P. agrestis we created a hypothesis framework, and tested it in single mating experiments where the copulations were interrupted to ensure predefined dura- tion, Three different scenarios of time-use were constructed (Fig. 1). From each scenario, specific predictions can be formulated and tested. We consider hypothesis A to be the null- scenario, in which sperm is transferred throughout the entire copulation time and it is all used for fertilization. In this scenario sperm transfer is insufficient if copulation time is limited which prevents the fertilization of all ova. Indeed, copulation duration is pos- itively associated with sperm transfer in a number of arthropods (Dickinson 1986; Arnqvist & Danielsson 1999; Stalhandske 2001), but in other cases, like in the spider Micrathena gracilis (Walckenaer 1805), this relationship does not exist (Bukowski & Christenson 1997). A prediction from Hypoth- esis A is that shorter than natural copulation time would result in a reduced number of fer- tilized eggs and offspring. In the following hypotheses (B and C) the sperm transfer rate Hypothesis: Copulation pattern; Figure 1 . — Hypotheses concerning copulation pattern. There are two considered components to the pattern: 1. (shaded areas) periods when sperm used for fertilization is transferred; 2. (empty areas) periods when either no sperm is transferred or no such sperm that would be used for fertilization ( = non-fertilizing period). Numbers indicate some pos- sible functions of the no-transfer/non-fertilizing pe- riod: 1. removing earlier male sperm and/or plug; 2. assessing female virginity; 3. in-copula court- ship; 4. transferring extra sperm; 5. transferring ma- terials to accelerate oogenesis; 6. transferring nutri- tive materials (nuptial gift). See details in text; for further possible functions see Eberhard (1996). does not limit fertilization, but rather the pro- longed copulation duration is maintained by other factors. With hypothesis B we propose the scenario that the sperm transfer is timed for the end of the copulation. In that case, the first part of the copulation then serves other purposes, such as copulatory courtship. Just as courtship influences female choice, copulatory courtship can influence female choice regarding postco- pulation events. In spiders, sperm transfer and fertilization are well separated in time, and during the period between transfer and fertil- ization females may manipulate which male’s sperm is used for fertilization via “cryptic fe- male choice” (Watson & Lighton 1994; Eber- hard 1997; Schafer & Uhl 2002). In Linyphi- idae the first part of copulation, the phase without sperm transfer, is often referred to as pseudo-copulation (Helsdingen 1965). Males are able to distinguish virgin from non-virgin females during courtship and pseudo-copula- tion (Robinson 1982; Suter 1990). This phase of mating can also be used to remove the sperm (Schafer & Uhl 2002) placed by a pre- 410 THE JOURNAL OF ARACHNOLOGY JZ U 01 a. cn c 13 1^ Q. m o Q 100 80 60 40 20 0 lOmin 40iTiin OOrnin Control Copulation treatment groups Figure 2. — The number of spiderlings in a clutch as a function of copulation length. vious mate in the female's genital tract. From Hypothesis B we can predict that copulation inten'upted early will result in no sperm trans- fer, and consequently in no hatched offspring from the egg sacs later. In Hypothesis C the volume of sperm nec- essary to fertilize the eggs is transferred at the beginning of copulation and the rest of the time is used for other activities. The remain- ing copulatory time can, similar to Hypothesis B, serve the purpose of copulatory courtship, or it may simply engage the female long enough, so that competing mates will have re- duced chances of copulation (mate guarding). Another possibility is that these activities de- crease female receptiveness to the sexual ad- vances of other males (Eberhard 1996). Hav- ing transferred the necessary amount of sperm to fertilize all eggs, sperm transfer might con- tinue. Surplus sperm might be advantageous if females mate multiple times, because then a greater volume of sperm can be used in nu- merical sperm competition (Elgar 1998). In some species after sperm transmission, males create mating plugs in the genital opening of the female to prevent further copulations (Masumoto 1993; Knoflach 1998). Males may transfer substances to the female that facilitate rapid oviposition, thus leaving less time for the female to meet and mate with competing males prior to egg deposition (Yamaoka & Hirao 1977). Nutritive substances might also be transferred to the female genital tracts (Su- ter & Parkhill 1990), which increase offspring size, thus increasing parental fitness. Hypoth- esis C predicts that in a single copulation, the female's reproductive output should remain unchanged after the first part of copulation. On the other hand, the various possible func- tions of the remaining copulation time gen- erate additional predictions for offspring size and period until oviposition. We interrupted copulations after three dif- ferent time intervals to distinguish between the above hypotheses. Here we report the re- lationship between the artificially set copula- tion duration and reproductive output. These experiments show which of the originally pro- posed hypotheses are rejected or supported and cast some light on the function of various phases of copulation. METHODS The experiment was conducted from April to September, 1998 at the Plant Protection In- SZIRANYI ET AL.^LONG COPULATION IN PARDOSA AGRESTIS 411 stitute of the Hungarian Academy of Sciences (near Budapest). Pardosa individuals were hand-collected in juvenile or subadult stages in April to ensure virgin adults for the exper- iment. Animals were kept separately in the laboratory^ where they were reared to adult- hood under standard conditions (25 °C, long daylight (L:D — 16:8), and Drosophila me- lanogaster ad libitum was provided as food). Basic rearing conditions are explained in Kiss & Samu (2002). Interrupted copulation experiment.- — Following maturation, adult males and fe- males were divided randomly into four groups. Pairs from each group were put to- gether in 17 cm -diameter Petri dishes, and the occurrence o-f copulation was monitored. In the first three treatment groups copulation was interrupted after 10 min {n = 16), 40 min {n = 22), and 90 min {n = 16) respectively. The pairs in the fourth group (Control) were left undisturbed until they finished copulation {n = 20). To establish the distribution of unin- terrupted copulation duration, additional ob- servations on copulation length were per- formed (w = 42), in which the reproductive consequences were not observed. After copulation, females were kept in the laboratory under the conditions presented above. We recorded the time between copu- lation and ovipositioe (in some cases, exact time of oviposition could not be recorded, which resulted in smaller sample sizes for that variable. — 9, — 7, ~ *7? ^con- trol “ 8); whether females abandoned or con- sumed their egg sac, and whether the hatching was successful. We calculated the ratio of abandoned egg sacs as the number of egg sacs abandoned in an interraption treatment / all egg sacs produced in the given treatment. We monitored mothers with egg sacs for hatching daily. If hatching was successful we counted the number of offspring (thus offspring num- ber counts are based only on successful hatch- es), and placed a sample of 10 spiderliegs into 70% ethanol. Later we estimated the prosomal area (length„„ X widthj„ax) on this random sample of ten spiderlings of each brood using scaled digital pictures. For the size measure- ment, we chose to measure the prosoma be- cause it is less prone to the current feeding status (e.g. cannibalizing a littermate) of the spiderling. Voucher specimens were deposited in the public collection of the Plant Protection Institute, Hungarian Academy of Sciences. RESULTS The duration of spontaneously ended (un- interrupted) copulation events was over 2.5 hours (mean “ 165.7 min; S.D. = 53.6; range ^ 90-319 min). Interruption of the copulation did not affect the number of offspring from hatched egg sacs (Fig. 2; ANOVA: E 3^ 40 = 0.46, P ^ 0.71; homogeneity of variances as- sumption tested by LeveeeN Test: F 3^ 40 = 2.13, P = 0.11). The prosomal area of the offspring did not differ significantly between copulation deration treatments with the effect of mothers nested within treatment (ANOVA, main treatment effect: F 3^ 47^ = 0.35, P = 0.79). However, the effect of mothers on off- spring size was highly significant (ANOVA, effect of mothers nested within treatment: F 38,478 6.11, P < O7OOOI). The time between copulation and egg sac production was not significantly different among the treatments (overall mean 20.6 days, S.D. = 3.88, AN- OVA: F 3^27 ™ 0.10, F = 0.96). Egg sac aban- donment, on the 'Other hand, occurred uneven- ly among the treatments (test of homogeneity: = 8.62; d.f. = 73, F = 0.035). As Fig. 3 illustrates, abandonment ratio was particularly high (0.56) in the 10 rnie interruption group, significantly higher than in the other treat- ments (10 minutes vs. other treatments lumped, Fisher’s Exact Test: n =■ 74, F = 0.01). Among the longer than 10 minutes cop- ulation treatments and the control group, the abandonment ratio was equally low (on av- erage 0.2, test of homogeneity: = LOl; d.f. - 57, F - 0.6). DISCUSSION In the present study, we wanted to establish the pattern of sperm transfer during the long copulation of F. agrestis. Our first hypothesis, Hypothesis A (Fig. 1), was that copulation takes longer because sperm transfer rate is slow. This hypothesis can be rejected, because the interruption experiment demonstrated that sperm transferred even in the first 10 minutes of copulation can be sufficient to fertilize all eggs of a female; offspring numbers were in- dependent of copulation treatments. These re- sults also cause Hypothesis B to be rejected, because this hypothesis predicted zero repro- ductive output for short copulation treatments. 412 THE JOURNAL OF ARACHNOLOGY Copulation treatment groups Figure 3. — The ratio of abandoned egg sacs in the copulation duration treatments. The interruption results are consistent only with Hypothesis C, which proposes a copu- lation pattern in which the transfer of sperm needed to fertilize the eggs takes place within a short period at the beginning of the copu- lation. However, egg sac abandonment was signif- icantly more frequent in the 10 min copulation treatment than in any of the other treatments. Since egg sacs are abandoned when they are sterile (Kiss 2003), this suggests that there is variability in the first 10 minutes of mating. That is, in some cases, during the first 10 min- utes enough sperm was transferred to fertilize all eggs of the future egg sac, while in other cases, no sperm was transferred and the egg sacs were sterile. We can speculate that this can happen if the event of sperm transfer is fast, even compared to the 10 min time scale. Sperm transfer seemed to occur with c. 50% probability during the first 10 min, and with near certainty during the first 40 min of the copulation. Since we had no direct observa- tion on the ratio of sterile and fertilized eggs in the abandoned egg sacs, we can only infer from previous observations that they were likely to be sterile. We note a 20% baseline abandonment which occurred in all longer in- terruption treatments, including the control, and which could be either a natural phenom- enon, or an artifact caused by the experimen- tal situation. A similar phenomenon has been described in the salticid Phidippus johnsoni (Peckham & Peckham 1883), in which copu- lation duration also had an all-or-none effect on fertility, with short copulations more fre- quently resulting in the ‘none’ outcome, and long copulations more frequently resulting in a fully fertilized clutch (an ‘all’ outcome), whereas no half-sized clutches occurring (Jackson 1980). Jackson (1980) did not pro- vide an alternative explanation to the quick release of a large amount of sperm, with an increasing probability over time. Thus, the basic copulation pattern of P. agrestis corresponds to Hypothesis C: rapid transfer of an amount of sperm that is enough to fertilize all eggs takes place at some point during the first part of the copulation, while the final phase, which is the longer portion of time spent in copula, does not serve the direct purpose of transfemng sperm to fertilize the eggs. Given this pattern, the question remains concerning the function of the final phase of copulation. In Fig. 1 we list a number of such possible functions. Of these, some can be ex- SZIRANYI ET AL.— LONG COPULATION IN PARDOSA AGRESTIS 413 eluded based on the present experiment. Since spiderlings were of equal size irrespective of copulation duration, there is no evidence that the male transferred any nutrients during the final phase of the copulation in order to in- crease parental fitness through offspring size (Suter & Parkhill 1990; Walker et al. 2003). The transfer of materials that could accelerate oogenesis can also be ruled out on the basis of similar egg sac formation times across the copulation duration treatments. We did not find any evidence for the presence of a mating plug. Although, we could rule out a number of proposed functions for the final phase of the copulation, several other functions still re- main possible. If surplus sperm are trans- ferred, the sperm could be used to out-com- pete the sperm of other possible mates. The final phase of copulation may also serve as copulatory courtship or as mate guarding. These functions are not mutually exclusive, and any combination of them is possible. To summarize, the copulation pattern estab- lished for P. agrestis seems to be paradoxical in a single mating situation, because much shorter copulations are sufficient to result in full fertilization. Any possible function of a long copulation that was not experimentally excluded here seems to be related to male- male competition and/or female choice, and gains meaning only in a multiple mating sit- uation. Therefore, we tentatively conclude that prolonged copulation is a sexually selected trait in P. agrestis. To establish the functional details and exact adaptive advantages, multi- ple mating studies are needed. ACKNOWLEDGMENTS We want to express our thanks to Mrs. Er- ika Botos for technical assistance. OTKA No. T48434 and F030264 provided financial sup- port, the Plant Protection Institute of the Hun- garian Academy of Sciences made research locations and tools available. We are grateful to Drs. Ann Rypstra, Shawn Wilder, Spren Toft, Paula Cushing, Dan Mott and two anon- ymous referees for revising the text and giving us useful advice. E Samu and B. Kiss were Bolyai fellows of the Hungarian Academy of Sciences. LITERATURE CITED Arnqvist, G. & 1. Danielsson. 1999. Postmating sexual selection: the effects of male body size and recovery period on paternity and egg pro- duction rate in a water strider. Behavioral Ecol- ogy 10:358-365. Bukowski, TC. & TE. Christenson. 1997. Deter- minants of sperm release and storage in a spiny orbweaving spider. Animal Behaviour 53:381- 395. Dickinson, J.L. 1986. Prolonged mating in the milkweed leaf beetle Lcibidomera clivicollis (Co- leoptera: Chrysomelidae): a test of “sperm-load- ing” hypothesis. Behavioral Ecology & Socio- biology 18:331-338. Eberhard, W.G. 1996. Eemale Control: Sexual Se- lection by Cryptic Female Choice. Princeton University Press, Princeton. Eberhard, W.G. 1997. Sexual selection by cryptic female choice in insects and arachnids. Pp. 32- 57. In The Evolution of Mating Systems in In- sects and Arachnids. (J. C. Choe & B. J. Crespi, eds.). Cambridge University Press, Cambridge. Elgar, M.A. 1998. Sperm competition and sexual selection in spiders and other arachnids. Pp. 307- 339. In Sperm Competition and Sexual Selec- tion. (T. R. Birkhead & A. P. Moller, eds.). Ac- ademic Press, New York. Jackson, R.R. 1980. The mating strategy of Phidip- pus johnsoni (Araneae, Salticidae): IT Sperm competition and the function of copulation. Jour- nal of Arachnology 8:217-240. Kiss, B. & E Samu. 2000. Evaluation of population densities of the common wolf spider Pardosa agrestis (Araneae: Lycosidae) in Hungarian al- falfa fields using mark-recapture. European Jour- nal of Entomology 97:191-195. Kiss, B. & E Samu. 2002. Comparison of autumn and winter development of two wolf spider spe- cies (Pardosa, Lycosidae, Araneae) having dif- ferent life history patterns. The Journal of Ar- achnology 30:409-415. Kiss, B. 2003. The autecology of Pardosa agrestis (Westring, 1864). PhD dissertation, Veszprem University (Hungary) Pp. 1-115. (In Hungarian) Knoflach, B. 1998. Mating in Theridion varians Hahn and related species (Araneae: Theridiidae). Journal of Natural History 32:545-604. Krupa, J.J. & A. Sih. 1998. Fishing spiders, green sunfish, and a stream-dwelling water strider: male-female conflict and prey responses to single versus multiple predator environments. Oecolo- gia 117:258-265. Masumoto, T 1993. The effect of the copulatory plug in the funnel-web spider, Agelena limbata (Araneae, Agelenidae). The Journal of Arach- nology 21:55-59. Robinson, M.H. 1982. Courtship and mating be- havior in spiders. Annual Review of Entomology 27:1-20. Samu, E, J. Nemeth, F. Toth, E. Szita, B. Kiss & Cs. Szinetar. 1998. Are two cohorts responsible for bimodal life history pattern in the wolf spider 414 THE JOURNAL OF ARACHNOLOGY Pardosa agrestis in Hungary? Proceedings 17* European Colloquium of Arachnology, Edin- burgh. Pp. 215-221. Samu, F. & Cs. Szinetar. 2002. On the nature of agrobiont spiders. Journal of Arachnology 30: 389-402. Samu, E, A. Sziranyi & B. Kiss. 2003. Foraging in agricultural fields: local ‘sit-and-move’ strategy scales up to risk-averse habitat use in a wolf spi- der. Animal Behaviour 66:939-947. Schafer, M.A. & G. Uhl. 2002. Determinants of pa- ternity success in the spider Pholcus phalangioi- des (Pholcidae: Araneae): the role of male and female mating behaviour. Behavioral Ecology and Sociobiology 51:368-377. Scheffer, S.J. 1992. Transfer of a larval mantispid during copulation of its spider host. Journal of Insect Behavior 5:797-800. Stalhandske, P. 2001. Nuptial gift in the spider Pi- saura mirabilis maintained by sexual selection. Behavioral Ecology 12:691-697. Stratton, G.E., E.A. Hebets, PR. Miller & G.L. Miller. 1996. Pattern and duration of copulation in wolf spiders (Araneae, Lycosidae). Journal of Arachnology 24:186-200. Suter, R.B. 1990. Courtship and the assessment of virginity by male bowl and doily spiders. Animal Behaviour 39:307-313. Suter, R.B. & V.S. Parkhill. 1990. Fitness conse- quences of prolonged copulation in the bowl and doily spider. Behavioral Ecolology & Sociobiol- ology 26:369-373. Walker, S.E., A.L. Rypstra & S.D. Marshall. 2003. The relationship between offspring size and per- formance in the wolf spider Hogna helluo (Ar- aneae: Lycosidae). Evolutionary Ecology Re- search 5:19-28. Watson, RJ. & J.R.B. Lighten. 1994. Sexual selec- tion and the energetics of copulatory courtship in the sierra dome spider, Linyphia litigiosa. An- imal Behaviour 48:615-626. Yamaoka, K. & T. Hirao. 1977. Stimulation of vir- ginal oviposition by male factor and its effect on spontaneous nervous activity in Bombyx mod. Journal of Insect Physiology 23:57-63. Manuscdpt received 2 November 2004, revised 12 September 2005. 2005. The Journal of Arachnology 33:415-425 LARVAL CHAETOTAXY IN WOLF SPIDERS (ARANEAE, LYCOSIDAE): SYSTEMATIC INSIGHTS AT THE SUBFAMILY LEVEL Beata Tomasiewicz: Zoological Institute, Department of Biodiversity and Evolutionary Taxonomy, Przybyszewskego 63/77, 51-148 Wroclaw, Poland. E-mail: beatatomas @ interia.pl Volker W. Framenau: Department of Terrestrial Invertebrates, Western Australian Museum, Locked Bag 49, Welshpool DC, Western Australia 6986, Australia. ABSTRACT. Studies into the systematics of wolf spiders have mainly employed morphological char- acters of adult spiders, in particular features of the male and female genitalia, and more recently mito- chondrial DNA sequence data. However, there is still no established phylogenetic framework for the Lycosidae, even at the subfamily level. This study uses a novel morphological character set, the chaetotaxy of lycosid larvae (presence and arrangement of setae and slit organs), to infer systematic information on seven species of wolf spiders that are currently listed in three subfamilies: Lycosinae [Alopecosa pulver- ulenta (Clerck 1757), Hogna antelucana (Montgomery 1904), Rabidosa rabida (Walckenaer 1837), Tro- chosa ruricola (DeGeer 1778)], Piratinae [Hygrolycosa rubrofasciata (Ohlert 1865), Pirata hygrophilus (Clerck 1757)], and Sosippinae (Sosippus californicus Simon 1898). Cheliceral and tarsal (legs 1 and 11) chaetotaxic patterns of the first postembryo showed equivalent chaetotaxic complexes amongst all species but revealed considerable differences between representatives of the three subfamilies. Sosippus califor- nicus showed the most complex pattern and P. piraticus the most reduced arrangement. In addition, it casts doubt on the previous listings of H. rubrofasciata in either the Lycosinae or Piratinae, as its chae- totaxic setae arrangement was more similar to S. californicus than to any other species investigated here. Keywords: Larval stages, chaetotaxic complex, Lycosinae, Piratinae, Sosippinae Chaetotaxy, the presence and arrangement of setae and other sensory structures on the integument of arthropods, has been widely used for systematic and taxonomic studies in a variety of groups, including insects (e.g. Alarie & Watts 2004) and arachnids such as mites (Tuzovsky 1987; Griffith et al. 1990) and pseudoscorpions (e.g. Chamberlin 1931; Harvey 1992). In contrast, investigations into the chaetotaxy of spiders are comparatively rare and have focused mainly on trichoboth- rial patterns. These have been argued to be a suitable feature in higher level systematics (Lehtinen 1980; Scioscia 1992). They may also serve as an important tool in identifica- tion at the species level. For example, the po- sition of the metatarsal trichobothrium has been used as an essential character in the iden- tification of central European Micryphantinae Bertkau 1872 (Wiehle 1960; Heimer & Nen- twig 1990). There is a considerable difference between the chaetotaxic pattern of immature and adult arthropods and larval chaetotaxy has been ar- gued to represent an excellent character set for systematic studies (Pomorski 1996). Larval morphology is of particular interest in holo- metabolic insects such as beetles (Kilian 1998; Borowiec & Swi^tojanska 2003) and butterflies (Kitching 1984, 1985) as different expressions of the same genotype can com- plement the morphological characters of adults (Alarie & Watts 2004; Grebennikov 2004; Ashe 2005). However, larval chaetotaxy has also been useful in phylogenetic and tax- onomic studies of arthropods with gradual larval development such as springtails, mites and pseudoscorpions (Nayrolles & Betsch 1993; Pomorski 1996; Griffith et al. 1990; Harvey 1992). Early stages of development may last for only a short period of time, which in many cases eliminates the devel- opment of distinct adaptive traits. In addition, the morphology of juveniles is less variable and complex than that of adults (Pomorski 1996). 415 416 THE JOURNAL OF ARACHNOLOGY Figures 1, 2. — Larval setae of wolf spiders. 1. Seta form [position: apical/etc leg/chelicera etc] of [speciesl; 2. Serrated seta from [position: apical/etc., leg/chelicera etc.] of [species]. Scale bar: 10 pm (Fig. 1), 5 pm (Fig. 2). Studies on chaetotaxic structures in imma- ture spiders are rare and initially focused on trichobothrial patterns (Emerit 1964). A recent study of the linyphiid spider Bathyphanthes eumenis (L. Koch 1879) included all sensory structures of the protonymph and showed that the arrangement of sensory organs such as se- tae, trichobothria and slit organs was constant in all examined specimens and may have the potential to serve in the identification of spi- ders at the generic and species level (Rybak & Pomorski 2003). The nomenclature of the chaetotaxic patterns developed for B. eumenis was subsequently used in a detailed compar- ative study including the wolf spider Trochosa ruricola (DeGeer 1778) (Lycosidae) (Rybak & Tomasiewicz 2005). Although this study showed considerable differences in chaetotax- ic pattern between both species, some body parts showed very similar setae distribution, which suggested homology for a large number of chaetotaxic complexes. Despite recent investigations into the sys- tematics of wolf spiders, there is still no ac- cepted phylogenetic framework for the Lycos- idae, even at the subfamily level (e.g. Dondale 1986; Zyuzin 1993; Vink et al. 2002). This problem can be attributed to a lack of well- defined morphological characters that could classify and separate particular genera and subfamilies. However, there appears to be a consensus that web-building wolf spiders, such as the genera Sosippus Simon 1888 (sheet-web) and Pirata Sundevall 1833 (tube- shaped retreat) represent more ancient evolu- tionary lines in comparison to genera within the Lycosinae Simon 1898 {Trochosa C.L. Koch 1847, Alopecosa Simon 1885, Rabidosa Roewer 1960 and Hogna Simon 1885) that are considered representatives of more recent evo- lutionary lineages (Dondale 1986; Zehethofer & Sturmbauer 1998; Vink et al. 2002). The genus Hygrolycosa Dahl 1908 was for- merly placed in the Lycosinae along with, amongst others, Alopecosa, Hogna and Tro- chosa (Dondale 1986). However, more re- cently, it was listed in a separate subfamily, Piratinae Zyuzin 1993, based on the shape and location of the embolus and the functional conductor in the male pedipalp (Zyuzin 1993). Current molecular evidence suggests that Hy- grolycosa is a sister taxon to Aulonia albi- mana (Walckenaer 1805) in a clade that also includes Pirata, Venonia Thorell 1894 (Ven- oniinae Lehtinen & Hippa 1979) and Xeroly- TOMASIEWICZ & FRAMENAU— LARVAL CHAETOTAXY IN WOLF SPIDERS 417 Table 1. — Nomenclature of chaetotaxic complexes on the larval integuments of A. pulverulenta, H. antelucana, H. rubrofasciata, P. hygrophilus, R. rabida, S. californicus and T. riiricolci. Abbreviation Chaetotaxic complex Illu.strations Chelicerae dorsal apical complex Figs. 1-4 Chpivi dorsal median complex Figs. 1-4 ChvAM ventral apico-median complex Figs. 5-8 ChvM ventral median complex Figs. 7-8 Tarsi I and II Tda dorsal apical complex Fig. 9 Tdai’ Tdaii. .. first, second, . . . dorsal apical complex Figs. 10-11 Tdm dorsal median complex Fig. 9 Tdmi’ Tqmii. . . first, second, . . . dorsal median complex Figs. 10-13 Top dorsal proximal complex Figs. 10-13 TvaI’ Tvaii. . . first, second, . . . ventral apical complex Figs. 12-14 Tvm ventral median complex Fig. 12 TvMh Tvmii. . . first, second, . . . ventral medial complex Figs. 13-14 Typ ventral proximal complex Figs. 13-14 cosa Dahl 1908 (Evippinae Zyuzin 1985) (N. Murphy et al. in press). The main objective of this study was to evaluate the significance of larval chaetotaxic patterns for systematic analyses in wolf spi- ders. More specifically, we used the ambigu- ous subfamily placement of H. rubrofasciata to assess its previous listings in either the Ly- cosinae or Piratinae by including representa- tives of these subfamilies in our comparative analysis. METHODS We analyzed the larval stages of seven spe- cies of wolf spiders currently listed in three different subfamilies: Lycosinae [Alopecosa pulverulenta (Clerck 1757), Hogna antelu- cana (Montgomery 1904), Rabidosa rabida (Walckenaer 1837), and Trochosa ruricola (DeGeer 1778)], Piratinae {Hygrolycosa rub- rofasciata and Pirata hygrophilus (Clerck 1757)], and Sosippinae (Sosippus californicus Simon 1898). We obtained immature stages through laboratory colonies (T. ruricola, A. pulverulenta, H. rubrofasciata and P. hygro- philus) or loan and donation of material from overseas collections {H. antelucana, R. rabida and S. californicus). Overall, we studied 64 specimens of T. rur- icola, 10 specimens each of H. rubrofasciata and P. hygrophilus and 5 specimens each of A. pulverulenta, H. antelucana, R. rabida, and S. californicus. There was no intraspecific var- iation in regard to the number of structures within chaetotaxic complexes, which allowed analysis of data without statistical consider- ation of variation. Specimens were transfen'ed to 5% KOH and cleared in distilled water. Subsequently, they were placed in chloramphenol and mounted in Swan medium (20 g distilled wa- ter, 60 g chloral hydrate, 15 g gum arabic, 3 g glucose, 2 g glacial acetic acid). All slides were examined under a phase contrast micro- scope (Nikon Eclipse E 600) with a drawing attachment. Scanning electron microscope (SEM) photographs were taken with a JEOL JSM-5800 LV at 15kV after spray-coating the specimen with gold. Voucher specimens of the species examined were lodged at the Museum of Natural History, Wroclaw (A. pulverulenta, P. hygrophilus, H. rubrofasciata) and the Cal- ifornia Academy of Sciences, San Francisco {H. antelucana, R. rabida, S. californicus). Larval stages. — We investigated the first immature stage that possesses chaetotaxic structures on the integument, i.e. the first pos- tembryo. These young spiders develop inside the egg-sac followed by the protonymph, which abandons the egg-sac (Vachon 1957). Vachon (1957) proposed the term ‘larva’ for the first postembryo, which corresponds to ‘stage D’ (Holm 1940), ‘prejuvenile (Ji 1)’ (Canard 1987), ‘larva “setose stage’” (Hallas 1988), and ‘IV instar’ (Galiano 1991). Con- 418 THE JOURNAL OF ARACHNOLOGY sequently, all references to Tarvae’ or ‘larval’ in this study refer to the first postembryo. Chaetotaxic structures. — Although nu- merous chaetotaxic structures such as spines, trichobothria, proprioreceptors in the form of hair plates, and chemoreceptors in the form of tarsal organs, and taste hairs exist in spiders (Foelix 1996; Rybak & Pomorski 2003), this study deals with setae and slit organs because only these structures were observed on the lar- val integument. In adult spiders, setae are tri- ply innervated hair-like structures that serve purely mechanical tasks (tactile receptors). They consist of a long exocuticular shaft of variable shape (including serrated and plu- mose), which is suspended in a slipper-shaped socket in which it can move (Rybak & Po- morski 2003). In contrast, spines are rigid structures that are regarded as hemolymph pressure receptors (Foelix & Chu-Wang 1973). In immature spiders, it is difficult to distinguish between spines and setae as the socket and the setae are generally not fully developed (Figs. 1, 2), although different types of setae may exist (Bond 1994). Con- sequently, within the scope of our study, we do not differentiate between setae and spines. Slit organs occur both in adult and larval spi- ders. They sense mechanical stress in the exo- skeleton caused by vibrations, gravity or the spider’s own movement and occur singly (‘slit sensillae’) or in groups where slits run parallel to each other (‘lyriform organs’) (Foelix 1996). In this study, the chaetotaxic structures on larval chelicerae and tarsi were grouped into distinct complexes. The nomenclature of these complexes follows Rybak & Pomorski (2003) and Tomasiewicz & Rybak (2005) (see also Table 1). RESULTS There were considerable differences in the number of chaetotaxic structures on the larval bodies of the investigated species, which al- lowed separating them into two main groups (Tables 2 & 3). While A. pulverulenta, H. an- teliicana, P. hygrophilus, R. rabida and T. riiricola possessed chaetotaxic structures only on the chelicerae, labium, maxillae, legs and pedipalps, H. ruhrofasciata and S. californi- ciis exhibited chaetotaxy on all body parts in- cluding sternum, carapace, abdomen and spin- nerets. Chelicerae and the tarsi of legs I and II showed distinct chaetotaxic patterns, which allowed a comparison between species and genera. These structures were most complex in S. californicus (Figs. 6, 10, 13, 16) and H, | ruhrofasciata (Figs. 5, 9, 12, 15) and most reduced in P. piraticus (Figs. 3, 7, 11, 14). Chelicerae dorsal. — All species possessed the apical complex Cho^. The number of setae within this complex differed between P. hy- grophilus (four setae; Fig. 3), a group com- prising T. ruricola, A. pulverulenta, R. rabida, and H. antelucana (seven setae; Fig. 4) and a group with H. ruhrofasciata and S. californb cus (10 setae; Figs. 5, 6). Hygrolycosa rub- rofasciata and S. californicus had an addition- al median complex Ch^m that consisted of three setae, which were long in S. californicus and very short in H. ruhrofasciata. All species had one slit sensilla in the median section of the chelicerae and two slit sensillae apically (Figs. 3-6). Chelicerae ventral. — All species showed an apico-median complex Chy^M that consist- ed of one or two setae in P. hygrophilus (Fig. 7), and four setae in all other species (Figs. 8-10). Hygrolycosa ruhrofasciata (Fig. 9) and S. californicus (Fig. 10) possessed a further structure Chy^, consisting of a single seta in H. ruhrofasciata and two setae in S. califor- nicus. The latter species showed an additional apical seta Chy^ that did not exist in any of the other lycosids. All species possessed two slit sensillae apically (Figs. 7-10). Tarsi of legs I and II dorsal. — All species examined showed two similar complexes, T^^ and Tdm in T. ruricola, H. antelucana, R. ra- bida, A. pulverulenta, and P. hygrophilus (Fig. 11), corresponding to Tdaih and Tp^iy in H. ruhrofasciata (Fig. 12) and Td^h and T^miv in S. californicus) (Fig. 13). Hygrolycosa rub- rofasciata (Fig. 12) and S. californicus (Fig. 13) showed seven more complexes in which the apical ones had a larger number of setae in S. californicus. All lycosids showed slit sensillae located laterally in the median part of the tarsi (Figs. 11-13). Tarsi of legs I and II ventral. — All spe- cies showed three identical complexes, Ty^j, Ty^ii, and Ty^ in T. ruricola, H. antelucana, A. pulverulenta and P. hygrophilus (Fig. 14), corresponding to Ty^i, Ty^ni, and Tyyu in H. ruhrofasciata and S. californicus (Figs. 15- 16). The complex Ty^j consisted of three setae in H. ruhrofasciata (Fig. 15) (as the equiva- lent complex Tyjvi in the other above-men- TOMASIEWICZ & FRAMENAU— LARVAL CHAETOTAXY IN WOLF SPIDERS 419 Figures 3-6. — Chaetotaxic pattern on dorsal side of the chelicerae in wolf spider larvae: 3. Pirata hygrophilus; 4. Alopecosa pulverulenta, Hogna antelucana, Trochosa ruricola, Rabidosa rabida; 5. Hy~ grolycosa rub rof as data; 6. Sosippus calif ornicus . Scale bar: 0,1 mm. Multiple scale bars in Fig. 4 reflect the comparative scale of the species in the given sequence. 420 THE JOURNAL OF ARACHNOLOGY Figures 7-10. — Chaetotaxic pattern on ventral side of the chelicerae in wolf spider larvae: 7. Pirata hygrophilus; 8. Alopecosa pulverulenta, Hogna antelucana, Trochosa ruricola, Rabidosa rabida; 9. Hy- grolycosa rubrofasciata; 10. Sosippus californicus. Scale bar: 0.1 mm. Multiple scale bars in Fig. 8 reflect | the comparative scale of the species in the given sequence. ' TOMASIEWICZ & FRAMENAU— LARVAL CHAETOTAXY IN WOLF SPIDERS 421 tioned lycosids), but included four setae in S. californicus (Fig. 16). Sosippus californicus and H. rubrofasciata showed six additional complexes (Tvah? Tyj^n, Tvyim, Tyjyjjy, Typ in H. rubTofciscicittt and TyAn? Tymn, Tyyjui, Tvmiv. Tvmv. Typ in Y. californicus (Figs. 15, 16). Although these two species showed the most similar chaetotaxic patterns, there are complexes (Ty^iv. Tyc in H. rubrofasciata and Tvmiv» Tvmv in 5. californicus) (Figs. 15, 16) among which it is difficult to establish ho- mology. Both species showed slit sensillae sit- uated medially near the apical part of the tarsi, which were absent in all other species (Figs. 14-16). DISCUSSION Our analysis of chaetotaxic patterns in wolf spiders showed distinct and regular complexes for all species examined. These complexes ap- pear to be similar to the arrangement in other spider families such as the Linyphiidae (Ry- bak & Pomorski 2003), which suggests that larval chaetotaxy may serve as a very useful character set in systematic studies if homolo- gies can be established on a higher taxonomic level. However, there was no difference of chaetotaxic patterns among any of the species currently included in the subfamily Lycosinae, In contrast to other arthropods, in particular insects (e.g. Deraaz et ah 1991; Alarie & Watts 2004), larval chaetotaxy does not seem to be suitable for the identification of taxa be- low subfamily level in wolf spiders. There were significant differences in the number of complexes of cheliceral and tarsal setae and the number and size within these complexes. Sosippus californicus showed the most complex pattern along with H. rubrofas- ciata that differed only in the absence of two setae on the ventral side of the chelicerae, the absence of the complex equivalent to Ty^iy in S. californicus, and a reduction in the number of setae in the apical and proximal complexes of the tarsi. On the other hand, all four species of Lycosinae and P. piraticus showed very similar setal arrangements (Table 2). Here, chaetotaxy was heavily reduced in comparison to Sosippus and Hygrolycosa, in particular in regard to the tarsal setae. Pirata piraticus had the lowest number of setae as complex CH^a and CHy^ had two setae less each than the equivalent complexes in the Lycosinae. This separation into two major groups, supported by the overall distribution of chaetotaxic com- plexes on the bodies of the spiders (Table 3), does not reflect current phylogenetic hypoth- eses for wolf spiders. Morphological (Dondale 1986) and molecular (Zehethofer & Sturm- bauer 1998; Vink et al. 2002; Murphy et al. in press) phytogenies consider the Lycosinae as the most derived lineage of wolf spiders, whereas the Piratinae and Sosippinae are thought to represent more basal evolutionary lines. Although we included a wide range of taxa from different currently recognized subfam- ilies our study is ambiguous in regards to the plesiomorphic condition for larval chaetotax- ic structures. Both Pirata and the sheet-web building Sosippus are thought to represent basal lineages in the evolution of wolf spi- ders but they differ considerably in their lar- val chaetotaxy. Preliminary studies on the chaetotaxy of Pisaura mirabilis (Clerck 1757) representing the Pisauridae, a putative sister taxon of the Lycosidae (Dondale 1986; Griswold 1993), show considerably reduced chaetotaxic patterns (Tomasiewicz unpub. data) supporting P. hygrophilus to display the plesiomorphic state. In this case, and in com- bination with current phylogenetic hypothe- ses (Murphy et al. in press), an increase in chaetotaxic structures has evolved twice within our sampled taxa, in Sosippus and Hy- grolycosa. The chaetotaxic pattern of H. rubrofascia- ta differs considerably from all other lycos- ine and piratinae species examined, the two subfamilies where it was previously listed (Dondale 1986; Zyuzin 1993) and our study suggests an alternative placement within the Sosippinae. However, current molecular data place H. rubrofasciata in a basal lineage within in the Lycosidae, close to the Venon- iinae, Piratinae and Evippinae (Murphy et al. in press), providing support for Zyuzin’s (1993) placement of the genus and at the same time rejecting chaetotaxic patterns as informative for the subfamilial placement of Hygrolycosa. This preliminary study shows that larval chaetotaxy may provide some additional mor- phological evidence that bears phylogenetic information in spiders although some discrep- ancies with common tenets of current phylo- genetic hypotheses in wolf spiders exist. It is not possible to distinguish species or even 422 THE JOURNAL OF ARACHNOLOGY Figures 11-13. — Chaetotaxic pattern on dorsal side of the tarsi of legs I and II in wolf spider larvae: 1 1. Pircita hygrophiliis, Alopecosa pulverulenta, Hogna antelucana, Trochosa ruricola, Rabidosa rabida; 12. Hygrolycosa rubrofasciata’, 13. Sosippus califoniicus. Scale bar: 0.1 mm. Multiple scale bars in Fig. 1 1 reflect the comparative scale of the species in the given sequence. Figures 14-16. — Chaetotaxic pattern on ventral side of the tarsi of legs 1 and II in wolf spider larvae: 14. Pi rata hygrophiliis, Alopecosa pulverulenta, Hogna antelucana, Trochosa ruricola, Rabidosa rabida', 15. Hygrolycosa rubrofasciata', 16. Sosippus califoniicus. Scale bar: 0.1 mm. Multiple scale bars in Fig. 14 reflect the comparative scale of the species in the given sequence. genera on the basis of this feature, but the analysis of chaetotaxic patterns may help to establish relationships among subfamilies or above. A more detailed analysis not only into the presence and absence but also the position and the shape of the setae, similar to a study in astigmatid mites (Griffith et al. 1990), may prove helpful in establishing a detailed and more informative character set based on larval chaetotaxy. Presence or absence of setae con- TOMASIEWICZ & FRAMENAU— LARVAL CHAETOTAXY IN WOLF SPIDERS 423 Table 2. — Number of setae per chaetotaxic complex on the chelicerae and tarsi of leg II and III of A. pulverulenta, H. antelucana, H. rubrofasciata, P. hygrophilus, R. rabida, S. californicus and T, ruricola. P. hygrophilus A. pulverulenta H. antelucana R. rabida, T. ruricola H. rubrofasciata S. californicus ChoA 1 1 10 10 ChoM — — 3 3 ChvAM 2 4 4 4 ChvM — — 1 2 Tda 3 3 — — Tdai — — 5 5 Tdao — — 2 3 TdAIII — — 3 2 Tdm 2 2 — — Tdmi — — 4 4 Tdmii — — 3 3 Tdmiv — — 2 2 Tdmv — — 2 2 Tdp — — 2 2 Tvai 2 2 2 2 Tvaii — — 4 4 Tvaiii — — 3 3 Tvm 3 3 — — Tvmi — — 3 4 Tvmii — — 4 5 Tvmiii — — 3 2 Tvmiv — — 4 3 Tvmv — — — 3 Tyc — — 2 — Typ — — 4 4 tain only a linaited amount of information since a reduction of structures may have easily occurred in multiple evolutionary lines. Dis= tinguishable morphological categories of setae exist in wolf spider larvae, for example smooth and serrated forms (Figs. 1,2). Future research could explore an expanded character set and subsequently code it as morphological matrix for a phylogenetic analysis similar to some studies of insects (e.g., Alarie & Watts 2004; Ashe 2005). This could then be incor- porated in an exhaustive morphological and molecular dataset for higher phylogenetic analysis in spiders. The analysis of larval chaetotaxy may bear considerable importance in interpreting struc- tures of mature spiders, in particular during character polarization as part of a phyloge- netic analysis (ontogenetic criterion, see Hee- nig 1966; Nelson 1978; Mabee 2000). For ex- ample, the study of setal arrangement during postembryonic development has been helpful in determining the phylogenetic migration of homological chelal trichobothria in pseudo- scorpions (Harvey 1992) and the setal ar- rangement in astigmatid mites (Griffith et al. 1990). Currently, it remains difficult to acquire lar- val material for morphological studies since larvae and juveniles are often discarded dur- ing the collection of spiders, and, if collected, the material may not represent a suitable de- velopmental stage for comparative studies. However, if larval chaetotaxy can be estab- lished as an important morphological tool in phylogenetic studies of spiders, the collection and preservation of spider larvae may receive much stronger support. ACKNOWLEDGMENTS We are grateful to Darrell Ubick (California Academy of Science) for the loan and dona- tion of specimens of H. anteiucana, R. rabida and 5. californicus for this study. Melissa Thomas and Mark Harvey provided helpful comments to improve this manuscript. TB was 424 THE JOURNAL OF ARACHNOLOGY Table 3. — Presence of chaetotaxic structures on the larval bodies of A, pulverulenta, H. antelucana, H. rubrofasciata, P. hygrophilus, R. rabida, S. californicus and T. ruricola. si - slit sensillae, ly - lyriform organs. A. pulverulenta H. antelucana, P. hygrophilus R. rabida, T. ruricola H. rubrofasciata S. californicus Chelicerae setae, si setae, si Labium setae setae Maxillae setae setae Sternum — setae, si Carapace — setae Pedipalps setae, ly, si setae, ly, si Legs setae, ly, si setae, ly, si Abdomen — setae Spinnerets — setae funded through the Zoological Institute of the Wroclaw University and VWF through the Australian Biological Resources Studies (ABRS) to Mark Harvey (Western Australian Museum) and Andy Austin (The University of Adelaide). LITERATURE CITED Alarie, Y. & H.S. Watts. 2004. Larvae of the genus Antiporus (Coleoptera: Dytiscidae) and phylo- genetic implications. Invertebrate Systematics 18:523-546. Ashe, J.S. 2005. Phylogeny of the tachyporine group subfamilies and basal lineages of aleo- charines (Coleoptera: Staphylinidae) based on larval and adult characteristics. Systematic En- tomology 30:3-37. Bond, J.E. 1994. Seta-spigot homology and silk production in first instar Antrodiaetus imicolour spiderlings (Araneae: Antrodiaetidae). Journal of Arachnology 22:19-22. Borowiec, L. & J. Swigtojahska. 2003. The first in- star larva of Cassida nebidosa L. (Coleoptera: Chrysomelidae: Cassidinae)- a model descrip- tion. Annales Zoologici 53:189-200. Canard, A. 1987. Analyse nouvelle du developpe- ment postembryonnaire des araignees. Revue Ar- achnologique 7:91-128. Chamberlin, J.C. 1931. The arachnid order Chelo- nethida. Stanford University Publications Uni- versity Series, Biological Sciences 7:1-284. Deruaz, D., J. Deruaz & J. Pichot. 1991. Corre- spondence analysis of larval chaetotaxy in the '"Anopheles macuUpennis complex” (Diptera, Culicidae). Annales de Parasitologie Humaine et Comparee 66:166-172. Dondale, C.D. 1986. The subfamilies of wolf spi- ders (Araneae: Lycosidae). Actas X Congreso In- ternacional de Aracnologia, Jaca, Espana 1:327- 332. Emerit, M. 1964. La trichothriotaxie et ses varia- tions au cour du developpement post-embryon- naire chez I’araignee Gasteracantha versicolor (Walck.) (Argiopidae). Comptes Rendus de FAcademie des Sciences Paris 258:4853-4854. ; Foelix, R.E 1996. Biology of Spiders. Oxford Uni- versity Press, New York. Foelix, R.E & I.-W. Chu-Wang. 1973. The mor- phology of the spider sensilla. I. Mechanorecep- tors. Tissue Cell 5:451-478. Galiano, M.E. 1991. Postembryonic development in ten species of neotropical Salticidae (Araneae). Bulletin of British Arachnological Society 8: 209-218. Grebennikov, V.V. 2004. Review of larval mor- phology of the beetle suborder Archostemata (In- ; secta: Coleoptera) with emphasis on first-instar chaetotaxy. European Journal of Entomology 101:273-292. Griffith, D.A., W.T. Atyeo, R.A. Norton & C.A. Lynch. 1990. The idiosomal chaetotaxy of astig- matid mites. Journal of Zoology, London 220:1- 32. Griswold, C.D. 1993. Investigations into the phy- logeny of the lycosoid spiders and their kin (Arachnida: Araneae: Lycosoidea). Smithsonian Contributions to Zoology 539:1-39. Hallas, S.E.A. 1988. Hatching and early postem- bryonic development in the Salticidae. Bulletin of British Arachnological Society 7:231-236. Harvey, M.S. 1992. The phylogeny and classifica- tion of Pseudoscorpionida (Chelicerata: Arach- | nida). Invertebrate Taxonomy 6:1373-1435. Heimer, S. & W. Nentwig. 1991. Spinnen Mitteleu- i ropas. Ein Bestimmungsbuch. Paul Parey, Berlin i and Hamburg. I Hennig, W. 1966. Phylogenetic Systematics. Uni- I versity of Illinois Press, Urbana. j Holm, A. 1940. Studien fiber die Entwicklung und | TOMASIEWICZ & FRAMENAU— LARVAL CHAETOTAXY IN WOLF SPIDERS 425 Entwicklungsbiologie der Spinnen. Zoologiska Bidrag fran Uppsala 19:1-289. Lilian, A. 1998. Morphology and phylogeny of the larval stages of the tribe Agathidiini (Coleoptera: Leiodidae: Leiodinae). Annales Zoologici 48: 125-220. Kitching, I.J. 1984. The use of larval chaetotaxy in butterfly systematics, with special reference to the Danaini (Lepidoptera: Nymphalidae). Sys- tematic Entomology 9:49-61. Kitching, I.J. 1985. Early stages and the classifica- tion of the milkweed butterflies (Lepidoptera: Danainae). Zoological Journal of the Linnean Society 85:1-97. Lethinen, P. 1980. Trichobothrial patterns in high level taxonomy of spiders. Proceedings of the Vlllth International Arachnological Congress, Vienna. Pp. 493-498. Mabee, P.M. 2000. The usefulness of ontogeny in interpreting morphological characters. Pp. 84- 114. In Phylogenetic Analysis of Morphological Data. (J.J. Wiens, ed.). Smithsonian Institution Press, Washington and London. Murphy, N.P., V.W. Framenau, S.C. Donnellan, M.S. Harvey, Y.-C. Park & A.D. Austin. In press. Phylogenetic reconstruction of the wolf spiders (Araneae, Lycosidae) using sequences from 12S rRNA, 28S rRNA and NADHl genes: implications for classification, biogeography and the evolution of web-building behavior. Molec- ular Phylogenetics and Evolution. Nayrolles, P. & J.M. Betsch. 1993. Pour une theorie de la description chetotaxique chez les collem- boles. Annales de la Societe Entomologique de France (N.S.) 29:5-15. Nelson, G.J. 1978. Ontogeny, phylogeny, palaeon- tology, and the biogenetic law. Systematic Zo- ology 27:324-345. Pomorski, R.J. 1996. The first instar larvae of On- ychiurinae — a systematic study (Collembola: Onychiuridae). Genus 7:1-102. Rybak, J. & R.J. Pomorski. 2003. Morphology of protonymph of Bathyphantes eumenis (L. Koch, 1879) (Araneae: Linyphiidae). Genus 14:585- 602. Rybak, J. & B. Tomasiewicz. 2005. The larval chaetotaxy of Bathyphantes eumenis (L. Koch 1879) (Araneae: Linyphiidae) and Trochosa rur- icola (De Geer 1778) (Araneae: Lycosidae) — a model description. Genus 16:129-143. Scioscia, C.L. 1992. Evolucion de la tricobotriotax- ia durante el desarrollo de tres Salticidae neo- tropicales (Araneae). EOS 68:99-206. Tuzovsky, P.V. 1987. Morphology and Postembry- onic Development of Water Mites. Nauka Pub- lications, Moscow. Vachon, M. 1957. Contribution a Petude du devel- oppement postembryonnaire des araignees. note: generalites et nomenclature des stades. Bul- letin de la Societe Zoologique de France 82:337- 354. Vink, C.J., A.D. Mitchell & A.M. Paterson. 2002. A preliminary molecular analysis of phylogenet- ic relationships of Australasian wolf spider gen- era (Araneae, Lycosidae). Journal of Arachnol- ogy 30:227-237. Wiehle, H. 1960. Spinnentiere oder Arachnoidea (Araneae). XL Micryphantidae — Zwergspinnen. Pp. 1-620. In Die Tierwelt Deutschlands und der angrenzenden Meeresteile nach ihren Merkmalen und nach ihrer Lebensweise. 47. Teil (Dahl, E, M. Dahl & H. Bischoff, eds.). Gustav Fischer Verlag, Jena Zehethofer, K. & C. Sturmbauer. 1998. Phylogenetic relationships of Central European wolf spiders (Araneae: Lycosidae) inferred from 12S ribosomal DNA sequences. Molecular Phylogenetics and Evolution 10:391-398. Zyuzin, A. A. 1993. Studies on wolf spiders (Ara- neae: Lycosidae). 1. A new genus and species from Kazakhstan, with comments on the Lycos- inae. Memoirs of the Queensland Museum 33: 693-700. Manuscript received 4 January 2005, revised 17 June 2005. 2005. The Journal of Arachnology 33:426-438 A REDESCRIPTION OF PORRHOMMA CAVERNICOLA KEYSERLING (ARANEAE, LINYPHIIDAE) WITH NOTES ON APPALACHIAN TROGLOBITES Jeremy A. Miller*: Department of Entomology, National Museum of Natural History, NHB-105, Smithsonian Institution, PO Box 37012, Washington, DC 20013-7012 U.S.A. ABSTRACT. The Appalachian troglobite Porrhomma cavernicola (Keyserling 1886) is redescribed. Porrhomma emertoni Roewer 1942 is a junior synonym (new synonymy). An unusual stridulatory organ with the plectrum on trochanter II and the striae on coxa I is found in both sexes of this species. Por- rhomma cavernicola is widespread in Appalachian caves. By contrast, Appalachian Nesticus (Nesticidae) troglobites tend to be highly endemic. This despite the fact that both groups of spiders are web-builders that may be found in the same caves. Porrhomma cavernicola is added to a previous phylogenetic analysis of linyphiid spiders. Implications of this analysis for the phylogenetic structure of linyphiid spiders is discussed. Keywords: Dispersal, phylogeny, stridulatory organ, Nesticus, Nesticidae There is a continuous gradation between epigean and troglobitic organisms. While a variety of spiders are known from cave en- trances or can be found both in and out of caves, true troglobites complete their entire life cycle in caves. In the Appalachian region, true troglobites belong to the Linyphiidae, Nesticidae, Dictynidae and Leptonetidae (Barr 1961; Holsinger & Culver 1988; Gertsch 1992; Peck 1998). Dictynid and leptonetid troglobites in Appalachia are understudied and will not be discussed further here (see Miller 2005). Porrhomma cavernicola (Keyserling 1886) is one of two linyphiid troglobites widespread and often syntopic in Appalachian caves. The other widespread linyphiid troglobite is Pha- netta subterranea (Emerton 1875) (Fig. 1). Anthrobia are also widespread in Appalachian caves, although Miller (2005) has concluded that there are at least two troglobitic Anthro- bia species instead of the one species previ- ously recognized. Some other linyphiid trog- lobites have more restricted distributions (e.g., some Islandiana species, Holsinger & Culver 1988; Gertsch 1992; Peck 1998; also some ’ Current address: Department of Entomology, California Academy of Sciences, 875 Howard Street, San Francisco, CA 94103 U.S.A. E-mail: jmiller@Calacademy.org undescribed species, N. Duperre, pers. comm.). Multiple linyphiid troglobite species can often be found in the same cave. By con- trast, troglobitic species of Nesticus (Nestici- dae) in Appalachian caves are never wide- spread, highly endemic, and rarely syntopic (Fig. 1; Gertsch 1984; Coyle & McGarity 1991; Hedin 1997b; Reeves 2000; Hedin & Dellinger 2005). About eight Appalachian Nesticus species appear to be troglobites (Gertsch 1984, 1992; Hedin 1997a; Hedin & Dellinger 2005). In both Nesticus (Hedin 1997a, b) and linyphiids, troglobitism seems to have occurred independently multiple times. Widespread troglobites are the exception to the rule. Troglobites cannot normally survive long under surface conditions so dispersal be- tween cave-islands across epigean seas must be rare (Barr 1967; Culver 1970, 1971, 1982; Barr & Holsinger 1985; Holsinger & Culver 1988). Thus widespread troglobites must ei- ther be a syndrome of multiple forms erro- neously lumped into a single species by tax- onomists, or genetically isolated populations that have not diverged because of insufficient time or low rates of change, or there must be some mechanism allowing gene flow between caves. Examination of specimens from across the range of Porrhomma cavernicola revealed no clear pattern of geographical variation that 426 MILLER— FOi?/?//OMMA CAVERNICOLA All 80° Pennsylvania V a/ £ r^'VA^- Virginia^^A aa Virginia AAA/ A North Carolina LINYPHIIDAE A Porrhomma cavernicola V Phanetta subterranea □ Anthrobia monmouthia O Anthrobia coylei NESTICiDAE Nesticus holsingeri Nesticus stygius Nesticus barrowsi ^ Nesticus dilutus (3 Nesticus furtivus O Nesticus Georgia O Nesticus barn ® Nesticus Jonesi 40° 35° 80° Figure 1. — Map showing distribution of nesticid and selected linyphiid troglobites in Appalachian caves. Records of Phanetta subterranea from Millidge (1984); of Anthrobia monmouthia and Anthrobia coylei from Miller (2005); of Nesticus species from Gertsch (1984) and Hedin (1997a, 1997b). would suggest cryptic species (Figs. 2-4). This contrasts with the conclusion reached concerning troglobitic Anthrobia. Anthrobia monmouthia was previously considered a widespread Appalachian troglobite, but new work has provided diagnoses and descriptions for two distinct species that were previously confused under this name (Miller 2005). Ef- forts are underway to study the population ge- netics of widespread linyphiid troglobites in Appalachia to evaluate gene flow among caves. METHODS Specimens were examined using a Leica MZ 16 dissecting microscope. Most illustra- tions of the genitalia were made using an Olympus BH-2 compound microscope fitted with a camera lucida. Specimens were tem- porarily cleared in methyl salicylate (Holm 1979), then positioned for illustration on a temporary slide using the method described in Coddington (1983). The illustration of the epi- gynum in ventral view was based on photo- graphs taken using a Nikon DXM 1200F dig- ital camera mounted on a Leica MZ 16. The photograph of the cleared epigynum was tak- en using the digital camera mounted on an Ortholux II compound microscope; multiple images were combined using Auto-Montage by Syncroscopy (version 4.01). SEM images were taken using the AMRAY 1800 at the Na- tional Museum of Natural History Scanning Electron Microscope Facility. All measurements are in millimeters and were taken using a reticle mounted in a Leica MZ APO dissecting microscope. The position of the first metatarsal trichobothrium (Tml) is expressed as the ratio of the distance between the proximal margin of the metatarsus and the root of the trichobothrium divided by the total length of the metatarsus (Denis 1949; Locket & Millidge 1953). Material examined.— When multiple con- secutive records in the material examined sec- tion were from the same locality, the locality 428 THE JOURNAL OF ARACHNOLOGY data after the first record is given in brackets as [same locality]. When data labels did not include geographic coordinates, I attempted to determine the approximate location using maps, gazetteers, and other literature. Once the location was inferred, the coordinates were included in [square brackets]; coordinates tak- en directly from the data label are given in (parentheses). When no coordinates could be determined for any cave within a county, a dot near the geographic center of the county was included in the map (Fig. 1). In most cases, coordinates will not be precise enough that readers will be able to locate caves without additional information. The map was created using ArcView version 8.3. Phylogenetic analysis. — Porrhomma cav- ernicola was coded into the phylogenetic data matrix of Miller (2005); no new characters were added to the analysis. The expanded analysis consists of 87 taxa coded for 176 characters, 172 of which are phylogenetically informative. The majority of characters con- cern the male genitalia, somatic morphology, and female genitalia; a few characters concern behavior and web architecture. See (Miller & Hormiga 2004) for descriptions of characters and character states. Porrhomma cavernicola was coded as follows: 0000001000 0401010- 01 1001000010 0000000101 1000000001 1100100000 0-00000-0 0012001000 0- 00000001 0100100100 000000000? 2100011 000 0041011111 3211111111 1111100000 0007000-00-201 1 1001 1? ?????. Analysis was conducted using heuristic searches in PAUP* (1000 replicates of random taxon ad- dition; Swofford 2001). All characters were treated as unordered and equally weighted. Abbreviations. — The following anatomical abbreviations are used in the text and figures: A = atrium; AC = aciniform gland spigot; AG = aggregate gland spigots; ALE = ante- rior lateral eye; AME = anterior median eye; ARP = anterior radical process; CD = cop- ulatory duct; CL = column; DP = dorsal plate of epigynum; DSA = distal suprategular apophysis; E = embolus; EM = embolic membrane; F = fundus; FE = femur; FL = flagelliform gland spigot; FD = fertilization duct; MT = metatarsus; PA = patella; PC = paracymbium; PLE = posterior lateral eye; PME = posterior median eye; PT = prote- gulum; R = radix; S = spermatheca; SPT = suprategulum; ST = subtegulum; T = tegul- um; TA = tarsus; TI = tibia; TLL = total leg length; Tml = position of first metatarsal tri- chobothrium; TmlV = fourth metatarsal tri- chobothrium; TP = tailpiece of radix; VP = ventral plate of epigynum. Institutional abbre- viations are given in the Acknowledgments. TAXONOMY Family Linyphiidae Blackwall 1859 Genus Porrhomma Simon 1884 Porrhomma Simon 1884: 360. Type species Liny- phia proserpina Simon 1873 {—Erigone convexa Westring 1851, synonymy in Holm 1944: 130, 133) by subsequent designation (Simon 1894; 701). Opistoxys Simon, 1884: 373. Type species Opistox- ys acuta Simon 1884 {=Linyphia microphthalma O. Pickard-Cambridge 1871, synonymy in Thaler 1975: 142) by monotypy. Synonymy in Thaler 1975: 142. Remarks. — The Holarctic genus Porrhom- ma consists of 3 1 species plus one subspecies (Platnick 2004). The species of the genus tend to be quite homogeneous, but within species, the genitalia tend to exhibit a high degree of variation. Species range in total length from about 1.2-3. 2. Most species are epigean, found mostly in cool, mesic habitats including forests, grasslands, and under stones. Some species are troglophilic [e.g., P. convexum (Westring 1861) and P. egeria Simon 1884], while others are troglobites [e.g., P. caverni- cola and P. rosenhaueri (L. Koch 1872)]. Porrhomma cavernicola (Keyserling 1886) Figs. 1-26 Linyphia incerta Emerton 1875:280, figs. 13-21 (3, $); Packard 1875:275; Packard 1888:57; Mac- Cook 1890:292, figs. 284-285; Simon 1894:690. Willibaldia incerta (Emerton); Keyserling 1886: 123; Marx 1890:531; Banks 1910:32. Willibaldia cavernicola Keyserling 1886:123-124, pi. 15, fig. 204 ($); Packard 1888:58; Marx 1890: 531; Comstock 1903:32; Banks 1907:739, Banks 1910:32; Comstock 1913:383; Comstock 1948: 397; Bonnet 1959:4721. Taranucnus cavernicola (Keyserling): Simon 1894: 690. Troglohyphantes cavernicola (Keyserling): Simon 1894:706; Crosby 1905:368, figs. 20-22 (3); Mclndoo 1910:304; Mclndoo 191 la; 183; Mc- Indoo 1911b:39L Troglohyphantes incertus (Emerton): Comstock 1903:32; Comstock 1913:383; Petrunkevitch 1911:272. Troglohyphantes cavernicolus (Keyserling); Com- miller— PORRHOMMA CAVERNICOLA 429 0.10 0.09 ® 0.08 E I 0.07 < 0.06 0.05 X ■ X o o ¥ Ik ♦ o X ▲ o o o A A AGA (5) ■ MO (1) ffl XIN (3) OVA (10) + MD (1) □ WV(18) □ ♦ KY{1) — 1 X Syntype f"'" ^ 0.070 0.075 0.080 0.085 Atrium width 0.090 0.095 0.100 3 4 1 .OV " ^ 1.25 - m X xX o I ♦ c © X X XIN (6) i 1.20 - ."fi ffl ■ ♦ KY(1) ■ MO (1) m S 1.15- O OVA (9) □ WV(18) X Syntype 1.10 - T- — J — 0.95 1.00 1.05 1.10 1.15 1.20 Male carapace length Figures 2-4. — Morphometric variation of selected features in Porrhomma cavernicola. Individuals grouped by state with sample size in parentheses. The syntype specimens of Willibaldia cavernicola are indicated by an asterisk symbol. 2. length and width of the atrium, female epigynum; 3. tibia I length and carapace length in female; 4. tibia I length and carapace length in male. 430 THE JOURNAL OF ARACHNOLOGY Figures 5-1 1. — Porrhomma caveniicola (Keyserling). 5-7. left male palp; 8—1 1. epigynum. 5. prolateral view; 6. retrolateral view; 7. embolic division, mesal view; 8. dorsal view; 9-1 1. ventral view. 5-9. from Sam Six Cave, Wythe County, Virginia; 10. from McFerrin Breakdown Cave, Greenbrier County, West Virginia; 1 1 . from El Rod Cave, Orange County, Indiana. Scale bars = 0. 1 mm. See text for abbreviations. stock 1903:32; Comstock 1913:383; Petrunkev- itch 1911:272; Elliott 1932:425. Willihaldi caveniicola (Keyserling): Banta 1907:62. Porrhomma incerta (Emerton): Berland 1931:384. Porrhomma caveniicola (Keyserling): Roewer 1942:603; Platnick 2004. Porrhomma emertoiii Roewer 1942:603 (replace- ment name for Liiiyphia incerta Emerton 1875, preoccupied by Linyphia incerta Walckenaer 1842, nomen dubium, see van Helsdingen 1972); Platnick 2004. NEW SYNONYMY. Porrhomma incertum (Emerton): Bonnet 1958: 3756. Justification of Synonymy. — Berland (1931) noted that the two nominal taxa appear to differ very little, but he did not synonymize them. After examination of the types and oth- er specimens, I found no morphological evi- dence to maintain two diagnosable species of troglobitic Porrhomma in the Appalachian re- gion. Selected morphometric characteristics j failed to show any regional pattern that might , indicate multiple species (Figs. 2-4). A sim- i ilar approach did reveal the presence of mul- ; tiple species in troglobitic Anthrohia (Miller 2005). Nomenclature. — Linyphia incerta Emerton 1875 is a primary homonym of Linyphia in- certa Walckenaer 1842 and is therefore per- manently invalid (International Commission on Zoological Nomenclature 1999, Article 57.2). Roewer (1942) proposed Porrhomma emertoni Roewer 1942 as a replacement name for L. incerta. Willihaldia caveniicola Key- serling 1886 has priority over P. emertoni. Types.— UNITED STATES: Kentucky: Barren County, Reynolds Cave (BMNH, i t ! UlUJER—PORRHOMMA CAVERNICOLA 1890.7.1.8242-8243, syntypes of Willibaldia cavernicola, examined), 1 c5^, 1 ?. General condition degraded; male abdomen missing, as are most legs for both specimens; female prosoma and abdomen disarticulated. Male: Carapace 1.05 long, 0.76 wide, tibia I 1.26. Female: Carapace 1.16 long, 0.78 wide, atri- um 0.078 long, 0.090 wide. Virginia: Augusta County, Fountain Cave [38°10'N, 78°55'W], Packard (MCZ, syntypes of Liny phia incerta, examined), 2 d, 4 $, 5 juveniles. Additional Material Examined. — UNIT- ED STATES: Georgia: Bartow County, King- ston Saltpeter Cave (34°12'N, 84°54'W), 2 June 1999, W. Reeves (USNM), 5 $. Indiana: Lawrence County, JJ’s Sister Cave (38°45'N, 86°36'W), 26 August 2004, J. Miller (USNM), 1 S ; Orange County, El Rod Cave (38°37'N, 86°31'W), 26 August 2004, P. Pa- quin, J. Miller, J. Lewis, N. Duperre (USNM), 4 d, 1 $, 3 juveniles; [same locality], inside cave, shallow cave, hand collecting, P. Paquin, N. Duperre (USNM, PP-2304), 2 $, juveniles. Kentucky: Carter County, Carter Cave, A Cave [38°22'N 83°07'W], Packard (MCZ), 1 d, 2 $. Maryland: Garrett County, Crabtree Cave, 24 September 1987, pool surface, right passage, D. Feller (USNM, 51B), 1 $; Wash- ington County, Fairview Cave, 30 September 1988, pool surface, mudslide area, D. Feller (USNM, 13 1C), 1 $. Missouri: Boone Coun- ty, Rocheport Cave, 3 miles below Rocheport [38°55'N, 92°30'W], 23 July 1905, C.R. Cros- by (MCZ), 1 d, 1 $. Virginia: Augusta Coun- ty, Fountain Cave [38°10'N, 78°55'W], (MCZ), 1 d ; Page County, Lurray [sic, Luray] Cave [38°39'N, 78°29'W], Kochele (USNM, 187), 2 $ [in two vials]; Page County, Luray Cave [38°39'N, 78°29'W], R.V. Chamberlin (MCZ, 59645), 1 d, 2 $; Montgomery Coun- ty, Aunt Nelli’s hole (37°12'N, 80°22'W), 5 September 2004, P. Paquin, J. Miller, N. Du- perre, R. Storey (USNM), 1 d, 1 juvenile; [same locality], 5 September 2004, inside cave, hand collecting, P. Paquin, N. Duperre (USNM, PP-4304), 1 $ ; Russell County, Car- top Cave, 26 November 1999, D. Hubbard (USNM), 3 d, 2 $ , 3 juveniles; Russell Coun- ty, Maggie Baker Cave, 17 September 1997, D. Hubbard (USNM), 3 9; Scott County, Abram’s Cave, 15 December 1999, D. Hub- bard (USNM), 1 d ; Scott County, Little Duck Cave, 28 November 1997, D. Hubbard 431 (USNM), 2 d, 1 9; Scott County, Queens Cave, 15 April 1997, D. Hubbard (USNM), 1 $ ; Washington County, Robinson Cave, 25 February 1997, D. Hubbard (USNM), 1 d; Wythe County, Sam Six Cave, 25 November 1998, D. Hubbard (USNM), 1 d, 2 9, 3 ju- veniles. West Virginia: Greenbrier County, McFerrin Breakdown Cave, 22 August 2004, visual, E. Saugstad, K. Schneider (USNM), 1 9 ; Mineral County, High Rock Fissure Cave, 30 October 1988, woodrat midden, rope drop, D. Feller (USNM, 133B), 1 d; Monroe Coun- ty, Steeles Cave, 1 1 June 2004, visual (USNM), 1 d, 2 9. Two additional vials had no locality data: 1 d, 1 9, Banks (MCZ, 57184, 1753); 1 9, Banks (MCZ, 59646). Additional Records. — The following re- cords were compiled from literature sources (Mclndoo 1910; Holsinger et al. 1976; Hol- singer & Culver 1988). UNITED STATES: Indiana: Lawrence County, Shawnee Cave [now called Donaldson Cave], 3 miles E Mitchell. Tennessee: Claiborne County: Jen- nings Cave [36°33'N, 83°30'W]; Hawkins County: Sensabaugh Saltpeter Cave [36°34'N, 82°39'W]. Virginia: Augusta County: Glade Cave, Madisons Saltpeter Cave; Bath County: Clark’s Cave [38°5'N, 79°39'W], Crossroads Cave, Porters Cave, Witheros Cave, Banes Spring Cave, Coon Cave; Craig County: New Castle Murder Hole Cave, Rufe Caldwell Cave; Frederick County: Beans Cave [39°9'N, 78°2UW]; Giles County: Clover Hollow Cave [37°19'N, 80°28'W]; Lee County: Unthands Cave, Fisher Cave; Page County: Luray Cav- erns [38°39'N, 78°29'W]; Ruffners Cave No. 1; Roanoke County: Dixie Caverns [37°9'N, 80°9'W]; Rockbridge County: Bell Cave [37°45'N, 79°22'W], Buck Hill Cave [37°36'N, 79°34'W]; Rockingham County: Three-D Maze Cave [38°30'N, 78°45'W]; Ta- zewell County: Gully Cave [37°2'N, 8r38'W], Lawson Cave [37°5^N, 8U2UW]; Wise County: Parsons Cave [36°5UN, 82°42'W]. West Virginia: Berkeley County: Whitings Neck Cave [39°30'N, 77°50'W]; Grant County: Klines Gap Cave [39°4'N, 79°14'W]; Greenbrier County: Bransfords Cave [38°0'N, 80°30'W], Higginbothams Cave [37°56'N, 80°24^W], Organ Cave [37°43'N, 80°26'W], Pollock Cave [37°45'N, 80°37'W], Pollock Saltpeter Cave; Monroe County: Fulton Cave [37°32'N, 80°27'W]; Pendleton County: Moyers Cave [38°34'N, 432 THE JOURNAL OF ARACHNOLOGY Figures 12-19. — Porrhomma cavernicola (Keyserling); 12-16, 18, 19. scanning electron micrographs; 12. photograph. 12-15. left male palp; 16, 17. epigynum; 18. epiandrous region of male; 19. coxa, tro- chanter I and II of female. 12. prolateral view; 13, retrolateral view; 14. embolic division, anow indicates DSA; 15. detail of tegiilum; 16. ventral view, arrow indicates socket in dorsal plate; 17. dorsal view, cleared; 18. arrow indicates epiandrous gland spigots; 19. arrows indicate striae on coxa I. 12-15. from High Rock Fissure Cave, Mineral County, West Virginia; 16, 19. from Fairview Cave, Washington County, Maryland; 17. from Cartop Cave, Russell County, Virginia; 18. from Sam Six Cave, Wythe County, Virginia. Scale bars = 0.01 mm (Fig. 18); 0.1 mm (Figs. 12-17, 19). See text for abbreviations. mU^EK—PORRHOMMA CAVERNICOLA 79°22'W], Mystic Cave [38°49'N, 79°26'W], Schoolhouse Cave [38°47'N, 79°47'W], Sen- eca Caverns [38°47'N, 79°2rW], Stratosphere Balloon Cave [38°46'N, 79°20'W]; Pocahon- tas County: Sharps Cave [38°25'N, 80°5'W]. Diagnosis. — Troglobite distinguished from other Porrhomma species in North America by the extreme reduction of the eyes (Figs. 20-22). Note that P. roserhaueri L. Koch 1872, a cave associated species from Europe and Russia, also has reduced eyes (Locket & Millidge 1953, Wiehie, 1956, Roberts 1993, Platnick 2004). Males of P. cavernicola have the ARP much thicker and more ventrally-di- rected compared to P. rosenhaueri (see Rob- erts 1993, fig. 59E). Description. — Male (from Sam Six Cave, Wythe County, Virginia): Total length 2.43. Carapace 1.19 long, 0.85 wide, orange, squa- mate to reticulate texture. Abdomen white. Eyes minute, laterals separated (see variation, below). Hairs on clypeus and ocular region relatively long (Fig. 20). Sternum 0.58 long, 0.58 wide, light orange. Coxa I with stridu- latory striae on posterior face (as in Fig. 19). Coxa IV separation 0.93 times their width. Chelicerae orange, with three promarginal teeth, four retromarginal teeth; stridulatory striae scale-like (as in Figs. 24, 25). Sulcus present on margin of carapace posterior to chelicerae (as in Fig. 23). Legs orange, tibia I 12.25 times longer than thick; Tml 0.43. Leg I: FE 1.31, PA 0.29, TI 1.23, MT 1.10, TA 0.70, TLL 4.63; leg II: FE 1.23, PA 0.29, TI 1.13, MT 1.10, TA 0.68, TLL 4.63; leg III: FE I. II, PA 0.25, TI 0.90, MT 0.86, TA 0.56, TLL 3.69; leg IV: FE 1.30, PA 0.28, TI 1.29, MT 1.13, TA 0.68, TLL 4.66. Epiandrous gland spigots present (Fig. 18). Anterior lat- eral spinnerets with five aciniform, flagelli- form, and two aggregate gland spigots (Fig. 26); posterior median spinnerets with minor ampullate, two aciniform gland spigots. Palpal tibia with one prolateral, two retrolateral tri- chobothria (Fig. 6); tibial apophysis absent. Retrobasal region of cymbium with short apophysis bearing long setae and glabrous re- gion along margin (Fig. 13). Paracymbium hook-like, proximal part clothed with macro- setae (Fig. 13). Subtegulum ectal to tegulum. Retrolateral part of tegulum partially covered in fine ridges (Fig. 15). Protegulum plus one other distal apophysis of the tegulum present (Fig. 15). Suprategulum continuous with te- 433 gulum, distal suprategular apophysis finger- like, projecting distally (Figs. 5, 14). Embolic division a plate-like radix with a short, pos- terior-projecting tailpiece and a long curved anterior radical process (Fig. 12). Basal part of embolus broadly articulated to radix by a membrane (Fig. 7); distal part of embolus par- tially wrapped by embolic membrane, which is covered in fine papillae (Fig. 14). Female (same locality as male): Total length 2.55. Carapace 1.13 long, 0.79 wide, orange, squamate to reticulate texture. Abdo- men white. Eyes minute, laterals separated (see variation, below). Hairs on clypeus and ocular region not as long as in male (Figs. 21, 22). Sternum 0.54 long, 0.55 wide, orange. Coxa I with stridulatory striae on posterior face (Fig. 19). Coxa IV separation 1.00 times their width. Chelicerae orange, with three pro- marginal teeth, four retromarginal teeth; strid- ulatory striae scale-like (Figs. 24, 25). Sulcus present on margin of carapace posterior to chelicerae (Figs. 22, 23). Legs orange, tibia I 12.13 times longer than thick; Tml 0.44. Pal- pal tibia with one prolateral, two retrolateral trichobothria; palpal tarsus with two dorso- mesal, two dorsoectal macroseta, four ventro- mesal, two ventroectal macrosetae, claw ab- sent. Leg I: FE 1.29, PA 0.29, TI 1.21, MT 1.05, TA 0.46, TLL 4.51; leg II: Fe 1.26, PA 0.29, TI 1.13, MT 0.99, TA 0.64, TLL 4.30; leg III: FE 1.09, PA 0.28, TI 0.86, MT 0.81, TA 0.54, TLL 3.58; leg IV: FE 1.31, PA 0.26, TI 1.28, MT 1.10, TA 0.66, TLL 4.61. Epi- gynum with deep circular atrium (Figs. 9-11, 16). Dorsal plate with socket (Fig. 16). Sper- mathecae crescent-shaped (Figs. 8, 17). Fer- tilization ducts arise from posterior part of spermathecae, long, straight, recurved termi- nally (Fig, 8). Copulatory ducts follow com- plex path, heavily sclerotized proximally, wid- er and less sclerotized close to atrium (Fig. 8). Chaetotaxy: Femur I with one or two dor- sal, two prolateral macrosetae; femur II with one or two dorsal macroseta. Tibia I and II with two dorsal, one prolateral, one retrola- teral macrosetae; tibia III and IV with two dorsal macrosetae. TmlV absent. Tracheae: Haplotracheate, four unbranched trunks confined to abdomen. Variation. — Some or all eyes may be ab- sent; eye loss not always bilaterally symmet- rical. Variation in epigynum illustrated in Figs 2, 9-11. 434 THE JOURNAL OF ARACHNOLOGY Figures 20-26. — Porrhomma cavernicola (Keyserling); scanning electron micrographs. 20. male pro- soma; 21-23. female prosoma, box in 22 defines area of image 23; 24, 25. female chelicera, box in 23 defines area of image 25; 26. anterior lateral spinneret of male. 20, 22. lateral view; 21. anterior view; 23. sulcus; 24. lateral; 25. detail. 20, 26 from Cartop Cave, Russell County, Virginia; 21, 24, 25. from Queens Cave, Scott County, Virginia; 22, 23. from Fairview Cave, Washington County, Maryland. Scale bars = 0.01 mm (Figs. 23, 25, 26); 0.1 mm (Figs. 20-22, 24). See text for abbreviations. Natural History. — Mclndoo (1910) report- ed that P. cavernicola . are found only in total darkness, where the atmosphere is satu- rated. . . ” Mclndoo described the web as a sheet, slightly curved downward with the spi- der on the underside. Distribution. — Known from caves in Geor- gia, Indiana, Kentucky, Maryland, Missouri, mLLBR—PORRHOMMA CAVERNICOLA 435 outgroup (5) Stemonyphantes Tennessee, Virginia, and West Virginia (Fig. 1)= DISCUSSION Phylogenetic Relationships and Charac- ter Evolution. — Analysis of the expanded data matrix (Miller 2005 plus P. cavernicola; see also Miller & Hormiga 2004) yielded a single most parsimonious tree (L = 931, Cl = 0.23, RI = 0.59; with four autapomorphic characters excluded: L = 927, Cl = 0.23; Fig. 27). The topology is identical to that found in Miller (2005) with Porrhomma placed sister to a clade consisting of Mynogleninae plus Er- igoeinae. Porrhomma has traditionally been placed in the Linyphiinae (e.g., Merrett 1963; Millidge 1977; Brignoli 1983). Porrhomma does not form a monophyletic group with the linyphiines included in the analysis (Fig. 27). Admittedly, this analysis suffers from sparse taxon sampling among non-erigonine liny- phiids so the conclusions presented here must be considered preliminary. More robust taxon sampling from a variety of non-erigonine lin- yphiids plus the addition of molecular se- quence data is called for. Miller & Hormiga (2004) added taxa and characters to a previous analysis of erigonine relationships (Hormiga 2000). Considering only taxa common to both studies, relation- ships changed dramatically from one study to the next. Miller & Hormiga (2004) investi- gated whether the addition of taxa, characters, or both were primarily responsible for the changes in the tree. They concluded that most of the changes were due to the addition and modification of characters, not the addition of taxa. Miller (2005) added four taxa from the genus Anthrobia to the Miller & Hormiga (2004) matrix. Consistent with the conclu- sions of Miller & Hormiga (2004) about the relative insensitivity of their topological re- sults to the addition of taxa, relationships of taxa included in both analyses were identical. For this study, one additional taxon has been added. Again, relationships among previous- ly-included taxa are unchanged. Hormiga (1999) reported the presence of lateral sulci on the margin of the prosoma in both males and females in Porrhomma (Figs. 22, 23), as well as Bathyphantes Menge 1866, Diplostyla Emerton 1882, Kaestneria Wiehle 1956, Pacifiphantes Eskov & Marusik 1994, and Vesicapalpus Millidge 1991. Hormiga — Microlinyphia - Linyphia - Bolyphantes — Tenuiphantes - Porrhomma - Mynogleninae (2) Erigoninae (75) Figure 27. — Summary of phylogenetic analysis results showing the position of Porrhomma. Labels representing multiple terminals have the number of taxa in parentheses. See Miller (2005) for a more detailed tree figure (without Porrhomma)', see Mill- er & Hormiga (2004) for characters and states. ( 1 999) pointed out that these sulci represent a derived trait, potentially supporting the mono- phyly of genera exhibiting these sulci. Males of Porrhomma cavernicola retain the triplet, one flagelliform and two aggregate gland spigots necessary for making araneoid sticky silk (Eig. 26; Coddington 1989). The triplet is not found in males of true linyphi- ines, but is retained in the enigmatic genus Stemonyphantes Menge 1866, most erigoni- nes, and in the two mynoglenine genera that have been investigated (Hormiga 2000). It would be useful to investigate the male spin- nerets in Bathyphantes and other genera known to have lateral sulci. Males of P. cavernicola have epiandrous gland spigots (Fig. 18); the loss of these spig- ots is considered a synapomorphy of Erigon- inae (Miller & Hormiga 2004). Porrhomma cavernicola lack a tarsal claw on the female pedipalp. Previous phylogenetic analyses concluded that the loss of the tarsal claw was a synapomorphy for Erigoninae (e.g., Hormiga 2000; Miller & Hormiga 2004). The distribution of the tarsal claw on the tree (Fig. 27) makes this conclusion am- biguous. Either the claw was lost indepen- dently in the branches leading to erigonines and Porrhomma, or the claw was regained in mynoglenines. 436 THE JOURNAL OF ARACHNOLOGY An Unusual Stridulatory Organ. — Bish- op (1925) described the trochanter Il-coxa I stridulatory organ (Fig. 19). Although he at- tributed the organ to members of the genus Troglohyphantes, not Porrhomma, Bishop was almost certainly observing P. cavernico- la. In 1925, P. cavernicola (under two names) was placed in Troglohyphantes; no other Por- rhomma species was placed in Troglohyphan- tes at that time (Platnick 2004). Legendre (1963) reviewed sound production in spiders, including the trochanter Il-coxa I organ. Cit- ing Bishop (1925), Legendre attributed this organ to Troglohyphantes in Europe. How- ever, this organ has not been described for Troglophyphantes in its current circumscrip- tion (Deeleman-Reinhold 1978, Platnick 2004) or for any other linyphiid genus I am aware of. The organ can be found in at least some epigean Porrhomma species (Scharff, pers. comm.). No epigean Porrhomma has ever been classified as in Troglohyphantes. ACKNOWLEDGMENTS Nikolaj Scharff, Julian Lewis, and Matjaz Kuntner made helpful comments on an earlier draft of the manuscript. Pierre Paquin, Nadine Duperre, Julian Lewis, and Rob Story helped with field work. Thanks to Janet Beccaloni of The Natural History Museum, London, UK (BMNH), Gonzalo Giribet and Laura Leiben- sperger of the Museum of Comparative Zo- ology, Harvard, USA (MCZ), and Jonathan Coddington of the National Museum of Nat- ural History, Smithsonian Institution, Wash- ington, D.C., USA (USNM) for the loan of specimens. Thanks also to William Shear, Daniel Feller, Katie Schneider and David Cul- ver for supplying additional specimens, which have been deposited in the USNM, Gustavo Hormiga, Nikolaj Scharff, and Michael Saar- isto participated in enlightening discussions about the anatomy and phylogenetic relation- ships of Porrhomma. Ingi Agnarsson helped with literature research on stridulatory organs. Special thanks to Pierre Paquin for help and encouragement. Scott Whittaker of the USNM Electron Microscope Eacility helped with SEM work. Thanks to Dana DeRoche for technical and curatorial support. This work was supported by a Smithsonian Institution postdoctoral fellowship. Travel to the Inter- national Congress of Arachnology in Gent was supported by a NSF — ATOL grant (DEB 0228699) to Ward Wheeler, Lorenzo Prendini, Jonathan Coddington, Gustavo, Hormiga and Petra Sierwald. Institutional support was pro- vided by Jonathan Coddington and the USNM. LITERATURE CITED Banks, N. 1907. A preliminary list of the Arachnida of Indiana, with keys to families and genera of spiders. Annual Report of the Department of Ge- ology and Natural Resources of Indiana 31:715- 745. Banks, N. 1910. Catalog of Nearctic Spiders. Bul- letin of the United States National Museum 72: I-III, 1-80. Banta, A.M, 1907. Araneida. Pp. 59-67. In, The Fauna of Mayfield’s Cave. Publications. Carne- gie Institution of Washington 67:1-117, pL 1. Barr, TC. 1961. Caves of Tennessee. State of Ten- nessee Department of Conservation and Com- merce, Division of Geology Bulletin 64:1-567. Barr, T.C. 1967. A new Pseudanophthalmus from an epigean environment in West Virginia (Cole- optera: Carabidae). Psyche 74:166-172. Barr, TC. & J.R. Holsinger. 1985. Speciation in cave faunas. Annual Review of Ecology and Sys- tematics 16:313-337. Berland, L. 1931. Campagne speologique de C. Bo- livar et R. Jeannel dans PAmerique du Nord (1928). Archives de Zoologie Experimentale et General 71:383-388. Bishop, S.C. 1925. Singing spiders. New York State Museum Bulletin 260:65-69. Bonnet, P. 1958. Bibliographia Araneorum. Volume 2. Part 4, N-S. Les Freres Douladoure, Toulouse. Bonnet, P. 1959, Bibliographia Araneorum. Volume 2. Part 5, T-Z. Les Freres Douladoure, Toulouse. Brignoli, RM. 1983. A Catalogue of the Araneae Described Between 1940 and 1981. Manchester University Press. Coddington, J.A, 1983. A temporary slide mount allowing precise manipulation of small struc- tures. Verhandlungen des naturwissenschaftlich- en Vereins in Hamburg 26:291-292, Coddington, J.A. 1989. Spinneret spigot morphol- ogy: evidence for the monophyly of orb- weaving spiders, Cyrtophorinae (Araneidae), and the group Theridiidae plus Nesticidae. Journal of Ar- achnology 17:71-91. Comstock, J.H. 1903. A Classification of North American Spiders. Comstock Publishing Com- pany, New York. Comstock, J.H. 1913. The Spider Book. Garden City, New York. Comstock, J.H. 1948. The Spider Book. Revised and Edited by W.J. Gertsch. Cornell University Press, Ithaca. Coyle, EA. & A.C. McGarity. 1991. Two new spe- cies of Nesticus spiders from the southern Ap- MILLER— PORRHOMMA CAVERNICOLA 437 palachians (Araneae, Nesticidae). Journal of Ar- achnology 19:161-168. Crosby, C.R. 1905. The spiders of the Rochport Cave, Mo. The Canadian Entomologist 37:367- 369. Culver, D.C. 1970. Analysis of simple cave com- munities. I. Caves as islands. Evolution 24:463- 474. Culver, D.C. 1971. Caves as archipelagoes. Bulletin of the National Speleological Society 33:97-100. Culver, D.C. 1982. Cave Life. Evolution and Ecol- ogy. Harvard University Press, Cambridge. Deeleman-Reinhold, C.L. 1978. Revision of the cave-dwelling and related spiders of the genus Troglohyphantes Joseph (Linyphiidae), with spe- cial reference to the Yugoslav species. Slovenska Akademija Znanosti in Umetnosti 23:1-220. Denis, J. 1949. Notes sur les erigonides. XVI. Essai sur la determination des femelles d’ erigonides. Bulletin de la Societe d’Histoire Naturelle de Toulouse 83:129-158. Elliott, ER. 1932. Revision of and additions to the list of Araneae (spiders) of Indiana. Proceedings of the Indiana Academy of Science 41:419-430. Emerton, J.H. 1875. Notes on spiders from caves in Kentucky, Virginia and Indiana. American Naturalist 9:278-281. Gertsch, W.J. 1984. The spider family Nesticidae (Araneae) in North America, Central America, and the West Indies. Texas Memorial Museum Bulletin 31:1-91. Gertsch, W.J. 1992. Distribution patterns and spe- ciation in North American cave spiders with a list of the troglobites and revision of the cicuri- ans of the subgenus Cicurella. Texas Memorial Museum, Speleological Monograph 3:75-122. Hedin, M.C. 1997a. Molecular phylogenetics at the population/species interface in cave spiders of the southern Appalachians (Araneae: Nesticidae: Nesticus). Molecular Biology and Evolution 14: 309-324. Hedin, M.C. 1997b. Speciational history in a di- verse clade of habitat- specialized spiders (Ara- neae: Nesticidae: Nesticus): Inferences from geo- graphic-based sampling. Evolution 16:1929- 1945. Hedin, M.C. & B. Dellinger. 2004. Descriptions of a new species and previously unknown males of Nesticus (Araneae, Nesticidae) from caves in eastern North America, with comments on spe- cies rarity. Zootara 904:1-19. Holm, A. 1944. Revision einiger norwegischer Spinnenarten und Bemerkungen iiber deren Vor- kommen in Schweden. Entomologisk Tidskrift. 65:122-134. Holm, A. 1979. A taxonomic study of European and east African species of the genera Pelecopsis and Trichopterna (Araneae, Linyphiidae), with descriptions of a new genus and two new species of Pelecopsis from Kenya. Zoologica Scripta 8: 255-278. Holsinger, J.R. & D.C. Culver. 1988. The inverte- brate cave fauna of Virginia and a part of eastern Tennessee: Zoogeography and ecology. Brimley- ana 14:1-162. Holsinger, J.R., R.A. Baroody & D.C. Culver. 1976. The invertebrate cave fauna of West Virginia. West Virginia Speleological Survey Bulletin 7: 1-82. Hormiga, G. 1999. Cephalothoracic sulci in liny- phiine spiders (Araneae, Linyphiidae, Linyphi- inae). Journal of Arachnology 27:94-102. Hormiga, G. 2000. Higher level phylogenetics of erigonine spiders (Araneae, Linyphiidae, Erigon- inae). Smithsonian Contributions to Zoology 609:1-160. International Commission on Zoological Nomen- clature. 1999. International Code of Zoological Nomenclature, fourth edition. International Trust for Zoological Nomenclature, London. Keyserling, E. 1886. Die Spinnen Amerikas: Ther- idiidae. Volume 2, part 2. Bauer & Raspe, Niirn- berg. Legendre, RR. 1963. L audition et Remission de sons chez les araneides. Annee Biologique 2: 371-390. Locket, G.H. & A.E Millidge. 1953. British Spi- ders. Volume 2. Ray Society, London. MacCook, H.C. 1890. American Spiders and their Spinning Work. Vol. 2. Published by the author. Academy of Natural Sciences of Philadelphia, Philadelphia. Marx, G. 1890. Catalogue of the described Araneae of temperate North America. Proceedings of the United States National Museum 12:497-594. Mclndoo, N.E. 1910. Biology of the Shawnee Cave spiders. Biological Bulletin 19:303-323. Mclndoo, N.E. 1911a. Notes on some arachnids from Ohio valley caves. Biological Bulletin 20: 183-186. Mclndoo, N.E. 191 lb. The lyriform organs and tac- tile hairs of araneids. Proceedings of the Acad- emy of Natural Sciences of Philadelphia 63:375- 418, pi. XXX-XXXIII. Merrett, P. 1963. The palpus of male spiders of the family Linyphiidae. Proceedings of the Zoolog- ical Society of London 140:347-467. Miller, J.A. 2005. Cave adaptation in the spider ge- nus Anthrobia (Araneae, Linyphiidae, Erigoni- nae). Zoologica Scripta. Miller, J.A. & G. Hormiga. 2004. Clade stability and the addition of data: A case study from eri- gonine spiders (Araneae: Linyphiidae, Erigoni- nae). Cladistics 20:385-442. Millidge, A.E 1977. The conformation of the male palpal organs of linyphiid spiders, and its appli- cation to the taxonomic and phylogenetic anal- ysis of the family (Araneae: Linyphiidae). Bui- 438 THE JOURNAL OF ARACHNOLOGY letin of the British Arachnological Society 4:1- 60. Millidge, A.F. 1984. The erigonine spiders of North America. Part 7. Miscellaneous genera (Araneae, Linyphiidae). Journal of Arachnology 12:121- 169. Packard, A.S. 1875. The invertebrate cave fauna of Kentucky and adjoining states. American Natu- ralist 9:274-278. Packard, A.S. 1888. The cave fauna of North Amer- ica, with remarks on the anatomy of the brain and origin of the blind species. Memoirs of the National Academy of Science 4:1-156. Peck, S.B. 1998. A summary of diversity and dis- tribution of the obligate cave-inhabiting faunas of the United States and Canada. Journal of Cave and Karst Studies 60:18-26. Petrunkevitch, A. 1911. A synonymic index-cata- logue of spiders of North, Central and South America with all adjacent islands, Greenland, Bermuda, West Indies, Terra del Fuego, Gala- pagos, etc. Bulletin of the American Museum of Natural History 29:1-791. Platnick, N. 2004. The World Spider Catalog. Ver- sion 5.0. Available online at: http:// research, amnh. org / entomology / spiders / catalog / index, html. American Museum of Natural History, New York. Reeves, W.K. 2000. Invertebrate cavernicoles of the Great Smoky Mountains National Park, USA. Journal of the Elisha Mitchell Scientific Society 116:334-343. Roberts, M.J. 1993. The Spiders of Great Britain and Ireland, Part 1, Text, Compact edition. Vol- ume 2: Linyphiidae and Check List. Harley Books, Martins, Essex. Roewer, C.F 1942. Katalog der Araneae von 1758 bis 1940. Volume 1. Natura, Bremen. Simon, E. 1884. Les arachnides de France. Vol 5. i Paris. Simon, E. 1894. Histoire Naturelle des Araignees. Deuxieme edition. Volume 1. Librairie Encyclo- pedique de Roret, Paris. Swofford, D.L. 2001. PAUP*. Phylogenetic Anal- i ysis Using Parsimony (*and Other Methods). , Version 4.0b 10 for 32-bit Microsoft Windows. Sinauer, Sunderland, MA. [ Thaler, K. 1975. The genus Opistoxys Simon — a « new synonym in linyphiid spiders. Bulletin of ! the British Arachnological Society 3:142-144. ; van Helsdingen, PJ. 1972. An account of money • spiders from down under (Araneida, Linyphi- idae). Zoologische Mededelingen (Leiden) 47: 369-390. Wiehle, H. 1956. Spinnentiere oder Arachnoidea (Araneae). 28. Familie Linyphiidae-Baldachin- spinnen. Tierwelt Deutschlands 44:i-viii, 1-337. Manuscript received 19 August 2004, revised 17 June 2005. 2005. The Journal of Arachnology 33:439-444 THE FOSSIL SPIDER FAMILY LAGONOMEGOPIDAE IN CRETACEOUS AMBERS WITH DESCRIPTIONS OF A NEW GENUS AND SPECIES FROM MYANMAR David Penney: Earth, Atmospheric and Environmental Sciences, The University of Manchester, Manchester, Ml 3 9PL, United Kingdom. E-mail: david.penney@ manche sten ac.uk ABSTRACT. The spider family Lagonomegopidae was described a decade ago from two specimens in Upper Cretaceous Siberian amber from the Taimyr Peninsula, and placed in the superfamily Palpimano- idea. Lagonomegopidae is known only from Cretaceous amber. Undiscovered extant species are considered unlikely because of their frequent occurrence in Cretaceous ambers and their absence in Tertiary fossil resins. One aim of this paper is to bring the existence of this family to the attention of neo-arachnologists. Burlagonomegops eskovi new genus and species is described from Cretaceous amber of Myanmar (Burma) and Lagonomegops americanus new species is assigned to a previously described, but unnamed specimen from Cretaceous New Jersey amber. Keywords: Burma, Mesozoic, paleontology, Palpimanoidea It is seldom the case that systematists work- ing on extant spiders acknowledge the exis- tence of fossil spiders in published papers on their particular group of interest. However, this is not universal and I am encouraged by the increased frequency with which reference to fossils now occurs. The 2U^ European Col- loquium of Arachnology, Russia 2003, hosted the first special symposium dedicated to pa- leoarachnology (see Logunov & Penney 2004), which was well attended. It is often true that fossil spiders preserved in shales and other sediments can be difficult, if not impos- sible to place in the framework of higher level extant spider taxonomy and systematics. However, this is not always the case with am- ber-preserved spiders. Marusik & Penney (2004) noted that fossil and Recent arachnol- ogical taxonomy cannot be considered as to- tally independent disciplines. The importance of considering fossils became evident when the fossil genus Archaea Koch & Berendt 1854, first described from Baltic amber (and placed in Archaeidae, a new family erected for the fossils) was shown to be a senior synonym of the extant genus Eriauchenius O. Pickard- Cambridge 1881 (originally placed in Theri- diidae) described from Madagascar by Simon (1895). More recently, the new name Theri- dion sulawesiense Marusik & Penney 2004 was erected for the extant spider species T. simplex Thorell 1877 from Sulawesi because that name was preoccupied by T. simplex Koch & Berendt 1854 from Baltic amber. Fossil spiders in Cenozoic ambers have been known for centuries. The first major work with formal descriptions appeared in the mid nineteenth century (Koch & Berendt 1854). In contrast, it was only a decade ago that the first spider inclusion in Mesozoic am- ber was described, by Eskov & Wunderlich (1995) of Santonian age from Siberia. How- ever, it is only within the last few years that new descriptions of Cretaceous amber spiders have been published, for example in fossil res- ins of Turonian age from New Jersey (Penney 2002, 2004fl), Barremian age from the Isle of Wight (Selden 2002), Upper Neocomian-bas- al Lower Aptian age from Lebanon (Penney & Selden 2002; Penney 2003a; Wunderlich & Milki 2004 [not 2001 as cited by Poinar & Milki 2001]), Albian age from Myanmar (Penney 2003b, 2004b) and Campanian age from Canada (Penney 2004c). Spiders have been listed as present (and occasionally fig- ured) in Mesozoic amber faunas from Canada (Me Alpine & Martin 1969), the Caucasus (Es- kov & Wunderlich 1995), France (Schliiter 1978; Neraudeau et al. 2002; Perrichot 2004), Alava, Spain (Alonso et al. 2000) and Astu- rias, Spain (Arbizu et al. 1999) but none of these have yet been formally described. 439 440 THE JOURNAL OF ARACHNOLOGY The enigmatic spider family Lagonomego- pidae was first described by Eskov & Wun- derlich (1995) from two specimens in Upper Cretaceous Siberian amber from the Taimyr Peninsula, and placed in the superfamily Pal- pimanoidea based on the presence of peg teeth, the absence of teeth on the cheliceral promargin, the trichobothrial pattern and the spineless legs. Penney (2002) described an ad- ditional specimen from New Jersey amber as Lagonomegops sp. indet. and Penney (2004c) described Grandoculus chemahawinensis Penney 2004 from Canadian amber. Wunder- lich (2004) provided the same figures and de- scriptions of the specimens originally de- scribed by Eskov & Wunderlich (1995). Platnick’s (2004) catalog did not include fossil taxa and the publications in which this fossil family is described may not be immediately obvious (or available) to some arachnologists, because one is a private journal published in Germany, two are paleontological and the fourth is a privately published book. The main aim of this paper is to bring to the attention of the arachnological community the existence of the enigmatic spider family Lagonomego- pidae, which is currently only known from amber, but which may have undiscovered ex- tant species in the southern hemisphere, as in the Archaeidae mentioned above. In addition, new specimens are described for the first time from Cretaceous amber of Myanmar (Burma). METHODS Material. — Two specimens preserved in Burmese amber (burmite) (for details of lo- cality and stratigraphy, see Zherikhin & Ross [2000], Grimaldi et al. [2002], Cruickshank & Ko [2003]) held in the Department of Ento- mology at the American Museum of Natural History (AMNH). AMNH Bu-707 is pre- served in a small piece (4X3X3 mm) of clear yellow-orange amber with no syninclu- sions, but with numerous small air bubbles; AMNH Bu-1353 is preserved in a small piece (9X5X5 mm) of clear yellow-orange am- ber containing several fracture planes and a male Diptera (Microphorinae) syninclusion. Methods. — Prior to being received by the author the amber had been set in a clear plas- tic resin and cut and polished to reveal the inclusions. All measurements were made us- ing an ocular graticule and are in mm. Draw- ings were done under incident light with a camera lucida attached to an Olympus SZH stereomicroscope and photographs were taken with a Nikon DIX digital camera attached to a Wild M8 stereomicroscope. Abbreviations used in the figures. — a = air bubble, ab = abdomen, car = carapace, L/ R 1-4 = left and right walking legs 1-4, p = pedipalp, s = spine, t = trichobothrium. SYSTEMATIC PALEONTOLOGY Remarks. — It is appreciated that fossil spi- ders are taxonomically subequal to the extant fauna (Eskov 1990) and the certainty with ! which pattern-based species can be recognized in the fossil record is less than that for extant organisms (Smith 1994). When I described the second known occurrence of the family La- gonomegopidae, from New Jersey amber (Penney 2002), I was reluctant to diagnose it as a species and refrained from naming it. However, given the recent discovery that this family represents a regular component of Cre- taceous faunas from several geographically distinct amber deposits, I feel it is now justi- fiable to place the specimens within a provi- sional taxonomic framework. Unfortunately , all specimens identified to date are immature. The genitalia are unknown for this family so the taxonomy is based on somatic characters. I Superfamily Palpimanoidea Remarks. — See Penney (2004c) for a dis- j cussion of the systematic placement of La- gonomegopidae in this superfamily. Family Lagonomegopidae Eskov & Wunderlich 1995 Distribution. — Fossil species in Creta- ceous ambers from Siberia, New Jersey, Myanmar and Canada. Recent species not known. ! Lagonomegops Eskov & Wunderlich 1995 ' Type species. — Lagonomegops sukatche- vae by original designation and monotypy. Holotype, juvenile, PIM 331 1/564, held in the Paleontological Institute of the Russian Acad- emy of Science, Moscow. Not examined be- cause the current location of these specimens within the PIM collections is unknown (K. Es- kov pers. comm. 2004). Distribution. — Fossil species in Creta- ceous ambers from Siberia and New Jersey. ' Recent species unknown. PENNEY— FOSSIL CRETACEOUS AMBER LAGONOMEGOPIDAE 441 Figures 1-4. — Burlagonomegops eskovi new species. Holotype, AMNH Bu-707, juvenile, Burmese amber. 1, 2. anterior view. 3, 4, dorsal view. 3-4, Scale lines = 0.5 mm Lagonomegops americanus new species Lagonomegops sp. indet: Penney 2002: 711, pL 1 fig. 2, text-fig. 2. Material examined.- — -Holotype juvenile, U.S.A.: New Jersey amber, 1995, K. Luzzi (AMNH NJ-556 (KL-297)). Diagnosis. — Lagonomegops americanus can be distinguished from L. sukatchevae by the possession of the following combination of characters: tarsi longer than metatarsi, a single dorsal spine distally on femur 1. Etymology.— The specific epithet is after America, the provenance of the fossil. Distribution and age.— New Jersey amber; Turonian, Upper Cretaceous (Grimaldi et al. 2000). Burlagonomegops new genus Type species. — Burlagonomegops eskovi new species. Etymology. — Bur derived from Burma, the former name of Myanmar, and lagonomegops, the type genus of the family. Diagnosis. — Burlagonomegops differs from the other genera in this family by having the carapace distinctly longer than wide and in possessing tarsal trichobothria. Description. — See description of the type species below. Distribution. — Fossil species in Creta- ceous amber from Myanmar. Recent species not known. Burlagonomegops eskovi new species Figs. 1-8 Lagonomegopidae: Grimaldi et al. 2002: 29, fig. 18e (AMNH BU-707). Material examined. — Holotype juvenile, Burmese amber, MYANMAR, Kachin: Tanai Village (on Ledo Road 105 km NW of Myit- kyna), 2000, by the Leeward Capitol Corpo- ration (AMNH Bu-707). Paratype: 1 juvenile, same data as holotype (AMNH Bu-1353). 442 THE JOURNAL OF ARACHNOLOGY Figures 5-8. — Burlagonomegops eskovi new genus and species. Paratype, AMNH Bu-1 353, juvenile, Burmese amber. 5, 6. anterior view. 7, 8. dorsal view. Scale lines = 0.5 mm. Etymology. — The specific epithet is a pa- tronym in honor of Dr. Kirill Eskov (Paleon- tological Institute, Moscow) in recognition of his contributions to paleoarachnology and his audible joy and excitement upon first viewing the paratype under a microscope. Diagnosis. — As for genus. Description (based on both holotype and paratype). — Body length 1.8; carapace 0.8 long, 0.5 wide between the eyes when viewed dorsally. With distinct, long setae, sides rounded in the thoracic region, cephalic region distinct and with a slightly procurved anterior edge (Figs. 3-4), lacking a fovea. Two large eyes, situated in flank positions anteriorly (Figs. 1-8). When viewed anteriorly, distance between clypeal margin and a hypothetical line joining these eyes at their centres 0.2; a second pair of smaller eyes are located mid- way between the large eyes and the end of the clypeal margin (Figs. 2, 6), width of clypeal margin 0.4, with long, curved setae projecting inwards from both sides. Chelicerae twice as long as wide, with long setae projecting downwards, not possible to determine whether peg-teeth are present or absent. Sternum 0.4 long, 0.3 wide between coxae 2, truncate an- teriorly and with sparse, long setae. Fang short, unmodified, labium as long as broad, maxillae longer than broad and converging. Opisthosoma oval (Figs. 3-4, 7-8), 1.0 long, 0.4 wide; spinnerets unmodified and in a com- pact group at the distal tip (Figs. 7-8). Leg formula unknown because neither specimen is preserved in a manner conducive to making accurate measurements, all seg- ments setose. Legs 1 and 2 appear approxi- mately equal in length, 2.0, leg 4 may be slightly longer and leg 3 is distinctly shortest. Leg spines thin and week, visible dorso-dis- tally on femora 1, 2 and 4 and the patellae of the pedipalp and legs 1, 2 and 3. Trichoboth- PENNEY— FOSSIL CRETACEOUS AMBER LAGONOMEGOPIDAE 443 ria: tibia 1 with paired (tibiae 2-4 with at least one), each metatarsus with one long in the dis- tal half and each tarsus with one long median and one short distal (FigSo 2, 6). Tarsi with three claws. Remarks. — Although both preserved in Burmese amber, each specimen appears to have undergone different diagenetic/tapho- nomic processes, to such an extent that at first sight they appear to be quite different from one another. The best preserved specimen is the holotype, the paratype seems to have un- dergone some somatic distortion in carapace shape anteriorly and in the legs, which appear thin, stretched and twisted. In addition, the majority of setae have not been preserved in the paratype. Distribution and age. — Burmese amber, Myanmar (Burma); Albian, Lower Cretaceous (Cruickshank & Ko 2003). DISCUSSION The known geological range of lagonome- gopids now spans approximately 25 Ma, from 100 Ma Burmese amber into the Campanian (Canadian amber; Penney 2004c). The youn- ger end of the known range is 75 Ma, shortly before the Cretaceous-Tertiary (K/T) bound- ary dated at 65 Ma, This boundary marks the mass extinction event that wiped out the di- nosaurs and numerous other groups. Spider in- clusions in Tertiary ambers are extremely common and the lack of Lagonomegopidae in these fossil resins, when considered against their frequent occurrence in Mesozoic resins, suggests they may have become extinct during this event, in contrast to many other spider families which survived it (Penney et al. 2003). However, undiscovered extant species of Lagonomegopidae may exist, as was sug- gested by Eskov & Wunderlich (1995), but their absence in Tertiary resins makes this un- likely. It is more probable, given the general habitus and frequent occurrence of lagonom- egopids in Cretaceous ambers that they oc- cupied a similar niche to the Recent Salticidae (the most species-rich family today), which are extremely frequent in Tertiary ambers but have not been described from the Cretaceous, Thus, the lagonomegopids may represent a primitive lineage which gave rise to the Sal- ticidae or they may have been ecologically re- placed by them. The discovery of mature la- gonomegopids with clearly visible genitalia should help resolve this problem and confirm or reject their superfamilial placement in Pal- pimanoidea. ACKNOWLEDGMENTS I thank D. Grimaldi of the American Mu- seum of Natural History, New York for pre- paring and providing the Burmese and New Jersey amber specimens for research purposes and P. A. Selden for his comments on the manuscript. The Royal Society is thanked for a conference travel grant, the Leverhulme Trust for research funding and the conference organizers for hosting an excellent congress. LITERATURE CITED Alonso, J., A. Arillo, E. Barron, J.C. Corral, J. Gri- malt, J.F. Lopez, R, Lopez, X. Martinez-Delclos, V. Ortuno, E. Penalver & P.R. Trincao. 2000. A new fossil resin with biological inclusions in Lower Cretaceous deposits from Alava (northern Spain, Basque-Cantabrian basin). Journal of Pa- leontology 74:158-178. Arbizu, M, E. Bernardez, E. Penalver, & M.A. Prie- to. 1999. El ambar de Asturias. Pp. 245-254. In Alonso, J., J. Corral & R. Lopez (eds) Proceed- ings of the world congress on amber inclusions. Estudios del Museo de Ciencias Naturales de Alava 14 (Numero Especial 2). Alava, Spain. Cambridge, O.-P. 1881. On some new genera and species of Araneidea. Proceedings of the Zoolog- ical Society of London 1881:765-775. Cruickshank, R. D. & K. Ko. 2003, Geology of an amber locality in the Hukawng Valley, Northern Myanmar. Journal of Asian Earth Sciences 21: 441-455. Eskov, K.Y. 1990. Spider palaeontology: present trends and future expectations. Acta Zoologica Fennica 190:123-127. Eskov, K.Y & J. Wunderlich. 1995 (for 1994). On the spiders from Taimyr ambers, Siberia, with the description of a new family and with general notes on the spiders from the Cretaceous resins. Beitrage zur Araneologie 4:95-107. Grimaldi, D., A. Shedrinsky & T. Wampler. 2000. A remarkable deposit of fossiliferous amber from the Upper Cretaceous (Turonian) of New Jersey. Pp. 1-76. In Grimaldi, D. (ed.) Studies on fossils in amber, with particular reference to the Creta- ceous of New Jersey. Leiden, Backhuys. Grimaldi, D., M.S. Engel & P. Nascimbene. 2002. Fossiliferous Cretaceous amber from Burma (Myanmar): Its rediscovery, biotic diversity, and paleontological significance, American Museum Novitates 336 1:1-71. Koch, C.L. & G.C. Berendt. 1854. Die im Bernstein befindlichen Crustaceen, Myriapoden, Arachni- 444 THE JOURNAL OF ARACHNOLOGY | den und Apteren der Vorwelt. Edwin Groening, Berlin. Logunov, D.V. & D. Penney (eds.). 2004. European Arachnology 2003. Proceedings of the 2U* Eu- ropean Colloquium of Arachnology (St. Peters- burg, Russia, August 2003). KMK Scientific Press, Moscow. Marusik, Y.M. & D. Penney. 2004. A survey of Baltic amber Theridiidae (Araneae) inclusions with descriptions of six new species. Pp.201- 218. In Logunov, D.V. & D. Penney (eds). Eu- ropean Arachnology 2003. Proceedings of the 21®^ European Colloquium of Arachnology (St, Petersburg, Russia, August 2003). KMK Scien- tific Press, Moscow. Mcalpine, J. F. & J.E.H. Martin. 1969. Canadian amber — a paleontological treasure chest. Cana- dian Entomologist 101:819-838. Neraudeau, D., V. Perrichot, J. Dejax, E. Masure, A. Nel, M. Philippe, P. Moreau, F. Guillocheau & T Guyot. 2002. A new fossil locality with insects in amber and plants (likely Uppermost Albian): Ar- chingeay (Charente-Maritime, France). Geobios 35:233-240. Penney, D. 2002. Spiders in Upper Cretaceous am- ber from New Jersey (Arthropoda, Araneae). Pa- laeontology 45:709-724. Penney, D. 2003a. A new deinopoid spider from Cretaceous Lebanese amber. Acta Palaeontolo- gica Polonica 48:569-574. Penney, D. 2003b. Afrarchaea grimaldii, a new species of Archaeidae (Araneae) in Cretaceous Burmese amber. Journal of Arachnology 31:122- 130. Penney, D. 2004a. New spiders in Upper Creta- ceous amber from New Jersey in the American Museum of Natural History (Arthropoda, Ara- neae). Palaeontology 47:367-375. Penney, D. 2004b. A new genus and species of Pi- sauridae (Araneae) in Cretaceous Burmese am- ber. Journal of Systematic Palaeontology 2:141- 145, pL4. Penney, D. 2004c. Cretaceous Canadian amber spi- der and the palpimanoidean nature of lagonom- egopids. Acta Palaeontologica Polonica 49:579- 584. Penney, D. & P.A. Selden. 2002. The oldest liny- phiid spider, in Lower Cretaceous Lebanese am- ber (Araneae, Linyphiidae, Linyphiinae). Journal ' of Arachnology 30:487-493. Penney, D., C.P. Wheater & P.A. Selden. 2003, Re- i sistance of spiders to Cretaceous-Tertiary extinc- [ tion events. Evolution 57:2599-2607. Perrichot, V. 2004. Early Cretaceous amber from south-western France: insight into the Mesozoic 1 litter fauna. Geologica Acta 2:9-22. j Platnick N.L 2004. The world spider catalog, ver- sion 4.5. American Museum of Natural History, i online at http://research.amnh.org/entomology/ ; spiders/catalogS 1-87/index. html. i Poinar, G.O. Jr, & R. Milki, 2001. Lebanese amber: 1 the oldest insect ecosystem in fossilized resin. I, Oregon State University Press, Corvallis. ' Schliiter, T. 1978. Zur Systematik und Palokologie ^ harzonservieter Arthropoda einer Taphozonose i aus dem Cenomanian von NW-Frankreich. Ber- ■ liner Geowissenschaftliche Abhandlungen (Se- ries A) 9:1-150. Selden, P.A. 2002. First British Mesozoic spider, from Cretaceous amber of the Isle of Wight, southern England, Palaeontology 45:973-984. Simon, E. 1895. Histoire naturelle des araignees, volume 1, part 4. Paris, pp 701-1084. Smith, A.B. 1994. Systematics and the fossil re- cord: documenting evolutionary patterns. Black- well Science, Oxford. Thorell, T. 1877. Studi sui Ragni Malesi e Papuani. I. Ragni di Selebes raccolti nel 1874 dal Dott, O. Beccari. Annali del Museo Civico di Storia Na- ! turali di Genova 10:341-637. Wunderlich, J. & R. Milki. 2004. Description of the : extinct new subfamily Microsegestriinae (Ara- | neae: Segestriidae) in Cretaceous Lebanese am- [ ber. Beitrage zur Araneologie 3b: 1867-1 873. Wunderlich, J. 2004, Fossil spiders in amber and i copal. Verlag J. Wunderlich, Hirschberg-Leuter- shausen. ! Zherikhin, V.V. & A.J. Ross. 2000. A review of the j history, geology and age of Burmese amber (Burmite). Bulletin of the Natural History Mu- f seum, London (Geology Series) 56:3-10. Manuscript received 20 August 2004, revised 31 March 2005. 2005. The Journal of Arachnology 33:445-455 THE GENERIC RELATIONSHIPS OF THE NEW ENDEMIC AUSTRALIAN ANT SPIDER GENUS NOTASTERON (ARANEAE, ZODARHDAE) B.C. Baehr: Queensland Museum, RO. Box 3300, South Brisbane Queensland 4101, Australia. E-mail: BarbaraB@qm.qld.gov.au ABSTRACT. A revision of the new endemic Australian genus Notasteron revealed two species, Notas- teron Carnarvon new species (male), Notasteron lawlessi new species (female, male). The genus is char- acterized by a strongly reticulated, shield-shaped sternum with steep lateral margins and a posteriorly situated boss. The male palp has a semicircular and undulated distal tegular apophysis and the female epigyne has long, convoluted copulatory ducts. Possible relationships of Notasteron with genera of the Asteron complex, Habronestes, Hetaerica, Malinella and Storosa, are analyzed with NONA and also reconstructed using the Hennigian method. The results indicate that the new genus does not belong to the Asteron complex but is the sister genus of Hetaerica. Notasteron lawlessi is quite common and occurs throughout the eastern part of Australia, whereas N. Western Australia. Keywords: Taxonomy, new species, cladistics The arachnid family Zodariidae is one of the most dominant ground-living spider fam- ilies in Australia (Churchill 1998). Most spe- cies can be easily recognized by their bright yellow or orange spots on a dark brown ab- domen and their annulated legs. With now 232 described and an estimated 350-400 total spe- cies, Australia has one of the richest known zodariid spider faunas worldwide. Rudy Jocque’s generic revision of the Zodariidae (1991) initiated intensive studies of the Aus- tralian zodariid fauna (Jocque 1991, 1995a, b; Jocque & Baehr 1992, 2001; Baehr & Jocque 1994, 1996, 2000). With funding from the Australian Biological Resources Study Partic- ipatory Program, 130 additional new species, including this revision, were described within the last three years (Baehr & Jocque 2001; Baehr 2003a, b, c, 2004a, b; Baehr & Chur- chill 2003). The two species of the new genus Notas- teron described here, were initially thought to belong to the Asteron complex because they share a similar abdominal pattern and the same general palp structure as the derived genera of the Asteron complex, Basasteron (Rainbow 1920), Cavasteron Baehr & Jocque 2000, Euasteron Baehr 2003, Holasteron Baehr 2004a, Masasteron Baehr 2004b, Min- asteron Baehr & Jocque 2000 and Spinasteron Carnarvon is only found in the Carnarvon region of Baehr & Churchill 2003. These genera are characterized by a thin semicircular embolus and an enormous semicircular distal tegular apophysis (DTA). The peculiar structure of the distal tegular apophysis is unique within the Australian zodariids but it occurs also in Tenedos O.P-Cambridge 1897 (Jocque & Baert 2002), a large South American genus. The use of a scanning electron microscope revealed significant differences in the struc- ture of the sternum, the labium, the endites and the coxae between Notasteron and the As- teron complex. The sternum of the species in the Asteron complex is flat or only slightly convex, shiny, finely reticulated and has a smooth, rebordered margin (Fig. 1 1). The spe- cies of Notasteron have a strong reticulated sternum with a weak boss posteriorly and a steep lateral margin (Fig. 12), Species of the Asteron complex have a triangular labium, but it is more rectangular with a narrow base and a broadly rounded tip in Notasteron. The en- dites within the Asteron complex are trian- gular and medially straight whereas the No- tasteron has medially concave endites. These characters are shared partly with Hetaerica Rainbow 1916 and Storosa Jocque 1991 (Figs. 13, 14). METHODS Descriptions are based on material stored in 70% ethanol. Epigynes were cleared in lactic 445 446 THE JOURNAL OF ARACHNOLOGY Figures 1-4. — Notasteron lawlessi. 1. body dorsal. 2, 4. carapace; 2. lateral; 4. frontal; 3. sternum and coxae. Scale bar = 1 mm. acid. Descriptions were generated with the aid of Intkey (Dallwitz et al. 1998) and shortened where possible. Location data and maps were created with Biolink version 1.5 (CSIRO En- tomology, Canberra, Australia; http://www. biolink.csiro.au/). Descriptions of spination and color patterns follow that in the revision of Euasteron (Baehr 2003a). Abbreviations of characters: ALE = anterior lateral eyes, AME = anterior median eyes, C = concavity on re- trolateral part of cymbium, CL/CW = cara- pace length / width, CD = copulatory duct, CO = copulatory opening, DTA = distal teg- ular apophysis (in previous papers called dor- sal tegular apophysis = conductor), DtiA = dorsolateral tibial apophysis, E = embolus, EB = embolus base, EP = external prong on dorso-retrolateral tibial apophysis, IP = inter- nal prong on dorso-retrolateral tibial apophy- sis, PE = prolateral extension of DTA, PLE BAEHR— NEW SPIDER GENUS NOTASTERON 447 Figures 5-10. — Notasteron spp., right male palps. 5, 7. ventral; 6, 8. lateral; 9, 10. epigyne. 9. ventral; 10. dorsal (cleared). 5, 6. N. Carnarvon-, 7-10. N. lawlessi. Scale bar = 0.5 mm (male palps), 0.1 mm (epigyne). Abbreviations: DTA = dorsal tegular apophysis; E = embolus; EP = external prong on dorso- retrolateral tibial apophysis (DtiA); IP = internal prong on dorso-retrolateral tibial apophysis (DtiA); PE = prolateral extension of DTA; PR = prong as ventral part of RDTA; RDTA = retrolateral extension of DTA; RE = retrolateral extension on cymbial flange; VtiA = ventral tibial apophysis. == posterior lateral eyes, PME = posterior me- dian eyes, PR = prong as ventral part of RDTA, RE = retrolateral extension on cym- bial flange, RDTA = retrolateral extension of DTA, S = spermatheca, SL/SW = sternum length/ width, VtiA = ventral tibial apophysis. Abbreviations of institutions from which material was borrowed: Australian Museum, Sydney (AM); American Museum of Natural History, New York (AMNH); Museum Vic- toria, Melbourne (MV); South Australian Mu- seum, Adelaide (SAM); Queensland Museum, Brisbane (QM); Western Australian Museum, Perth (WAM). SYSTEMATICS Family Zodariidae Thorell, 1881 Notasteron new genus Type species. — Notasteron lawlessi new species. Etymology. — The generic name reflects the fact that Notasteron is not a genus of the As- teron complex, and is considered neuter in gender. Diagnosis. — Species of Notasteron resem- ble those of Hetaerica in having a sternum with steep lateral margins and of Storosa in having a sternum with a posterior boss in males, but can be distinguished by the undu- lated prolateral part of DTA in the male palp (Figs. 5-8, 12) and the long convoluted epi- gynal ducts (Figs. 9, 10). Description. — Medium sized spiders (4.80-5.70) with oval, roughly reticulated, lat- erally rebordered carapace, widest between coxae II and III, narrowed in front to about 0.53 of maximum width. Profile flattened, with the highest point behind fovea (Fig. 2). Color of carapace, sternum and chelicerae or- ange to sepia brown; endites and labium sepia 448 THE JOURNAL OF ARACHNOLOGY brown, distally white. Abdomen sepia brown; dorsally with two pairs of white patches on top and one above the spinnerets; ventrally dark brown. Legs brown. Eyes (Figs. 1, 2, 4) in three rows (2-4-2). ALE in first row, AME, PLE in second, PME third row. AME smallest. Clypeus curved downwards, height about 2.3 times the diameter of ALE. Chilum single. Chelicerae with longitudinal boss and lateral condyle, few setae in front and a dense row of setae on distal promargin, with one tooth on promargin (Fig. 4). Endites broad, median margin concave with anteromesal scopula, no serrula, labium inverted u-shaped, basally constricted. Sternum shield-shaped with straight anterior margin, roughly reticu- lated and punctated, posteriorly with weak boss, lateral margin steep (Fig. 12). Legs with few spines on pairs I and II, more numerous on III and IV. Metatarsal preening brush on metatarsi II and III weakly developed. Paired tarsal claws with eight teeth on inner side, un- paired claw toothless, on onychium. Abdomen oval with two sigilla. Anterior lateral spinner- ets on common base, posterior median and posterior lateral spinnerets tiny, situated in one transverse row behind anterior spinnerets. Colulus represented by group of setae. Tra- cheal spiracle, tiny slit-like, covered by tiny sclerotized lip. Male palp (Figs. 5-8): Cymbium with dor- sal apical scopula, retrolaterally with straight, rectangular extension (RE). Cymbium base re- trolaterally with concavity C (Fig. 8). DTA semicircular, distal part folded containing em- bolus, with short undulated PE; RDTA with well developed tip and prong. Embolus base hidden behind RDTA, embolus thin, semicir- cular. Tibia: VtiA bipartite, internal prong long needle-shaped, external part flat, rebor- dered along lateral margin; DtiA bipatite, IP spatulate as long as EP (Figs. 6, 8). Epigyne (Figs. 9, 10): Epigyne with m- shaped copulatory openings, long convoluted copulatory ducts and small, globular sperma- thecae. Distribution. — The two known species of Notaste ran have a disjunct distribution. One species is quite common and occurs through- out the eastern part of Australia whereas the more derived species is only found in the Car- narvon region of Western Australia. Notasteron lawlessi new species (Figs. 1-4, 7-10, 12, 16) Type material. — Holotype male: AUS- TRALIA: Queensland: Taroom, “Boggo- moss” Station, 25°25'S, 150°0UE, 11 Novem- ber 1996, P. Lawless, pitfall (QM S37401). Paratypes: AUSTRALIA: Queensland: 1 male, Barakula State Forest, Hellhole Creek, open woodland, 26°20'S, 150°42'E, 13-15 October 2004, C. Burwell, pitfall (QM S67697); 1 male, Taroom District, BS24, 25°25'S, 149°58'E, 12 November 1996-Jan- uary 1997, P. Lawless, pitfall (QM S37214); 2 males, 2 females, Expedition Range Nation- al Park, ‘Amphitheatre’ yards, 440 m, 25°13'S, 149°01'E, 27 September 1997-4 March 1998, G. Monteith, D. Cook, pitfall (QM S44250, S44798); 1 female, Langlo Crossing, 3 km NW., 26°07'S, 145°39'E, 4 May 2001, G.B. Monteith, pyrethrum (QM S60615); 26 males. Lake Broadwater, via Dal- by, site 2, 5, 9, 27°2US, 15r06'E, 17 May 1985-25 February 1986, M. Bennie, pitfall (QM S47388-91, S47613-15); 2 males, same data (WAM T63078); 1 male, 1 female, Mount Gayndah, summit, 25°36'S, 151°32'E, 18 December 1998-27 January 1999, G. Monteith, pitfall (QM S55161); 1 male, 1 fe- male, Mount Gayndah, 25°35'S, 151°32'E, 16 November 2000, N. Platnick, hand collecting (AMNH); 2 males, 3 females. Mount Gayn- dah, 25°36'S, 151°32'E, 21 November 1998, R. Raven, vibration (QM S51313); 3 males, Mount Stuart, 23°05'S, 148°41'E, 12 Decem- ber 1999, D. Hannah, tree clearing, pitfall (QM S60741, S60743, S60744); 1 male, Mount Debatable, 1.5 km NE., 25°37'S, 151°34'E, 11 October-19 December 1998, G. Monteith, pitfall (QM S47569); 3 males, 1 fe- male, Mount Pleasant, site 32.2, 24°52'S, 146°23'E, 30 October 1999, D. Hannah, tree clearing, pitfall (QM S60742, S60754); 3 males, Thylungra, site 3, 26°05'S, 143°27'E, October 1995, T. Churchill, pitfall (QM S60752); 9 males, Fairview, 24°19'S, 147°01'E, 6 November 1998, D. Hannah, tree clearing, pitfall (QM S60749, S60745); 4 males, Narrien, 22°53'S, 146°49'E, 1998, D. Hannah, tree clearing, pitfall (QM S60747); 4 males, Oakleigh, 26°5US, 15r27'E, 1998, D. Hannah, tree clearing, pitfall (QM S60746); 1 male, Mulga gradient, pitfall traps 14-20, Site 5, October 1995, T. Churchill, pitfall (QM BAEHR— NEW SPIDER GENUS NOTASTERON 449 Figure 11—14. — Sterna and mouthparts. 11. Masasteron queeslandicum; 12. Notasteron lawlessi; 13. Hetaerica scenica; 14. Storosa obscura. S60751); 13 males, Meta Park, 1998, tree clearing, pitfall (QM S60748); 1 female, Fleurs, site58/2, February 1999, tree clearing, pitfall (QM S60750); 3 females, Texas, 16 km S., 28°56'S, 15r08'E, 25 January 2002, B. Baehr, N. Platnick, R. Raven, vibration (QM S60616); 1 female, Wycheproof, 23°38'S, 146°5UE, 1998, tree clearing, pitfall (QM S60753); New South Wales: 1 female, Kelvin SF, 8 km N. of Kelvin, 30°45'S, 150°20'E, 23 November-14 December 2001, H. Doherty, M. Elliot, pitfall (AM KS82173); 3 males, 2 females, Dowe SF, 30°47'S, 150°30'E, 23 No- vember-14 December 2001, L. Wilkie, H. Smith, pitfall (AM KS82166, KS82168, KS82 170-72); 1 male, 2 km from Tamworth on Tintinhull Rd, 31°04^S, 150°57'E, 15 No- vember-6 December 2001, H. Doherty, M. Elliot, pitfall (AM KS82167); 1 male, be- tween Kootingal and Tamworth, Crown Res. 200m past tip, 31°04'S, 15r02'E, 15 Novem- ber-6 December 2001, G, Carter, pitfall (AM KS82169); 1 male, Gubatta, 33°36'S, 146°3UE, 6-14 December 1999, D. Driscoll, pitfall (QM S53897); 1 male, Morton Plains Station, 1.5 km NE. of Enngonia, 29°05'S, 146°12'E, 15 October 1991, R. Harris (SAM KS32557); 1 female, Pulletop, strip site 3P, 450 THE JOURNAL OF ARACHNOLOGY 34°01'S, 146°04'E, 3-8 November 1999, D. Driscoll, pitfall (QM S53736); 1 male, Pulle- top, reserve, 33°58'S, 146°05'E, 3-8 Novem- ber 1999, D. Driscoll, pitfall (QM S53840); 9 males, 3 females, Pulletop, roadside, 34°01'S, 146°04'E, 12 October-8 November 1999, D. Driscoll, pitfall (QM S52523, S52642, S52732, S53778, S52919); 2 males, Pulletop site lOP, 33°55'S, 146°06'E, 12 October-8 November 1999, D. Driscoll, pitfall (QM S53279, S52817); 1 male, Rankins Springs, 33°45'S, 146°19'E, December 1999, D. Dris- coll, pitfall (QM S45819); 3 males. Round Hill Nature Reserve sitelR, 33°03'S, 146°13'E, 19-23 December 1999, D. Driscoll, pitfall (QM S52587); 2 males. Round Hill Na- ture Reserve, site 4R, 32°59'S, 146°05'E, 2 November-23 December 1999, D. Driscoll, pitfall (QM S52700, S531 15); 2 males. Round Hill Nature Reserve site 6R, 32°59'S, 146°03'E, 2-8 November 1999, D. Driscoll, pitfall (QM S52746); 3 males, Taleeban, 33°55'S, 146°28'E, 3-8 November 1999, D. Driscoll, pitfall (QM S53930, S52680); 1 male, 2 females, Taleeban site 4T, 33°57'S, 146°26'E, 23 February- 18 October 1999, D. Driscoll, pitfall (QM S53101, S53039, S52131); 4 males, 2 females, Taleeban, road- side, 33°52'S, 146°25'E, 12 October-10 No- vember 1999, D, Driscoll, Pitfall (QM S52650, S53127, S53128, S53531); 3 males, Taleeban, roadside site 8T, 33°53'S, 146°28'E, 1-10 November 1999, D. Driscoll, pitfall (QM S52661, S53753); 1 male, Taleeban sitelOT, 33°57'S, 146°24'E, 3-10 November 1999, D. Driscoll, pitfall (QM S53761); Vic- toria: 2 males, Meringur, 5 km ESE., site 1 14, 34°24'S, 141°23'E, November 1985, A.L.Yen, drift fence pitfall (MV); South Australia: 3 males, 8 females, Danggali Conservation Park, Sandford Dam, 33°22'S, 140°54'E, 22- 23 November 1996, D. Hirst, vibration (SAM NN 1739 1-401); 3 males, 6 females, Danggali Conservation Park, 3 km N. Tomahawk Dam, 33°19'S, 140°43'E, 24-26 November 1996, J.A. Forrest, D. Hirst (SAM NN 17402-09, NN 17411); 1 male, same locality, 4 Septem- ber 1996, D. Hirst (SAM NN17410); 1 male, 1 female, 8 km NNE. Mount Woodroffe, 26°15'S, 13r47'E, 13-17 October 1994, Pi- tjantjara Lands Survey, J.A. Forrest, pitfall (SAM NN 1 1435, NNl 1439); 1 male, Gluepot Res., 8.5 km W.-WNW. Gluepot Homestead, 33°44'S, 140°02'E, 26 November-6 Decem- ber 2000, Gluepot survey, Sitella Camp (SAM NNl 7390); 2 males, 12.5 km E. Mitchell Nob, 26°08'S, 131°57'E, 20-21 October 1994, J.A. Forrest, pitfall (SAM NNl 1436-7); Northern Territory: 1 male,l female, Illamurta Spring, 24°18'S, 132°4UE, 26 March 1993, D. Hirst, pitfall (SAM NNl 7388-89); 1 female, Dan- gali Conservation Park, 1.5 km S. 3LO Dam, i 33°17'S, 140°55'E March 2001, J.A. Forrest, D. Hirst, vibration (SAM NN17412). Etymology. — The specific name is a pa- tronym in honor of Phillip Lawless, formerly of the Queensland Museum, the collector of the holotype. Diagnosis. — This species can be distin- guished from the other species of the genus by the blunt tip of the RDTA in the male palp. Description. — Male (holotype): Total length 4.88. Cephalothorax 2.48 long; 1.80 wide; 0.80 high; cl/cw 1.37; sternum 1.20 long; 1.00 wide; sl/sw 1.20; abdomen 2.40 long; 1.52 wide. Color: body orange brown, to sepia brown, endites and labium distally white, abdomen dorsally with weak scutum and two pairs of white patches on top and with one above the spinnerets. Legs brown. Eyes: AME smallest; eye group width 0.50 of head- width; AME 0.10; ALE 0.12; PME 0.12; PEE 0.12; AME-AME 0.04; AME- ALE 0.04; PME-PME 0.02; PME-PLE 0.10; ALE-PLE 0.04; eyes group AME-PME 0.34; AME- AME 0.24; PME-PME 0.26. Clypeus 0.28 high. Male palp (Figs. 7, 8): RDTA with blunt tip and strong prong, equal in length. DtiA IP spatulate as long as EP, EP peg-shaped. Female (paratype): Total length 5.48. Cephalothorax 2.60 long; 1.64 wide; 0.96 high; cl/cw 1.58; sternum 1.20 long; 1.08 wide; sl/sw 1.11; abdomen 2.88 long; 1.68 wide. Coloration as male, but no scutum. Eyes: AME smallest; eye group width 0.45 of headwidth; AME 0.09; ALE 0.12; PME 0.12; | PLE 0.12; AME-AME 0.04; AME- ALE 0.04; PME-PME 0.04; PME-PLE 0.10; ALE-PLE ! 0.04; eyes group AME-PME 0.34; AME- i AME 0.22; PME-PME 0.28. Clypeus 0.28 high. Legs: female palpal claw strong with eight teeth. Epigyne (Figs. 9, 10): CO m- shaped, half way between epigastric fold and j end of epigyne. CD curled horizontally and vertically back to circular S. j Variation. — There is some variation in the | body color from dark brown to light orange ; BAEHR— NEW SPIDER GENUS NOTASTERON 451 Preferred tree 3 6L31Ci77Ri65 Figure 15, 16. — 15. Phylogeny of some genera related to Notasteron, based on character matrix in Table 2. Cladogram of supposed relationships in sense of Hennig (1966), consistent with NONA (fast optimi- zation). 16. Records of the genus Notasteron in Australia, rectangle N. lawlessi, circle N. Carnarvon. brown, but the color pattern of the abdomen is the same. Distribution. — This species occurs in South Australia, southern part of Northern Territory, New South Wales, Victoria and Queensland (Fig. 16). Notasteron Carnarvon new species (Figs. 6, 16) Type material. — Holotype male: AUS- TRALIA: Western Australia: Francois Peron National Park, W. of Monkey Mia along road 452 THE JOURNAL OF ARACHNOLOGY Table 1. — Characters and character states scored for the cladistic analysis. Character number Character Character state 0 AME size 0, smallest; 1, largest 1 Labium shape 0, triangular; 1, inverted u-shaped 2 Endites medial margin 0, straight; 1, concave 3 Sternum surface 0, smooth, shiny; 1, strongly reticulated, or/and punctated 4 Sternum profile 0, flat, rebordered; 1, elevated, with steep lateral margin (in males) 5 Sternum posteriorly 0, flat; 1 , with posterior boss (in males) 6 Palp DTA shape 0, short; 1 , cone-shaped, with enrolled lat- eral margin; 2, with stalk and enrolled tip; 3, semicircular with distinct RDTA and PE 7 Base of embolus 0, part of tegulum; 1, separated from teguh um as a chitinous plate 8 Base of embolus, shape 0, unmodified; 1, conical; 2, flattened; 3, hidden behind RDTA 9 Retrolateral cymbial flange (RE) 0, rectangular; 1, with rounded extension; 2, retrolateral concavity 10 TBE direction 0, pro-laterad; 1, baso-laterad; 2, retro-lat- erad 11 DTA, prolateral extension (PE) length 0, PE absent; 1, inside cymbium; 2, reach- ing tibia 12 Prolateral extension (PE) shape 0, PE absent; 1, about 14 of circle; 2, about 1/2 of circle; 3, about 14 of circle lateral margin undulated 13 VTA 0, present; 1, reduced to a tiny spine; 2, absent to Denham, 25°47'32"S, 113°4U37"E, 7 No- vember 1998, J.M. Waldock, vehicle vibration (WAM T54483). Paratypes: AUSTRALIA: Western Australia: Kennedy Range National Park, 24°31'25"S, 114°57'55"E, 14 January-7 April 1995, W. Muir, wet pitfall (WAM T54696); 3 males, Nerren Nerren Station, 27°03'S, 114°35'E, 11 January-11 May 1995, P. West et al., wet pitfall (WAM T44494); 1 male, same locality, 11 May-18 August 1995, N. Hall (WAM T54699); 1 male, same data (QM S67696); 1 male, Nerren Nerren Station, 27°03'24"S, 114°35'21"E, 25 August-16 Oc- tober 1994, J.M. Waldock et aL, wet pitfall (WAM T44496). Etymology. — The specific name is a noun in apposition taken from the region in which this species occurs. Diagnosis. — This species can be distin- guished from A. lawless! by the sharp tip of RDTA in the male palp. Description. — Male (holotype): Total length 5.60. Carapace 3.00 long; 2.20 wide; 1.08 high; cl/cw 1.36; sternum 1.36 long; 1.16 j wide; sl/sw 1.17; abdomen 2.60 long; 1.80 wide. Color: body sepia brown, endites and i labium distally white. Abdomen dorsally with two pairs of white patches on top and one above the spinnerets. Legs brown. Eyes: AME smallest; eye group width 0.44 of headwidth; AME 0.11; ALE 0.14; PME 0.14; PEE 0.14; AME-AME 0.04; AME-ALE 0.04; PME- PME 0.04; PME-PLE 0.12; ALE-PLE 0.04; eyes group AME-PME 0.40; AME-AME 0.26; PME-PME 0.32. Clypeus 0.32 high, i Male palp (Figs. 5, 6): RDTA with sharp tip and flattened prong. Embolus base flattened; ‘ DtiA IP spatulate as long as EP, EP bent ven- - trally. Female: Unknown. ' Distribution. — Found only in the Carnar- | von region of Western Australia (Fig. 16). PHYLOGENETIC ANALYSIS j As the primary tool for the phylogenetic analysis, I used the methods originally pro- ? posed by Hennig ( 1 966) and further explained BAEHR— NEW SPIDER GENUS NOTASTERON by Sudhaus & Rehfeld (1992, p.l37). Only homologous characters were considered where I was able to determine plesiomorphic and apomorphic character states. I followed the technique of Watrous & Wheeler (1981), using an outgroup, preferably a sister taxon, to determine the polarity of the character states. Based on these character states, I at- tempted to deduce the phylogenetic history of the species-groups. The result of this phylo- genetic analysis (seesu Heneig 1966) was tested with NONA version 2,0 (Goloboff 1997) using the heuristic search option and following settings: 5,000 random taxon addi- tion replications (mult*N), 5 starting trees per replication, and multiple tree-bissectioe-re- connection (TBR) branch swapping. The NONA bootstrap consensus tree was calculat- ed with 1,000 replications, 10 search replica- tions, and 5 starting trees per replication. I an- alyzed the same data set using unordered character states and fast optimization. Unsup- ported nodes were collapsed. The analysis is restricted to selected genera of the Asteron complex and the genera Habronestes, Hetaer- ica, Notasteron and Storosa. It is used here to define where Notasteron should be placed in the Zodariinae. The genera are represented by one species reflecting the “gruedplan” in the sense of Heneig (1966) for the considered character states. Yeates (1995) called this the exemplar method. The species Pentasteron simplex Baehr & Jocque 2001, Basasteron leucosemum (Rainbow 1920), Habronestes jocquei Baehr 2003, Storosa obscura Jocque 1991 and Hetaerica scenica (Koch 1872) were chosen for the relatively primitive male palpal morphology of the considered genera. Euasteron enterprise Baehr 2003 was added as a representative of the putatively derived genus of the Asteron complex to see how ro- bust the cladogram behaved. Maiinella zebra (Thorell 1881), the only known Australian species from a paleotropical zodariid genus, was selected first as the outgroup taxon, then replaced by Pentasteron simplex Baehr & Jocque 2001, the most basal representative of the analysis. Character assessment. — The sternum, la- bium and endites provide few distinguishing characters which appear to be informative at the genus and higher level. The eye pattern and particularly the male palps provide syea- pomorphic features of high value for phylo- 453 Table 2. — Character cladistic analysis. matrix for species used in Taxon 0-4 5-9 10-13 Pentasteron simplex 00000 00000 0000 Basasteron leucosemum 00000 03110 0112 Euasteron enterprise 00000 03121 1222 Habronestes jocquei 00000 02122 1000 Mallinella zebra 10000 00112 1000 Notasteron lawlessi 01111 13130 2132 Hetaerica scenica 01111 01000 0001 Storosa obscura OHIO 10000 0000 genetic examinations on species-group level. Derived characters of single species (autapo- morphies, e.g., epigynes) are not discussed here. Characters and their states are listed in Table 1. Sternum: The surface of the sternum is smooth and shiny (character 3/0), the profile is flat, the lateral margin is rebordered (char- acter 4/0) in all species of the Asteron com- plex and in Habronestes. In contrast, all ex- amined species of Notasteron, Hetaerica, and Storosa have a strongly reticulated and/or punctate sternum (character 3/1) with steep lateral margin in Notasteron and Hetaerica (character 4/1), and with a posterior boss in Notasteron and Storosa (character 5/1). Mouthparts: Within the Zodariinae, the shape of the labium and the endites are distin- guishing characters at the genus or higher lev- el. All species of the Asteron complex and Habronestes possess a triangular labium (character 1/0) and endites with a straight me- dial margin (character 2/0). The labium of He- taerica and Storosa is inverted u-shaped (character 1/1) and the endites of these genera as well as Notasteron are concave on the me- dial margin (character 2/1). Eyes: In all Zodariinae, both eye rows are so strongly procurved that they appear to be in three rows (2-4-2): ALE in the first row, AME and RLE in the second row, and the third row consists only of PME (Figs. 1, 2, 4). In all Notasteron, Hetaerica and Storosa spe- cies, the AME are smaller than the other eyes. This is also the case for the “grundplan” spe- cies Basasteron leucosemum, Pentasteron simplex, Habronestes jocquei and Euasteron enterprise (character 0/0), In Maiinella zebra, the AME are largest (character 0/1). The in- crease in size of the AME seems to be derived 454 THE JOURNAL OF ARACHNOLOGY but has presumably happened convergently quite frequently in different genera; e.g., in the genus Habronestes in the Habronestes mace- donensis group (Baehr 2003c) and at least four times in the “derived” genera of the As- teron complex (some species of Euasteron, about half of the species of Masasteron and almost all species of Spinasteron and Holas- ter on). Male palp: Most characters used in this phylogenetic study are taken from the male palp. Keeping in mind the primitive type of the male palp represented in Storosa obscura and Pentasteron simplex, the palps of Notas- teron appear quite derived. The most spectac- ular change happens in the undulating of the semicircular DTA (character 12/3) and that the embolus base is hidden under the RDTA (character 8/3). The most plesiomorphic short and membranous DTA and a short straight embolus occur in Storosa obscura and Pen- tasteron simplex (character 6/0) whereas a cone-shaped DTA with an enrolled margin is synapomorphic for all Hetaerica species (character 6/1). The slender stalk-like DTA with an enrolled distal tip is unique for all Habronestes species (character 6/2). The main synapomorphy for the genus Notasteron and the derived genera of the Asteron complex (Basateron, Cavasteron, Euasteron, Holaster- on, Minasteron, Spinasteron, Tropasteron and Masasteron) is the large semicircular DTA with marginal fold, well developed retrolater- al-(RDTA) and prolateral extension (PE) (character 6/3). As this kind of DTA is unique in the Australian Zodariinae, it could be thought that the above-mentioned derived genera of the Asteron complex and Notasteron are monophyletic. The base of the embolus is separated from the tegulum in the genera No- tasteron and Habronestes, and in all derived genera of the Asteron complex. Whereas in Habronestes and in the Asteron complex, the embolus base is uncovered it is hidden in No- tasteron (characters 7/1, 8/3). The position of the transbasal area of the embolus (TBE) pro- vides a derived character state for the genus Notasteron (character 10/2). The retrolateral cymbial flange (RE) is rectangular and straight in the basal condition (character 9/0) as for Notasteron, Storosa, Hetaerica, Pentas- teron, Basasteron but has a deep groove (character 8/2) as a synapomorphy for all Ha- bronestes species. In the derived genera of the Asteron complex the flange consists of a rounded extension (character 8/1). Results. — The cladistic analysis of the data matrix (Table 2) with NONA including 14 characters and seven taxa resulted in three most parsimonious trees (length 26 steps, Cl = 80, RI = 68). Euasteron enterprise was added to the data matrix representing the more derived genera of the Asteron eomplex. In this case, six most parsimonious trees were found (length 31 steps. Cl = 77, RI = 65). In all trees, Notasteron was matched with only three of eight taxa. The pair Notasteron-Basasteron is based exclusively on the advanced male palp character states as synapomorphies. The sister-group constellation Notasteron-Storosa is based on the sternum with posterior boss taken as a synapomorphy (5/1). From the analyses; phylogenetic analysis in the sense of Hennig, the cladistic analysis with 7 taxa and 8 taxa, and the bootstrap (“majority rules”), the congruent tree in all was taken as the pre- ferred tree (Fig. 15). The resulting tree shows that Notasteron is the sister genus of Hetaer- ica considering the somatic characters of the sternum, the endites and the labium as syna- pomorphic character states of the genera Sto- rosa, Hetaerica and Notasteron, and the steep lateral margin as the main synapomorphic character state for Hetaerica and Notasteron. Biogeography. — The Zodariidae are ground dwelling spiders that are not known to disperse aerially. The genus Notasteron in- cludes only two species to date. Both species were found in a disjunct distribution pattern. The basal species, N. lawlessi, is quite com- mon in the southern and eastern part of Aus- tralia (Fig. 16), like the most plesiomorphic species of the Asteron complex {Pentasteron simplex, Basasteron leucosemum and Euaster- on enterprise)', whereas the derived species, Notasteron Carnarvon, is only found in the Carnarvon region of Western Australia. ACKNOWLEDGMENTS This paper is dedicated to Valerie Davies for an excellent working atmosphere through the last years, as she is no longer working on spiders. I thank Mark Harvey and Julianne Waldock (Western Australian Museum, Perth), David Hirst (South Australian Muse- um, Adelaide), Graham Milledge (Australian Museum, Sydney), Robert Raven (Queensland Museum, Brisbane) and Peter Lillywhite BAEHR— NEW SPIDER GENUS NOTASTERON 455 (Melbourne Museum, Victoria) for loan of the material and great support of the work, I am most grateful to the following for providing critical comments on the manuscript: Robert Raven (Brisbane) and Volker Framenao (Perth), I also thank Robert Raven, who sup- ported me with the scanning electron micro- scope, I would like to thank my children, Jo- hanna and Ursula, for their patience. LITERATURE CITED Baehr, 2003a. Revisions of the new endemic genera Basasteron, Euasteron and Spinasteron of Australia (Araneae, Zodariidae): Three new gen- era of the Asteron-comple.'K, Memoirs of the Queensland Museum 49:1-27. Baehr, B. 2003b. Revision of the tropical genus Tropasteron gen. nov. of North Queensland (Ar- aneae, Zodariidae): A new genus of the Asteron- complex. Memoirs of the Queensland Museum 49:29-64. Baehr, B. 2003c. Revision of the Australian genus Habronestes L. Koch (1872) (Araneae: Zodari- idae). The species of NSW. Records of the Aus- tralian Museum 55:343-367. Baehr, B. 2004a. Revision of the new Australian genus Holasteron (Araneae, Zodariidae): taxon- omy, phylogeny and biogeography. Memoirs of the Queensland Museum 49:495-519. Baehr, B. 2004b. The systematics of a new endemic Australian genus of ant spiders Masasteron (Ar- aneae: Zodariidae). Invertebrate Systematics 18: 661-691. Baehr, B. & TB. Churchill. 2003. Revision of the endemic Australian genus Spinasteron (Araneae, Zodariidae): taxonomy, phylogeny and biogeog- raphy. Invertebrate Systematics 17:641-665. Baehr, B. & R. Jocque. 1994. Phylogeny and zoo- geography of the Australian genus Storena (Ar- aneae, Zodariidae). Spixiana 17:1-12. Baehr, B. & R. Jocque, 1996. A revision of Asteron, starring male palpal morphology (Araneae, Zo- dariidae). Proceedings of the XIII International Congress of Arachnology, Geneva, 3-8 Septem- ber 1995. Revue Suisse de Zoologie, hors serie 1:15-28. Baehr, B. & R. Jocque. 2000. Revisions of the gen- era in the A^tero/i-complex (Araneae, Zodari- idae). The new genera Cavasteron and M inaster- on. Records of the Western Australian Museum 20:1-30. Baehr, B. & R. Jocque. 2001. Revisions of the gen- era in the Asteron-compl&x (Araneae, Zodari- idae). The new genera Pentasteron, Phenasteron, Leptasteron and Subasteron. Memoirs of the Queensland Museum 46:359-385. Churchill, TB. 1998. Spiders as ecological indica- tors in the Australian tropics: family distribution patterns along rainfall and grazing gradients. Bulletin of the British Arachnological Society 11:325-330. Dallwitz, M.J., T.A. Paine & E.J. Zurcher. 1998. Interactive keys. Pp, 201-212. In Information Technology, Plant Pathology and Biodiversity. (Eds P. Bridge, P. Jeffries, D. R. Morse, and P. R. Scott.) CAB International, Wallingford. Goloboff, P.A. 1997. NONA version 2.0. Depart- ment of Entomology, American Museum of Nat- ural History, New York. Hennig, W. 1966. Phylogenetic Systematics. Uni- versity of Illinois Press, Urbana, Jocque, R. 1991. A generic revision of the spider family Zodariidae (Araneae). Bulletin of the American Museum of Natural History 201:1- 160. Jocque, R. 1995a. Notes on Australian Zodariidae (Araneae). I. New taxa and key to the genera. Records of the Australian Museum 47:117-140. Jocque, R. 1995b. Notes on Australian Zodariidae (Araneae). II. Redescriptions and new records. Records of the Australian Museum 47:141-160. Jocque, R. & B. Baehr. 1992. A revision of the Australian spider genus Storena (Araneae, Zo- dariidae). Invertebrate Taxonomy 6:953-1004. Jocque, R. & B. Baehr. 2001. Revisions of the gen- era in the A^-terow-complex (Araneae, Zodari- idae): Asteron Jocque and the new genus Pseu- dasteron. Records of the Australian Museum 53: 21-36. Jocque, R. & L. Baert. 2002. A revision of the Neo- tropical spider genera Tenedos O.P.-Cambridge and Ishania Chamberlin (Araneae, Zodariidae). Bulletin de Flnstitut Royal des Sciences Natu- relies de Belgique, Entomologie 72:67-173. Sudhaus, W. & K. Rehfeld. 1992, Einfuehrueg in die Phylogenetik und Systematik. Gustav Fischer Verlag, Stuttgart, Jena, New York. Watrous, L.E. & Q.D. Wheeler. 1981. The out- group comparison method of character analysis. Systematic. Zoology 30:1-11. Yeates, D.K. 1995. Groundplans and exemplars: path to the tree of life. Cladistics 11:343-357. Manuscript received 23 August 2004, revised 31 March 2005. 2005. The Journal of Arachnology 33:456-467 TARSAL SCOPULA SIGNIFICANCE IN ISCHNOCOLINAE PHYLOGENETICS (ARANEAE, MYGALOMORPHAE, THERAPHOSIDAE) Jose Paulo Leite Guadanucci: Museu de Zoologia da Universidade de Sao Paulo, Institute de Biociencias da Universidade de Sao Paulo, Av. Nazare, 481, Ipiranga, CEP: 04263-000 Sao Paulo, SP — Brazil. E-mail: zepaulo@artist.com.br ABSTRACT. Tarsal scopula condition and carapace length were studied for eighteen Ischnocolinae species. For cladistic analysis a matrix of 20 terminals and 30 characters of representatives of Ischnocol- inae, Theraphosinae, Aviculariinae, Harpactirinae and Trichopelmatinae were analyzed using Nona 2.0 computer software. The matrix was analyzed in four different ways: 1. each tarsal scopula (legs I-IV) coded as separate characters; 2. one character with six ordered states; 3. one character with six independent states; 4. without tarsal scopula character. The first two matrices result in one tree with the same indices (L = 72; Cl = 0.54; RI = 0,74) and topology: Part of Ischnocolinae is monophyletic {H. rondoni(S. longibulbi(l. algericus+Catumiri))) and the other representatives (Oligoxystre and Genus 1) form a distinct monophyletic group with Theraphosinae, Harpactirinae and Aviculariinae. There are no homoplasies in tarsal scopula evolution in the second cladogram. The other two cladograms show less resolution for the Ischnocolinae than the two first cladorams. The tarsal scopula condition appears to have no relation to spider size (t =—0.80433; P — 0.438247) and should be used in phylogenetic analysis of Ischnocolinae because it provides information on the character variability within the subfamily. Keywords: Phylogeny, South America, cladistics The condition of the tarsal scopula has had an important role in the systematics of the Is- chnocolinae Simon 1892. The scopula shows ontogenetic differentiation, being divided in all juvenile Theraphosidae and becoming en- tire in adults of some groups (Pocock 1897; Gerschman de Pikelin & Schiapelli 1973; Perez-Miles, 1994). The condition of the tar- sal scopula has been considered a good taxo- nomic tool and has already been used to di- agnose genera and species groups in Theraphosidae. Its use in phylogenetics is questionable since, within the Theraphosinae, the presence of a divided scopula is related to small sized species (Perez-Miles 1994). Char- acterized as theraphosids with a divided tarsal scopula (plesiomorphic state), Ischnocolinae is considered a paraphyletic group. Ischnocol- inae is the subfamily of Theraphosidae Tho- rell 1869 that shows the broadest geographic distribution, with species occurring in north- ern, central and eastern Africa, Seychelles, the Middle-East, the Mediterranean region, cen- tral and south Americas and the Antilles (Smith 1990; Rudloff 1997; Vol 2001). Con- sidering that Ischnocolinae was proposed as a subfamily based on a plesiomorphic character state (divided tarsal scopula), the situation of the group’s systematics is very confusing. Ra- ven (1985) considered Ischnocolinae a para- phyletic group that should have been revised at the generic level and grouped into mono- phyletic units. However, since the description of the type-genus Ischnocolus Ausserer 1871, only a few genera have been revised. Gersch- man de Pikelin & Schiapelli (1973) revised the subfamily as a whole but of the 42 genera included in this study, only 10 are in fact Is- chnocolinae representatives. The remaining 12 were subsequently synonymyzed or trans- ferred to other families and subfamilies. Rud- loff (1997) revised the genus Holothele Karsch 1879 but did not present a diagnosis of the genus and its species; the identification key does not include all the species and the structures are poorly illustrated. Smith (1990) published a taxonomic revision of European and African Ischnocolinae and presented de- scriptions and diagnoses for all genera. For over a century, the tarsal scopula state ‘'divided by a longitudinal band of setae” was used in Theraphosidae taxonomy (Perez-Miles 456 GUADANUCCI— TARSAL SCOPULA SIGNIFICANCE 457 1994). Gerschmae de Pikelin & Schiapelli (1973), following Ausserer (1871), considered the tarsal scopula condition an important tax- onomic character and stated that the divided tarsal scopula is present in all juvenile thera- phosids. Juvenile Theraphosinae Thorell 1870 have divided tarsal scopulae that become en- tire in the adult stage, in Ischnocolinae the scopulae remain divided into adulthood (Po- cock 1897; Gerschman de Pikelin & Schia- pelli 1973; Perez-Miles 1994). Although this ontogenetic differentiation was detected, the divided condition continued to be used caus- ing the inclusion of juvenile Theraphosinae within Ischnocolinae. This problem remained unresolved until Raven (1985) considered Is- chnocolinae a paraphyletic group that presents poorly developed tarsal scopulae. The tarsal scopula as a phylogenetic character was used for the first time by Perez-Miles (1992) in a preliminary cladistic analysis of the subfamily Theraphosinae. In this paper he shows that the entire tarsal scopula is synapomorphic for some genera of this subfamily. Later, Perez- Miles (1994) discussed the value of the tarsal scopula in Theraphosinae systematics and concluded that the scopula condition is related to spider size, although some exceptions exist. Considering that the role of the tarsal scopula in Theraphosidae systematic remains obscure, the goal of this study is to vary the use of the character “tarsal scopula” in a phylogenetic analysis for Ischnocolinae and discuss the re- sults. METHODS The material examined belongs to the fol- lowing institutions: lestituto Butantan, Sao Paulo (IBSP); American Museum of Natural History, New York (AMNH); Museo Argen- tine de Ciencias Naturales Bernardino Riva- davia, Buenos Aires (MACN); Museo de la Plata, La Plata (MLP); Museu de Zoologia, Universidade de Sao Paulo, Sao Paulo (MZSP); Zoological Museum University of Copenhagen (ZMUC); Moseu Paraense Emi- lio Goeldi, Belem (MPEG). Tarsal scopula condition was observed un- der a stereomicroscope. Following Perez- Miles (1994), a few isolated long and thin hairs in the tarsal scopula were not considered divided. Carapace length was used to estimate spider size. Below is a list of Ischnocolinae species used for tarsal scopulae condition and carapace length: Holothele rondoni (Lucas & Bticherl 1972): 1 d, Apiacas, Mato Grosso, Brazil (MZSP 18046); 1 d, Apiacas, Mato Grosso, Brazil (MZSP 18038); 1 d, Manicore, Amazonas, Brazil (MZSP 18990). Sickius longibulbi Soares & Camargo 1948: 4 d, Itirapina, Sao Paulo, Brazil (MZSP 22756). Genus 1: 3 d, Colinas do Sul, Serra da Mesa, Goias, Brazil (MZSP 18992). Catumiri petropolium Guadanucci 2004: 1 d, Pe- tropolis, Rio de Janeiro, Brazil (IBSP 8596). Catumiri chicaoi Guadanucci 2004: 1 d, Una, Ba- hia, Brazil (IBSP 9514). Catumiri uruguayense Guadanucci 2004: 1 d, Lav- alleja, Aguas Blancas, Uruguay (IBSP 9491). Catumiri argentinenese (Mello-Leitao 1941): 1 d, Jujuy, Yuto, El Pantanoso, Argentina (MACN 6424). Genus 2, sp. 1:3 d, laragua, Goias, Brazil (MPEG 1677) Genus 2, sp. 2: 1 d, Serra Norte, Para, Brazil (MPEG 1678). Genus 2, sp. 3: 1 d, Ilha Marajo Breves, Para, Bra- zil (MPEG 1679). Genus 2, sp. 4: 1 d (IBSP 11083), 1 d (IBSP 11086), 1 d (IBSP 1 1087), Pimenta Bueno, Ron- donia, Brazil; 1 d, Mineiros, Goias, Brazil (IBSP 8070). Genus 2, sp. 5: 1 d, Linhares, Espirito Santo, Brazil (IBSP 8654); 1 d, Linhares, Espirito Santo, Bra- zil (IBSP 7987); 1 d, Porto Seguro, Bahia, Brazil (IBSP 11084); 1 d, Ilheus, Bahia, Brazil (IBSP 11085). Oligoxystre new species 1: 1 d (IBSP 9488), 1 d (IBSP 9489), 1 d (IBSP 9486), 1 d (IBSP 9484), Central, Bahia, Brazil. Cladistic analysis. — The cladistic analysis was carried out using Nona version 2.0 (Go- loboff 1993). Search strategy was mult* with 100 replications. The data matrix included 20 terminal taxa and 30 characters and was con- structed with NDE (Nexus Data Editor) ver- sion 0.5.0 (Page 2001). The out-group was chosen based on the phylogenetic relation- ships of Mygalomorphae presented by Golo- boff (1993). In order to avoid an excess of missing entries, we preferred used a Bary- chelidae Simon 1889 rather than a Paratropi- didae Simon 1889 as the out group, since the last family presents some incomparable char- acters with Theraphosidae. Character polarity was read straight from the preferred dado- gram following Nixon & Carpenter (1993). Below is a list of species used in the cla- distic analysis: 458 THE JOURNAL OF ARACHNOLOGY Reichlingia annae Holothele rondoni Sickius longibuibi Ischnocolus algaricus Catumiri petropolium Catumiri chicaoi Catumiri uruguayensa Catumiri argentinensa Genus 1 Tapinauchenius sp Pterinochiius murinus Avicularia avicularia Euathlus vulpinus Vitalius vellutinus Oiigoxystre sp4 Oligoxystre sp5 Oiigoxystre sp1 Oligoxystre sp2 Oiigoxystre sp3 Barychelidae Figures 1 72; Cl - 0 72; Cl - 0. -2.- .54; 54; Reichlingia annae Holothele rondoni Sickius longibuibi Ischnocolus algericus Catumiri petropolium Catumiri chicaoi Catumiri uruguayense Catumiri argentinense Genus 1 Tapinauchenius sp Pterinochiius murinus Avicularia avicularia Euathlus vulpinus Vitalius vellutinus Oligoxystre sp4 Oligoxystre sp5 Oligoxystre spl Oligoxystre spl Oligoxystre sp3 —Relationship hypothesis between Ischnocolinae and other Theraphosidae groups. 1 . (L = RI = 0.74). Tarsal scopula coded in four characters (22-25), one for each leg. 2. (L = RI = 0.74). Tarsal scopula coded as one character with six ordered states. Table 1. — Matrix composed of 20 terminals and 31 characters. GUADANUCCI— TARSAL SCOPULA SIGNIFICANCE 30 o o o o o O CO o o CM VO VO VO VO ITi in in in m 29 o - - - - 28 o o T=j - - - 27 o o o o o o o o o o o o o o o o o o 26 o o o o o CM o o o o 25 o o o o o o o o o o o CM CM CM CM - 24 o o o o o o o o o o 23 o o o o o O CM o o CM CM CM CM CM CM CM CM CM CM 22 o o o o o o — < o o ^=H l=i ^==( o o o o o o o o o o o CM o o o o o o o 20 o o o o o o o o o o o - o\ o o o o o O o o o o o o o o o o o w o o o o o o o o o o o o o o o o O' 1 1 o 1 o O 1 o 1 1 1 o 1 ! 1 1 1 VO o o 04 o m rn fO cn o CM o o o CM o o o o o IT) o o o o o o - CM o o o o o o CM CM CM o o CM CM CM CM CM - - - m o o o o o o o o o o o CM o o o 1 o O C'- o o o o o o 7=H o o o o o o o 1 O e-- o o o o o o o o o o o o o o o o o o o o o o o o o o OV o o o - o o o M o o o o o o o o o o o o o o o o o o O' o o o o o o o o o o o o o o o o o o VO o o o o o o o o o o o o o o o o o o - IT) o o o o o o o o o o o o o o o o o o o o o o o 1 o 1 o o o o o o o o o o o o CO o o o o o o o o o o C\1 o o o o 1 - 1 o o o o o o o o o o o o 1 “ 1 o 7=^ o 1 o o o o o O '-H o o o o o CM o o o o o o o « ^ § - K d.2 e-^a^ os ^ ;5:l taf.lgE 11 io cO X ^ W3 d a «3 s ^3 i o g .« ”li N 0. s 3| e ^ s I ^ IT) C^ CO S S p^ a d( di Oh ^ M M M Cfl M “ I s s s s s s ^33 3 3 3 s s ^ s ^ -> o o o o o Q ^ -S’ c3 § 3 o o o S 459 460 THE JOURNAL OF ARACHNOLOGY BARYCHELIDAE: Reichlingia annae (Reichling 1997): 1 (5', 1 9, New River Lagoon, Orange Walk, Belize (AMNH). THERAPHOSIDAE: Avicularia avicularia (Lin- naeus 1758) (Aviculariinae): 1 S, Jacare, Rio Trombetas, Oriximina, Para, Brazil (MZSP 5687); 1 9, Jacare, Rio Trombetas, Oriximina, Par4 Brazil (MZSP 5687). Eiiathlus vulpinus (Karsch 1880) (Theraphosinae): 5 c?, Osorno, Chile (IBSP 3817-A); 4 9, Osorno, Chile (IBSP 3817-B). Catumiri petropolium Guadanucci 2004 (Ischno- colinae): 1 6, Petropolis, Rio de Janeiro, Brazil (IBSP 8596); 1 6, Petropolis, Rio de Janeiro, Brazil (IBSP 8606). Catumiri chicaoi Guadanucci 2004 (Ischnocolinae): 1 (?, Una, Bahia, Brazil (IBSP 9514); 1 9, Una, Bahia, Brazil (IBSP 9514). Catumiri uruguayense Guadanucci 2004 (Ischno- colinae): 1 (?, Lavalleja, Aguas Blancas, Uruguay (IBSP 9491); 1 9, Lavalleja, Aguas Blancas, Uruguay (IBSP 9507). Catumiri argentinenese (Mello-Leitao 1941) (Is- chnocolinae): 1 6 , Jujuy, Yuto, El Pantanoso, Ar- gentina (MACN 6424); 1 9, Catamarca, Argen- tina (MLP 14608). Genus 1 (Ischnocolinae): 1 6, Fazenda Sandoval, Porto Nacional, Tocantins, Brazil (IBSP 8585); 1 9 , Fazenda Sandoval, Porto Nacional, Tocantins, Brazil (IBSP). Holothele rondoni (Lucas & Biicherl 1972) (Is- chnocolinae): 1 (5, Apiacas, Mato Grosso, Brazil (MZSP 18046); 1 9, Tucurui, Para, Brazil (IBSP). Ischnocolus algericus Thorell 1875 (Ischnocoli- nae): 1 (3, 1 9, El Araish, Marocco (ZMUC 620, 628). OUgoxystre new species 1 (Ischnocolinae): 1 c3, Central, Bahia, Brazil (IBSP 9487); 1 9, Toca da Esperanga, Jussara, Bahia, Brazil (IBSP 8549). OUgoxystre new species 2 (Ischnocolinae): 1 6', Chapada dos Guimaraes, Mato Grosso, Brazil (IBSP 9495); 1 9, Chapada dos Guimaraes, Mato Grosso, Brazil (IBSP 9504). OUgoxystre new species 3 (Ischnocolinae): 1 6, Sao Domingos, Goias, Brazil (IBSP 8625); 1 9, Serra da Mesa, Mina§u, Goias, Brazil (IBSP 9467). OUgoxystre new species 4 (Ischnocolinae): 1 c3, Tucuriu, Para, Brazil (IBSP 9459); 1 9 , Tucurui, Para, Brazil (IBSP 7936). OUgoxystre new species 5 (Ischnocolinae): 1 9, Toca da Esperan^a, Central, Bahia, Brazil (IBSP 8553). Pterinochilus murinus Pocock 1897 (Harpactiri- nae): 1 c3, Africa (IBSP); 1 9, Kenya (IBSP). Sickius longibulbi Soares & Camargo 1948 (Is- chnocolinae): 1 d, Parnafba, Mato Grosso do Sul, Brazil (IBSP 8019); 1 9, Votuporanga, Sao Pau- lo, Brazil (IBSP 8693). Tapinauchenius sp. (Aviculariinae): 1 6, Tucurui, Para, Brazil (IBSP 4925-A); 1 9, Rio Marupi, Para, Brazil (IBSP 4676). Vitalius vellutinus (Mello-Leitao 1923) (Theraphos- inae): 1 (3, Porto Cabral, Rio Parana, Teodoro Samapaio, Sao Paulo, Brazil (MZSP 14953); 1 9, Teodoro Samapaio, Porto Cabral, Rio Parana, Teodoro Sampaio, Sao Paulo, Brazil (MZSP 3150). CLADISTIC ANALYSIS Below is a list of the characters used to con- struct the data matrix. The matrix was ana- lyzed in four different ways: 1. tarsal scopula of each leg was coded as a separate character (characters 22-25) and these were treated as ordered; 2. the four tarsal scopula were coded as a single character with six ordered states (character 30); 3. the four tarsal scopula were coded as a single character with six indepen- dent states (character 30); 4. the character(s) of the tarsal scopula were deactivated in the matrix. The optimization option was ACCT- RAN. Abbreviations: L = length of character; Cl = consistency index; RI = retention index. 0. Male tibial spur (L = 3; Cl = 0.66; RI = 0).--0, present; 1. absent; 2. present, formed by thick spines. The great diversity of male tibial spur morphology might be related to re- productive isolation. The tibial spur is the first structure that touches the female and could act as a mechanism for the female to recognize a conspecific male (Coyle 1985; Eberhard 1985; Jackson & Pollard 1990). Since the structures that compose the tibial spur are under inde- pendent evolution, the '‘tibial spur” is coded in three different characters. 1. Apical megaspiee in the tibial spur (L = 3; Cl = 0.33; RI — Oj.—O. present; 1. absent. 2. Prolateral branch of tibial spur (L = 3; Cl — 0.33; RI = 0).— 0. present; 1. absent. 3. Metatarsus I of males (L = 4; Cl = 0.25; RI = 0.66).— -0. straight; 1. dorsoven- trally curved. This character shows great var- iation in the degree of curvature and in the taxa in which it is present. 4. Flexion of metatarsus I of males (L = 1; Cl = 1; RI = 1).— -0. flexes outside the prolateral branch of tibial spur; 1. flexes be- tween the two branches of tibial spur. The way that the metatarsus flexes is related to the po- sition of the tibial spur. GUADANUCCI— TARSAL SCOPULA SIGNIFICANCE 461 Barychelidae 3 S 18 28 Reichlingia annae Holothele rondoni Sickius longibuibi Ischnocoius algericus Genus 1 Pterinochilus murinus Euathlus vulpinus Vital! us vellutinus Tapinauchenius sp Avicularia avicularia Catumiri petropolium Catumiri chicaoi Catumiri uruguayense Catumiri argentinense Oligoxystre 4 Oligoxystre 5 Oligoxystre 1 Oligoxystre 2 Oligoxystre 5 Barychelidae Catumiri petropolium Catumiri chicaoi Catumiri uruguayense Catumiri argentinense Ischnocoius algericus Oligoxystre 1 Oligoxystre 4 Oligoxystre 5 Oligoxystre 2 Oligoxystre 3 Holothele rondoni Sickius longibuibi Pterinochilus murinus Genus 1 Euathlus vulpinus Vitalius vellutinus Reichlingia annae Tapinauchenius sp Avicularia avicularia Figures 3—4. — Relationship hypothesis between Ischnocolinae and other Theraphosidae groups. 3. (L — 79; Cl = 0.48; RI = 0.56). Tarsal scopula coded as one character with six unordered states. 4. (L = 57; Cl = 0.57; RI = 0.72). Tarsal scopula deactivated. 462 THE JOURNAL OF ARACHNOLOGY Figure 5, — Variation possibilities of Ischnocolinae tarsal scopulae. States correspond to character 30. Full line corresponds to divided scopulae with setae; dashed line corresponds to entire scopulae with band of setae. 5* Ventral depression of the palpal tibia of males (L = 1; Cl = 1; RI = 1)*=0. straight or slightly curved, occupying more than half of the article; 1. sigma-like, occu- pying half the article. Autapomorphic for L algericus. Diagnostic for the genus Ischno- colus (Raven 1985). 6. Palpal bulb (L ^ 1; Cl = 1; RI = 1).— 0. apical keel absent; 1. apical keel present. The presence of keels on the bulb is a syna- pomorphy of Theraphosinae (Raven 1985; Perez-Miles et al. 1996; Bertani 2000). How- ever, small keels were observed on the bulb of some Oligoxystre species. These are not considered homologous to the Theraphosinae bulb keels. 7. Prolateral keels on the palpal bulb (L — 1; Cl = 1; RI = 1), — 0. absent; 1. present. Synapomorphy of Theraphosinae (Raven 1985; Perez-Miles et al. 1996; Bertani 2000). 8. Subtegulum (L = 1; Cl = 1; RI = 1).— - 0. narrow, not extending over the tegulum; 1. wide, extending over the tegulum. Synapo- morphy of Theraphosinae (Raven 1985; Per- ez-Miles et al. 1996). 9. Ventral region of the cymbium (L = 3; Cl = 0.33; RI = 0.6).— 0. as wide as long; 1. longer than wide. This character is very common among Ischnocolinae although it does not represent a synapomorphy for this group. 10. Size of the lobes of cymbium (L = I; Cl = I; RI = I),— 0. similar; 1. different. 11. Lobular state of spermathecae (L = 3; Cl = 0.33; RI = 0.71).— 0. unilobular (Fig. 7); 1. multilobular (Fig. 8). 12. Lateral lobe of spermathecae (L — 2; Cl = 0.5; RI ™ 0.5).— 0. absent (Fig. 7); 1. present (Fig. 8). 13. Maxillae (L = 2; Cl = 0.5; RI = 0.87) .“0. many cuspules (more than 50); 1. few cuspules (less than 45). 14. Labium (L = 6; Cl = 033; RI = 0.5) ordered.~0. cuspules absent; 1 . few cuspules (less than 10); 2. many cuspules (more than 15). 15. Labium shape (L = 5; Cl = 0.4; RI — 0.57).— 0. much wider than long (2.5-3 times wider); 1. almost as wide as long (less than 2 times wider); 2. longer than wide. 16. Tarsal claw (L = 5; Cl = 0.6; RI ^ 0,6)«— 0. bare, without teeth; 1. two rows of teeth; 2. median row of teeth; 3. prolateral row of teeth; 4. single tooth. 17. Tarsal claw with teeth (L = 1; Cl = 1; RI = 1) 0. present on all legs; 1. present on anterior legs (I-II). 18. Posterior sternal sigilla (L = 3; Cl = 0. 33. RI = 033).— 0. marginal; L submar- ginal. 19. Metatarsus I (L = 1; Cl = 1; RI ^ 1).— ”0. more than % of the article scopulate; 1. less than half of the article scopulate. 20. Metatarsus IV (L = 4; Cl = 0.25; RI = 0.62).— 0. less than half of the article scop- ulate; 1. more than half of the article scopu- late. 21. Legs spines (L = 3; Cl = 0.66; RI = 0) ordered.— 0. many spines, especially on tibia and metatarsus; L few reduced spines, on the apical region of tibia and metatarsus; 2. spines absent. The presence of several GUADANUCCI—TARSAL SCOPULA SIGNIFICANCE 463 spines was considered plesiomorphic for Theraphosoidina (Raven 1985). Representa- tives of Aviculariinae Simon 1874 show a re- duced number of leg spines. Bertani (2002) demonstrated that state 2 is a syeapomorphy for Aviculariinae semu stricto. 22. Tarsal scopula I (L = 2; Cl = 0.5; RI “ 0.85).— 0. divided by a longitudinal band of setae; 1. entire. 23. Tarsal scopula II (L = 4; Cl = 0.5; RI = 0.86) ordered.— 0. divided by a longi- tudinal band of setae; 1. entire with a longi- tudinal band of setae; 2. entire. In the present study a third state was identified which differs from state 0 in having type A setae (Roveer 1978; Perez-Miles 1994) mixed with type B setae (Rovner 1978; Perez-Miles 1994). In this state the tarsal scopula is not divided but there are lined setae forming a longitudinal band. 24. Tarsal scopula III (L = 1; Cl ^ 1; RI = 1).— 0. divided by a longitudinal band of setae; 1. entire. 25. Tarsal scopula IV (L = 2; Cl = 1; RI = 1) ordered.— 0. divided by a longitudinal band of setae; 1, entire with a longitudinal band of setae; 2, entire. 26. Clypeus (L = 3; Cl = 0.66; RI = 0.87).-— 0. absent; 1. present, narrower than the diameter of the anterior median eyes; 2. present, wider than the diameter of the ante- rior median eyes. 27. Urticatieg hair type III (L = 1; Cl = 1; RI = 1),— "0. absent; 1. present. Synapo- morphy of Theraphosinae (Raven 1985; Per- ez-Miles et al. 1996). 28. Apical article of posterior lateral spinnerets (L = 1; Cl = 1; RI = 1).— 0. domed or rounded; 1, digitiform. This char- acter is widely used to separate Barychelidae from Theraphosidae, the latter presenting the distal article of the PLS digitiform. However, among Barychelidae this character shows great variation making it impossible to place some representatives within either family. 29. Anterior maxillary projection (L = 1; Cl = 1; RI = 1).— 0. poorly developed; 1. developed. 30. Tarsal scopula I4V (L = 9; Cl = 0.66; RI = 0.92) (Fig. 5).— 0. all scopula di- vided; 1. only tarsal scopula I entire; 2. only tarsal scopula I entire and scopula II entire with a longitudinal band of setae; 3. only tar- sal scopula I and II entire; 4. tarsal scopula I- III entire; 5. tarsal scopula I~III entire and scopula IV entire with a longitudinal band of setae; 6. all tarsal scopula entire. Although state 1 was not observed in any of the speci- mens examined in this study, it was included in the matrix since it was observed in an on- togenetic series. The exuvia of a specimen of Grammostola actaeon (Pocock 1903) was ob- served and it demonstrated that from state 0 to 2, two steps must be counted since the tar- sal scopula II does not turn into undivided with a band of setae unless the scopula I is undivided. RESULTS The first cladogram (Fig. 1) refers to the tarsal scopula character separated into four in- dividual characters (characters 22-25). It re- sulted in a single tree (L = 72; Cl = 0.54; RI = 0.74). It shows that part of Ischnocolieae, represented by the taxa {Holothele ran- doni(Sickius longibulbi{lschnocolus algeri- cusdCatumiri))), is monophyletic. The re- maining Ischnocolieae form a distinct group with Harpactirinae Pocock 1897, Theraphosi- eae and Aviculariinae. This hypothesis sug- gests that part of Ischnocolinae is the sister- group to the polytomy presented by Raven (1985) solving in part the cladogram present- ed in that study (Fig. 6). The second cladogram (Fig. 2) refers to the tarsal scopula coded as a single character with six ordered states (character 30). It resulted in a single tree with the same topology and in- dices of the first cladogram but with different optimizations in the following nodes: Catu- miri petropoUum; {{Genus l{{Tapinauchemus sp.{{Ptermochilus murmus{Avicularia avicu- laria{Euathlus vulpinus+Vitalius vellutin- us){OUgoxystre spp.)))); {{Tapinaucherdus sp.((Fte™oc/w7M5' murinus{Avicularia avicu- laria{Euathlus vulpinus+Vitalius vellutin- us){Oligoxystre spp.))); {Pterinochilus murd nus{Avicularia avicularia{Euathlus vulpinusEV it alius vellutinus){OUgoxystre spp.)) and {{Pterinochilus murinus{Avicularia avicularia{Euathlus vulpinusEVitalius vellu- tinus)). Moreover it showed no homoplasies for the tarsal scopula character. The third cladogram (Fig. 3), which is a consensus of 16 trees, refers to the tarsal scop- ula coded as a single character with six in- dependent states. The only monophyletic groups in this tree are {Euathlus vulpi- 464 THE JOURNAL OF ARACHNOLOGY Figure 6. — Relationship hypothesis among Theraphosidae subfamilies (Raven 1985). nus^Vitalius veliutinus); (Tapinauchenius sp.-^-Avicularia avicularia); Catumiri spp. and Oligoxystre spp. The fourth cladogram (Fig. 4), which is a consensus of two trees, refers to the tarsal scopula character deactivated in the matrix. The only Ischnocolinae representatives that formed a monophyletic group are {Oligoxystre ^pp.i-lschnocolus algericus). The tarsal scopula I condition did not show any relation to spider size (t =-0.80433; P = 0.438247) (graphic 1). Divided tarsal scopula I is present in large species {H. rondoni) and entire scopula I in small (Genus 2 spp.). DISCUSSION Anterior-posterior gradation* — Some characters (e.g. number of spines, develop^ ment of teeth on the paired tarsal claws) show an anterior-posterior gradation that was de- scribed by Raven (1985). Concerning the tar- sal scopula, many species of Ischnocolinae present different states on legs I-IV, where there is a tendency towards the anterior legs Figures 7-8. — 7. Unilobular spermathecae {Catumiri argentinense). 8. Multilobular spermathecae, arrow showing the lateral lobe {Oligoxystre sp4). GUADANUCCI— TARSAL SCOPULA SIGNIFICANCE 465 presenting the apomorphic state. If the ple- siomorphic state (divided scopula) is present on leg I, then legs II-IV always show the same state. If the apomorphic state (entire scopula) is present on leg I, legs II-IV may or may not have the scopulae entire. The op= posite happens if we observe the state of the tarsal scopula on leg IV: if scopula IV is di- vided, it can be entire or divided on the an- terior legs. If scopula IV is entire, all the an- terior legs will present the apomorphic entire state. Moreover, if all the tarsal scopulae are divided on all legs, then leg IV will present the widest band of setae dividing the scopula. Character dependency. — Considering the ontogenetic differentiation (Gerschman de Pi- kelin & Schiapelli 1973), the anterior-poste- rior gradation (Raven 1985) and the variabil- ity of states (described above) of the tarsal scopula in the Theraphosidae it is reasonable to conclude that there is a dependency of this character between the legs. As such, a poste- rior tarsal scopula will not become entire un- less the anterior one is entire during the on- togenesis and an anterior tarsal scopula will not be divided if the posterior one is entire. An ontogenetic series composed of the exu- viae of a specimen of G. actaeon showing this sequenced transformation and all the combi- nations found within the Ischnocolinae (Fig. 5) confirms this character dependency. Tarsal scopula as a character. — Perez- Miles (1992) used the tarsal scopula as a char- acter on a preliminary cladistic analysis for Theraphosinae. In this analysis the plesiom- orphic state was ‘"at least one of the tarsal scopula divided”. It means that if a certain species has the tarsal scopula IV divided, it would be coded as plesiomorphic and the state of scopula I would be ignored. It would be interesting to study the condition of the ante- rior tarsal scopulae in the species that have the scopula IV divided. The use of the tarsal scopula in phyloge- netics was discussed by Perez-Miles (1994). According to him, results presented in that pa- per questioned the use of this character in cla- distic analysis since a close relation between body size and scopula condition (small sized species tend to possess divided scopula), in Theraphosinae and Harpactirinae adults, could suggest a functional adaptation or a develop- mental effect. However it is admitted that the role of this character in theraphosid evolution remains obscure. Perez-Miles (1994) ex- plained that the scopula of tarsus IV was the only one used in order to avoid ambiguity, since scopula division width increases on the hind legs (Raven 1985). It can be supposed that the tarsal scopula condition within Ther- aphosinae is either all legs with scopula di- vided or entire. Species of the genera HapaU opus Ausserer 1875 and Homoeomma Ausserer 1871 have all tarsal scopula divided (pers. obs.). Ischnocolinae is a very problematic group that lacks synapomorphies (Raven 1985) and its genera and species are mostly recognized by sexual characters (e.g. spermatheca and bulb morphology; presence, absence and mor- phology of structures of the tibial apophysis). So far, Ischnocolinae is considered a paraphy- letic subfamily and the results presented in this paper show that at least part of this group is supported by having more than half of the metatarsus IV occupied by scopula. The main difficulty to infer phylogenetic hypotheses for Ischnocolinae is the reduced number of char- acters that can be defined, since these spiders have a very homogeneous morphology. Dif- ferent from Theraphosinae, Ischnocolinae shows a great diversity of tarsal scopula states. If the divided scopula is related to small sized species, the large ones would be more likely to present the scopula I entire, which does not happen in Ischnocolinae (Fig. 9). Since the tarsal scopula condition is not related to spider size in Ischnocolinae (t = -0.80433; P = 0.438247), this character might have an important role in ischnocoline phylogenetics. This importance is evident when cladogram 4 is analyzed: monophyletic groups like {S. longibulbi+ Pterinochilus mu~ rinus) are based on characters that are very variable (clypeus and curvature of metatarsus I of males); the monophyletic group {{E. vul- pinusEV. vellutinus){R. annae(Tapmauchenius sp.+A. avicularia)) has R. annae (Trichopel- matinae) as the sister-group of Aviculariinae. This does not agree with the basal position of Trichopelmatinae within Theraphosidae pro- posed by Raven (1994). Furthermore, the po- sition of R, annae within this group does not agree with the monophyly of Theraphosinae + Aviculariinae proposed by Lucas (et al. 1991) and Perez-Miles (1992). 466 THE JOURNAL OF ARACHNOLOGY Figure 9. — Carapace length and tarsal scopula I condition for some Ischnocolinae species. It was possible to show all the variability of the tarsal scopula condition with the six state character. There was no mask on the var- iation of a character using the synthetic code, contrary to Pogue & Mickevich (1990). The difference between cladograms 1 and 2 is that the second one provided no homoplasies for the tarsal scopula character. Another differ- ence between the two optimizations is that the first cladogram has more synapomorphies on the nodes mentioned above. Since the char- acter dependency is admitted, these synapom- ophies (characters 22, 23, 24) might be false. The use of the tarsal scopula condition as six unordered states of only one character admits that the transformation from divided to entire scopula is independent in all pairs of legs. If this is true, we could find a spider with divid- ed anterior scopula and entire posterior scop- ula, or even scopula I and II entire, scopula III divided and scopula IV entire. The depen- dency of tarsal scopula condition between the legs means that the more entire the scopula the more apomorphic conditions (states) it will show. From these results it is possible to conclude that since there is relation between tarsal scopula condition and spider size, this character should be used in cladistic analysis. Additional Ischnocolinae taxa must be includ- ed in this analysis in order to provide a better knowledge of this character. ACKNOWLEDGMENTS I would like to thank Dr Ricardo Pinto-da- Rocha and Cristina Anne Rheims for their valuable suggestions on the manuscript and the following curators for loan of specimens: The material examined belongs to the follow- ing institutions (abbreviations and curators in parentheses): A.D. Brescovit (IBSP); N.I. Platnick (AMNH); C. Sciocia (MACN); L.A. Pererira (MLP); R. Pinto-da-Rocha (MZSP); N. Scharff (ZMUC); A.B. Bonaldo (MPEG). I also thank Dr Robert Raven for valuable comments and suggestions on the manuscript. Thanks also to the colleagues of the Labora- tory of Arachnology (LAL) of the Institute de Biociencias da Universidade de Sao Paulo for suggestions on the study. LITERATURE CITED Ausserer, A. 1871. Beitrage zur Kenntniss der Ar- achniden-familie der Territelariae Thorell (My- galidae Autor.). Verhandlungen der Zoologisch- Botanischen Gesellschaft in Wien 21:177-224. Bertani, R. 2000. Male palpal bulbs and homolo- gous features in Theraphosinae (Araneae, Ther- aphosidae). Journal of Arachnology 28:29-42. Bertani, R. 2002. Morfologia e evolugao das cerdas urticantes em Theraphosidae (Araneae). Tese de Doutorado, Instituto de Biociencias, Universida- de de Sao Paulo. Coyle, F.A. 1985. Observations on the mating be- havior of the tiny mygalomorph spider, Microh- GUADANUCCI— TARSAL SCOPULA SIGNIFICANCE 467 exura montivago Crosby & Bishop (Araneae, Di- pluridae). Bulletin of British Aracheological Society 6:328-330. Eberhard, W.G. 1985. Sexual selection and animal genitalia. Harvard University Press, Cambridge, Massachusetts. 244pp. Gerschman de Pikelin, B.S. & R.D. Schiapelli. 1973. La subfamilia “Ischnocolinae” (Araneae: Theraphosidae) Revista del Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”, En- tomologia 4:43-77. Goloboff, P.A. 1993. A reanalysis of mygalomorph spider families (Araneae). American Museum Novitates 3056:1-32. Guadanucci, J.P.L. 2004. Description of Catumiri n. gen. and three new species (Theraphosidae: Is- chnocolinae). Zootaxa 671:1-14. Jackson, R.R. & S.D. Pollard. 1990. Intraspecific interactions and the function of courtship in my- galomorph spiders: a study of Porrhothele anti- podiana (Araneae: Hexathelidae) and a literature review. New Zealand Journal of Zoology 17: 499-526. Lucas, S., P.I. da SilvaJr. & R. Bertani, 1991. The genus Ephebopus Simon, 1892. Description of the male of Ephebopus murinus (Walckenaer 1837). Spixiana 14:245-248. Nixon, K.C. & J.M. Carpenter. 1993. On outgroups. Cladistics 9:413-426. Page, R.D.M. 2001. Nexus Data Editor for Win- dows, version 0.5.0. Computer program available at http://taxonomy.zoology.gla.ac.uk/fiod/NDE/ nde.html Perez-Miles, E 1992. Analisis cladistico preliminar de la subfamilia Theraphosinae (Araneae; Ther- aphosidae). Boletin de la Sociedad Zoologica Uruguay (2° epoca) 7:11-12. Perez-Miles, E 1994. Tarsal scopula division in Theraphosinae (Araneae, Theraphosidae): its sys- tematic significance. Journal of Arachnology 22: 46-53. Perez-Miles, E, S.M. Lucas, P.I. da SilvaJr., & R. Bertani. 1996. Systematic revision and cladistic analysis of Theraphosinae (Araneae: Theraphos- idae). Mygalomorph 1:33-68. Pocock, R.L 1897. On the spiders of the suborder Mygalomorphae from the Ethiopian Region, contained in the collection of the British Muse- um. Proceedings of the Zoological Society of London 1897:724-774. Pogue, M.G. & M.F. Mickevich. 1990. Character definitions and character state delineation: the bete noire of phylogenetic inference. Cladistics 6:319-361. Raven, R.J. 1985. The spider iefraorder Mygalo- morphae (Araneae): cladistics and systematics. Bulletin of the American Museum of Natural History 182:1-180. Raven, R.J. 1994. Mygalomorph spiders of the Bar- ychelidae in Australia and the western Pacific. Memoirs of the Queensland Museum 35:291- 706. Roviier, J.S. 1978. Adhesive hairs in spiders; be- havioral functions and hydraulically mediated movement. Symposium of the Zoological Soci- ety of London 42:99-108. Rudloff, J.P. 1997. Revision der Gattung Holothele Karsch, 1879 nebst Aufstellung einer eeuen Gat- tung Stichoplastoris gen. nov. (Araneae, Thera- phosidae) und Wiedereinsetzung einiger weiterer Gattungen der Mygalomorphae. Arachnologisch- es Magazin 5(2): 1-19. Smith, A.M. 1990. Baboon Spiders. Tarantulas of Africa and the Middle East. Fitzgerald Publish- ing, London. Vol, F. 2001. Description de Pseudoligoxystre bo- livianus sp. et gen. n. (Araneae: Theraphosidae: Ischnocolinue), de Bolivie. Arachnides 50:3-10. Manuscript received 1 7 September 2004, revised 31 March 2005. 2005. The Journal of Arachnology 33:468-481 A PRELIMINARY STUDY OF THE RELATIONSHIPS OF TAXA INCLUDED IN THE TRIBE POLTYINI (ARANEAE, ARANEIDAE) Helen M. Smith: The Australian Museum, 6 College St, Sydney, New South Wales 2010, Australia and Faculty of Agriculture, Food and Natural Resources, The University of Sydney, New South Wales 2006, Australia. E-mail: hsmith@austmus. gov.au ABSTRACT. Poltys and the genera Cyphalonotus, Homalopoltys, Ideocaira, Kaira, Micropoltys and Pycnacantha have historically been considered members of the tribe Poltyini. There is little published information on most members of the group and their potential relationships in the context of recent advances in araneid systematics. Information is sought on possible relatives of Poltys. All araneid members of the group except Pycnacantha were added to the data matrix compiled by Scharff & Coddington (1997), which already contained Kaira. Homalopoltys was found to be a tetragnathid when males were identified and was not considered further. The full data matrix of 74 taxa and 82 characters was run in PAUP* and NONA. The resulting placement of Poltys was not well supported but it frequently occurred in association with members of a slightly modified version of the 'Hypsosinga clade’ of Scharff & Coddington, including Kaira. Cyphalonotus may be placed close to Araneiis and Ideocaira may also belong in the same area of the araneines. Micropoltys may belong in the sister clade to these two. Keywords: Poltys, Cyphalonotus, Ideocaira, Micropoltys, phylogenetic relationships. Spiders of the genus Poltys C.L. Koch 1 843 are distributed throughout the Old World, mostly in tropical and subtropical regions. The Australasian species mimic galls or dead twigs by day and exhibit morphological mod- ifications to enhance their cryptic disguise, making them rather odd-looking spiders. After some initial uncertainty over the affinities of the genus (Koch thought it might belong with taxa that are now included within Uloboridae) Simon (1895) placed Poltys in the subfamily Argiopinae as the nominative member of the tribe Poltyeae (here referred to as the Poltyini to conform with the International Code of Zoological Nomenclature). Also included by Simon were the genera Cyphalonotus Simon 1895, Homalopoltys Simon 1895, Kaira Cam- bridge 1889 and Pycnacantha Blackwall 1865. The genera Ideocaira Simon 1903 and Micropoltys Kulczyhski 1911 were described later, and their authors suggested that they might be related to Kaira and Poltys, respec- tively. More recently they were listed as part of the Poltyini (as ‘Poltyeae’) by Dippenaar- Schoeman & Leroy (1996). Archer (1951) recognized that the male pedipalp of Cyphal- onotus was far more complex than that of Pol- tys and proposed a new tribe, the Cyphalon- otini, for the former, later he decided it , belonged in the ‘Dolophini’ (Archer 1965). None of these tribes are currently in regular taxonomic use, and I am using the Poltyini grouping in the broadest sense, including all ! the above genera as the basis for this study. The phylogenetic analysis of araneid taxa by Scharff & Coddington (1997) was based on taxa selected from Simon’s tribes (or the earlier subfamily versions thereof), and Kaira was used as the representative of the Poltyini. The results suggested that Kaira should be placed in the "Hypsosinga clade’ in the mid- basal araneines. If Simon was correct in his affiliations of taxa this is where Poltys, and the remaining Poltyini taxa, should also be- long. However, Scharff & Coddington (1997) also found that some of Simon’s taxa were seriously polyphyletic. As Archer may have realized during his work on Cyphalonotus, the possibility of errors in Simon’s grouping of the Poltyini was compounded by his lack of knowledge of the males of almost all the gen- era in the tribe. Simon’s assemblage was ap- parently based on the irregular form of the abdomen, slightly unusual eye arrangements 468 SMITH— THE POLTYINI 469 i and the strong macrosetae on the legs of the I three genera which are now known to prey I mainly on moths (Kaira, Poltys and Pycna- I cantha) (Stowe 1986; Dippenaar-Schoeman & Leroy 1996). There is a confusing mixture of similarities and contradictions amongst char- acters within the genera of this putative group and also with respect to genera elsewhere in the Araneidae. These conflicts make the as- : sessment of the likely placement of Poltys within the Araneidae problematic. The primary motivation for this work was i to attempt to establish some possible relatives ; of Poltys which could provide a sensible out- : group taxon for an analysis of the Australasian Poltys taxa. Most of the other putative Pol- I tyini would not be suitable for this, even if they were closely related, because of the prob- lems of obtaining suitable recent material for destructive techniques such as the extraction of DNA. Nevertheless, I was still intrigued by some of the characters exhibited by these taxa and their superficial similarities to Poltys. Therefore, there were two goals to this study. The first aim was to test whether Poltys might indeed belong in the 'Hypsosinga clade’ of Scharff & Coddington 1997 (and if not, where). Secondly, to find out whether, without any changes or additions to the characters used, the Poltyini would emerge as a mono- phyletic grouping within the context of the taxa examined by Scharff & Coddington 1997. METHODS Taxa. — -The genus Pycnacantha was ex- cluded, as no male specimens were available. Kaira was recently revised by Levi (1993) and was included by Scharff & Coddington 1997 in their study. The other genera of Poltyini are generally poorly known and it was first nec- essary to identify males for Homalopoltys, Ideocaira and Micropoltys, which are de- scribed only from females. When Homalopol- tys males were found it became apparent that this taxon is in fact a tetragnathid. This genus was therefore excluded from further analysis here. The female type of Ideocaira transversa Simon 1903 has been examined, and unpub- lished drawings of the female type of Micro- poltys placenta Kulczynski 1911 were sup- plied by H. Levi. Unfortunately, none of the species in which males could be matched to females represented the type species of the ge- nus. For Cyphalonotus, the expanded pedipalp is from a different species to that used for scoring general characters (necessitated by the need to use material from the only vial which contained more than a single male). The struc- tures visible on the unexpanded pedipalp of the species against which other male and fe- male characters were scored appear to be sim- ilar; there are also no scoreable differences in the general attributes in the males of both spe- cies. Neither species has been identified, the type species, C. larvatus (Simon 1881), is re- corded from Congo and East Africa (Platnick 2005). This leaves Poltys illepidus C.L. Koch 1843 as the only type species used in this analysis. Although this is far from ideal, the nature of this data set, with a rather high pro- portion of taxa to characters, meant robust re- sults were unlikely even before adding addi- tional taxa (Scharff & Coddington 1997). Therefore, I did not expect to achieve precise results in this tentative exploration of these genera and any more rigorous analysis would need to address these issues. Abbreviations. — The following abbrevia- tions for morphological features were used throughout the text and figures: C = conduc- tor; CY = cymbium; E = embolus; MA = median apophysis; PC = paracymbium; PM = paramedian apophysis; R — radix; S = sti- pes; SEM = scanning electron microscope; T — tegulum; TA = terminal apophysis; TL = tegular lobe. The following abbreviations were used for repository institutions: AM ™ Australian Museum, Sydney, Australia; MNHNP “ Museum National d’Histoire Na- turelle, Paris, France; MRAC ~ Koninklijk Museum voor Midden Afrika, Tervuren, Bel- gium; NCAP = National Collection of Arach- nida, Pretoria, South Africa; NHRM ^ Swed- ish Museum of Natural History, Stockholm, Sweden; QM = Queensland Museum, Bris- bane, Australia; RMNH = National Museum of Natural History, Leiden, The Netherlands; UNAM ~ lestituto de Biologia, Universidad Nacional Autonoma de Mexico, Mexico D.E, Mexico; ZMB = Museum fiir Naturkunde, Zentralinstitut der Humboldt-Universitat, Ber- lin, Germany. Characters. — The character attributes for each of the selected taxa were examined and scored according to the methods of Scharff & Coddington 1997. The specimens examined are shown in Table 1 and attribute codings are 470 THE JOURNAL OF ARACHNOLOGY Figures 1-6. — Scanning electron micrographs of Poltys and Micropoltys: 1. Poltys illepidus from Trinity Park, male; expanded pedipalp, apico-dorsal view; 2. Poltys illepidus from Lakeland, male, pedipalp, prolateral. 3-6. Micropoltys sp. from W of Cape Kimberley, male: 3. Pedipalp, prolateral; 4, 5. Modified setal bases and sensory seta on carapace and sternum, respectively; 6. Prosoma, frontal view. See text for abbreviations. Scale bars Figs. 1, 2 (30 fxm). Figs. 3-5 (20 |xm). Fig. 6 (100 fxm). SMITH—THE POLTYINI 47 1 Figures 1 ~~12. —Poltys illepidus: 7-9. Male from Trinity Park: 7. General lateral view; 8. Ditto but at same scale as female; 9. Left pedipalp, prolateral. 10. Male from Rockhampton, left pedipalp, expanded, prolateral. 11. Female from Trinity Park: General lateral view. 12. Female from Brisbane, epigynum, ventral. See text for abbreviations. Scale bars Figs. 7, 11 (1 mm). Figs. 9-10, 12 (0,5 mm). shown in Table 2. The full list of characters is not repeated here but most characters are adequately illustrated in Figs. 1-30. Some characters, listed below, do require some com- ment on their interpretation in relation to the Scharff & Coddiegtoe 1997 analysis. Characters 11 and 12: Median apophysis of male pedipalp with bifid prong or threadlike spur. The apically directed hook-like portion of the Poilys MA is very distinctive (Figs. 1, 9). However, it does not conform totally to either of the diagnoses for these character states. Character 19: Stipes absent or present. In Micropoltys the sperm duct appears to pass from the radix, through the base of the distal haematodocha and straight into the embolus. There is apparently no sclerite as such be- tween the two, so this is scored absent [0] (Fig. 28). Character 23: Tip of male pedipalp embolus simple or with cap. Only Poltys and Micro- poltys pedipalps have been examined under SEM (Figs. 2, 3). There is eO' indication on either of these that any part is designed to break off, or has already done so. These are scored as simple [0] . The attributes of the oth- er genera are unknown so they are scored [?]. Character 30: Scape with pocket near tip, absent or present. Poltys illepidus have a broad turned-over rim along the whole of the posterior margin of the epigyee (Fig. 12). I have interpreted this as a (rather wide) pocket present [1]. Micropoltys females have at least a sharp depression which is tentatively also scored here as a pocket present [1] (Fig. 30). Characters 33 and 34: Coxa I hook and fe- mur II groove. Among these taxa, all of the males with similarly sized females have these features (e.g. coxal hook arrowed in Fig. 6, Micropoltys). Character 46: Clypeal tooth of females ab- sent or present. Both males and females of the Micropoltys species figured have a rather rounded clypeal tooth. The male is shown in Fig. 6, but the tooth is more developed in fe- males. This character is not present in Levi’s 472 THE JOURNAL OF ARACHNOLOGY Figures 13-19. — Cyphalonotus sp.: 13, 14. Male from Natal: 13. General lateral view; 14. Left pedipalp, prolateral; 15, 16. Male from Misahohe: Left pedipalp expanded, prolateral and retrolateral (different species to Fig. 14). 17 — 19. Cyphalonotus sp. from Natal, female: 17. General lateral view; 18, 19. Epigynum, ventral and lateral. See text for abbreviations. Scale bars Figs. 13, 17 (1 mm). Figs. 14-16, 18, 19 (0.5 mm). drawing of the type female of Micropoltys placenta but I have scored it as present [1]. Character 50: Ratio of lateral eye-median eye separation, < 1 or > 1. Poltys and Mi- cropoltys are unusual among araneids in that they have widely separated lateral eyes, so there is no lateral eye group as such (Figs. 7, 11, 26, 29). In applying this character to these genera I took the Scharff & Coddington 1997 instructions literally, and used the distance at the widest point, i.e. that to the posterior eye, so that the separation is scored as > 1 [1]. Characters 59 and 60: Abdominal shape. Both male and female Ideocaira triqueta Si- mon 1903 have strongly triangular abdomens, which are widest anteriorly (Fig. 24, female). The females of /. triqueta vary in their relative dimensions, some being wider than long and some the reverse. However, the female of /. transversa, the type species, is distinctly wid- er, so I have used this to decide the matter and scored Character 60 as wider [1]. Character 67: Tactile setal bases on cara- pace and abdomen, normal or gasteracanthine- shaped, Micropoltys has rather distinctive se- tal bases over much of the prosoma, including the basal chelicerae (Fig. 6). There are none on the dorsum of the abdomen, but they do occur around the pedicel on the venter. Some of these bases and the setae themselves (Fig. 4) are extremely similar to those figured by Scharff & Coddington 1997 and I have scored them as gasteracanthine-like [ 1 ] . Those on the sternum (Fig. 5) and around the eye region and chelicerae are further modified, with an anteriad-projecting lamella and deep pits on each side. Characters 74 and 75: Orb web and sticky spiral. Joseph Koh has provided me with a photograph of Cyphalonotus in an orb web. I cannot see anything to suggest that it is not a normal araneid web and so have scored Char- acter 75, sticky spiral, as present [0]. (This SMITH— THE POLTYINI 473 Table L — Details of specimens examined in this study. Sex & Species Locality data Coordinates Repository & No. Used for S' 9 Cyphalonotus sp. Natal, South Af- rica MNHNP 19654 All codings; Figs. 13, 14, 17-19 S Cyphalonotus sp. Misahohe, Togo 06°57'N, 00°35'E ZMB (unreg’d) Expanded pedi- palp; Figs. 15, 16 S 9 Meocaira triqueta Mzimhlava river mouth, Lusiki- siki district, Eastern Cape, South Africa 3r22'S, 29°35'E MRAC 166621 All codings; Figs. 20-25 9 Meocaira triqueta Port Elizabeth, Eastern Cape, South Africa 33°58'S, 25°35'E MNHNP 18508 Types (2), used to confirm ID 9 Meocaira transversa Natal, South Af- rica MNHNP 16334 Type S Micropoltys sp. Cape Kimberley, Queensland, Australia 16°16'S, 145°28'E AM KS86251 Pedipalp; Fig. 27 S Micropoltys sp. Cape Kimberley, Queensland, Australia 16°16'S, 145°28'E AM KS86252 Expanded pedi- palp, general codings; Figs. 26, 28 S Micropoltys sp. W of Cape Kim- berley, Queensland, Australia 16°15'S, 145°26'E AM KS86740 SEM; Figs. 3-6 9 Micropoltys sp. Cooktown, Queensland, Australia 15°29'S, 145°15'E AM KS57876 General codings; Fig. 29 9 Micropoltys sp. W of Cape Kim- berley, Queensland, Australia 16°15'S, 145°26'E AM KS57890 Epigynum; Fig. 30 S Poltys illepidus Trinity Park, N Cairns, Queensland, Australia 16°48'S, 145°427E AM KS86253 General codings; Figs. 7-9 S Poltys illepidus Rockhampton, Queensland, Australia 23°22'S, 150°29'E AM KS58033 Expanded pedi- palp; Fig, 10 S Poltys illepidus Trinity Park, N Cairns, Queensland, Australia 16°48'S, 145°42'E AM ex eggsac laid by KS86257 SEM; Fig. 1 S Poltys illepidus Lakeland, SW of Cooktov/n, Queensland, Australia 15°50'S, 144°53'E AM KS58017 SEM; Fig. 2 9 Poltys illepidus Trinity Park, N Cairns, Queensland, Australia 16°48^S, 145°42'E AM KS86258 General codings; Fig. 11 9 Poltys illepidus Brisbane, Queensland, Australia 27°30'S, 152°58'E QM S20786 Epigynum; Fig. 12 $ Poltys illepidus Edmonton, Queensland, Australia 17°0rS, 145°44'E AM KS86310 SEM (spinnerets, not figured) 474 THE JOURNAL OF ARACHNOLOGY Figures 20-25. — Ideocaira triqueta from Lusikisiki district: 20-22. Male: 20. General lateral view; 21, 22. Left pedipalp, prolateral and expanded, dorsal view. 23-25. Female: 23. General lateral view; 24. Abdomen, dorsal; 25. Epigynum, ventral. See text for abbreviations. Scale bars Figs. 20, 23, 24 (1 mm). Figs. 21, 22 (0.5 mm). Fig. 25 (0.25 mm). character makes no difference to the position of Cyphalonotus in the results). Character 78: Sticky-spiral (SS) localiza- tion: outer leg 1, inner leg 1 or leg 4. In the Poltys species I have observed spinning webs, leg 4 is mostly used to monitor the position of the spider with respect to the sticky spiral, especially closer to the hub where the distance between radii is very short (Smith unpub. data). I do not have notes on the behavior of P. illepidus itself, but the web is similar to the species I observed and I have therefore scored it as L4 [2]. These Poltys species also move around the web in a similar way to the larger nephilines (Scharff & Coddington 1997; Eber- hard 1982), constantly facing between the hub and the direction of travel. Like these nephi- line spiders, Poltys makes a finely meshed web, which probably influences the most ef- ficient way of moving around the web (Eber- hard 1982). Analysis. — The full set of data (74 taxa, 82 characters) was run in PAUP* (Swofford 2001) using a heuristic search with the com- mands: hsearch addseq= random nchuck“5 chuck- score =1 nreps=1000 randomize = trees; Table 2. — Character attribute codings for the newly added Poltyini taxa. See Scharff & Coddington (1997) for full list of characters. Character number 0 1 1234567890 1 2 1234567890 2 3 1234567890 3 4 1234567890 Cyphalonotus Ideocaira Micropoltys Poltys 1111110100 0111110000 0021110000 0000110000 0000011011 0000011010 0000011001 0000011110 0170000101 -170000101 0100000101 -100000101 1011000011 0011000011 0011000011 0000000011 SMITH— THE POLTYINI 475 Figures 26-30. — Micropoltys sp.: 26-28. Male from Cape Kimberley: 26. General lateral view; 27, 28. Left pedipalp, prolateral and expanded, apico-dorsal view. 29. Female from Cooktown, General lateral view; 30. Female from W of Cape Kimberley, Epigynum, ventral. See text for abbreviations. Scale bars Figs. 26, 29 (1 mm), Figs. 27, 28, 30 (0.25 mm). hsearch start=current nchuck = 0 chuck- score =0; The first line keeps only 5 trees from each island sampled, preventing the tree buffers from filling with thousands of trees and in- creasing the chances of finding all islands of trees. One thousand replicates are carried out, each time with the taxa added in a random order. The default branch swapping algorithm TBR (tree bisection reconnection) is used. The order of the resulting trees is randomized be- fore entering the second line of command. The second line swaps on the trees kept from the first search to completion. All data was also run in NONA (Goloboff 1993) using the standard commands, as rec- ommended by Miller (2000): mult* 1000; max*; or jump* 1; Before using any consensus method in Table 2. — Extended. Character number 4 5 5 6 6 7 7 8 8 1234567890 1234567890 1234567890 1234567890 12 0000100010 110000-000 0001000100 999'! 1 99999 ? ? 0000100011 110000-061 0001000100 9-1 99999999 ? ? 0000110001 110000-000 00000011?? 9999999999 ? ? 0000000011 110000-000 1001000100 2101000200 00 476 THE JOURNAL OF ARACHNOLOGY PAUP* it is desirable to check through the topologies and delete any with zero-length branches (Scharff & Coddington 1997). NONA’S algorithms are better in this regard but the program can still produce uncollapsed polytomies which are suboptimal once col- lapsed. Scharff & Coddington 1997 also ad- vocate the filtering of tree sets to remove those trees containing polytomies for which there is a more resolved solution present. With the solution present in another, otherwise identical tree, it is reasonable to support their interpretation as ‘soft’ polytomies, i.e. irreso- lution due to a lack of data, rather than ‘hard’ polytomies which is an assertion of simulta- neous cladogenesis (Coddington & Scharff 1996). The tree data set can be filtered in PAUP* but the removal of trees containing zero-length branches is more problematic. Two methods used here are the manual re- moval of the topologies with assigned zero- length branches from a saved PAUP* tree file, or alternatively using WinClada (Nixon 1999- 2002) by a process of collapsing unsupported nodes then removing suboptimal trees. The tree set produced by NONA can also be ‘cleaned up’ using WinClada, but cannot eas- ily be filtered. While tree data sets from either PAUP* or NONA can be imported into WinClada and back into NONA, once export- ed from PAUP* retrieving them is difficult. An Adams consensus (Adams 1972; imple- mented in PAUP*) was required to examine whether clades might be recovered which would otherwise not be found by more simple consensus methods. Consequently, the tree set primarily used is that produced by PAUP*’s filtering and the manual removal of topologies with zero-length branches. However, this is not the same as the set obtained by passing the filtered trees through the WinClada rou- tine. It was decided that both methodologies should be used to confirm that any conclu- sions drawn were supported in both cases. Strict, majority-rule and Adams consensus trees were produced in PAUP* and all topol- ogies were examined using WinClada. RESULTS PAUP* initially found 948 minimal length trees (300 steps). This was reduced to 376 trees by filtering and finally 156 trees after manual removal of topologies with zero length internal branches (referred to subse- quently as the ‘manual tree set’). After passing the filtered set through WinClada, 132 topol- ogies remained (the ‘WinClada tree set’). NONA found 344 initial trees using the jump*l command (length 300, as PAUP*), which is reduced to 232 trees after collapsing polytomies in WinClada. These topologies are the same as those in the PAUP* data set (shown by putting the unfiltered PAUP* tree set through WinClada: the same 232 trees are found). Using the max* swapping algorithm was less effective and only recovered 308 trees, or 192 trees post WinClada. All the consensus trees maintain the out- group structure and basal araneid placement of Chorizopes O.R-Cambridge 1870 found by Scharff & Coddington 1997 (fig. 82, Fig. 31). The araneines become a bush beyond this point in the strict consensus tree (Fig. 31), al- though with a few resolved terminal clades. All the Poltyini examined here are found with- in the Araneinae (sensu Scharff & Coddington 1997 except for Scoloderus Simon 1887). The majority-rule tree produced from the Win- Clada tree set is slightly less resolved than that shown from the manual tree set (Fig. 32): two additional levels are collapsed in the ar- aneines, so that Hypsosinga Ausserer 1871 and Dolophones Walckenaer 1837 are in the main araneine ‘bush’. The position of Poltys within the araneines is unresolved by all the consensus methods (Figs. 31-33). The character partition table from PAUP* indicates that Poltys pairs with Zygiella F.O.P.-Cambridge 1902 (31% of trees) or Kaira (15%) in the manual tree set, and there are several combinations of a clade involving Poltys and some or all of Zygiella, Kaira, Metepeira F.O.P.-Cambridge 1903, Sin- ga C.L. Koch 1836, and Larinia Simon 1874. Examining trees, these sub-arrangements add up to 61% of topologies. This group is all of the Scharff & Coddington 1997 'Hypsosinga clade’ (clade 44), except Hypsosinga itself and with the addition of Larinia, which also frequently came into this clade in the Scharff & Coddington 1997 analysis. In other topol- ogies there is usually a series of single taxon ‘steps’ in the basal araneines, in which Poltys occurs, often with other parts of the 'Hypso- singa clade’ emerging as adjacent steps. In many trees with this type of topology, Witica O.R-Cambridge 1895 and Arachnura Vinson 1863 are also present in the very base of the SMITH— THE POLTYINI 477 c Chorizopes Gasteracantha Aetrocantha Togacantha Isoxya Austracantha Macracantha Gastroxya Augusta Aspidolasius Caerostris Hypognatha Arkys Archemorus Encyosaccus Xylethrus Chaetacis Micrathena Mastophora Cyrtarachne Pasilobus Arachnura Witica Mecynogea Cyrtophora Neogea Argiope Gea Scoloderus Acanthepeira Anepsion Dolophones Hypsosinga Zygiella Kaira Metepeira Singa Larinia Neoscona Mangora Cercidia Pronous Aculepeira Araneus Cyphalonotus Bertrana Enacrosoma Alpaida Wixia Acacesia Metazygia Eustala Cyclosa Nuctenea Colphepeira Araniella Eriophora Verrucosa Poltys IVlicropoItys Ideocaira Figure 31. — Strict consensus of the Araneidae for the data of Scharff & Coddington 1997 and taxa from the Poltyini (in bold). Clade numbers show relevant areas of agreement with Scharff & Coddington 1997 (fig. 82). 478 THE JOURNAL OF ARACHNOLOGY 62 92 54 62 62 62 67 67 67 100 100 100 100 100 100 Scoloderus Acanthepeira Anepsion Zygiella Kaira Metepeira Singa Larinia Neoscona Mangora Cercidia Pronous Aculepeira Araneus Cyphalonotus ideocaira Bertrana Enacrosoma Alpaida Wixia Acacesia Metazygia Eustala Cyclosa Nuctenea Colphepeira Araniella Eriophora Verrucosa Poltys Micropoltys Dolophones Hypsosinga Figure 32. — Majority-rule consensus of the Araneinae for the data of Scharff & Coddington (1997) and taxa from the Poltyini (in bold). Numbers show the percentage of topologies containing the particular clade (>50% only). araneine branch. In the WinClada tree set, 55% of topologies placed Poltys with various permutations of this modified 'Hypsosinga clade’, and the figures for pairing with Zyg- iella or Kaira are 27% and 18%, respectively. Poltys never appears in clades with any other taxa in either tree set. The only Poltyini taxon to be resolved with- in the araneine ‘bush’ in the strict consensus is Cyphalonotus, which is the sister taxon to the Scharff Sl Coddington 1997 clade 60 of {Araneus Clerck 1757 + Aculepeira Cham- berlin & Ivie 1942) (Fig. 31). The majority- rule and Adams consensus trees both suggest Ideocaira may belong among or near the Scharff & Coddington 1997 clade 57 (but now also containing Cyphalonotus and possibly without Larinia) (Figs. 32-33). In every to- pology Ideocaira occurs in a trichotomy with Neoscona Simon 1864. The majority-rule tree shows Kaira as sister to Metepeira, as previ- ously found by Ssharff & Coddington 1997 (clade 47). Micropoltys is best resolved by the Adams tree which recovers a clade where it is sister to Alpaida O.P.-Cambridge 1889 + {Bertrana Keyserling 1884 + Enacrosoma Mello-Leitao 1932) (Scharff & Coddington j 1997 clade 64). Examination of the trees in- | dicates that Micropoltys is always found either ! at the base of this clade plus its sister clade, or at the base of its sister clade. These results are true for either tree set. DISCUSSION ! 1 Any topology resulting from a consensus method is simply a statement about areas of j agreement among trees (Swofford 1991). Fig- t ures 31-33, therefore, are not presented as an I actual suggestion of phylogeny, but merely . I serve to suggest the taxa among which these new additions might be placed. The question of whether Poltys should be included in the 'Hypsosinga clade’ remains uncertain. In these results it is most frequently associated with one or more of the genera I Zygiella, Kaira, Metepeira, Singa and Larinia, 1 most of which are indeed from this clade. However, the inclusion of P. illepidus in the data set destabilises the arrangement found by Scharff & Coddington 1997 and reduces the former clade to a loose association of genera with variable placement within the Araneinae. Despite this, one of these genera would pro- vide the best choice of outgroup given the cur- rent evidence. However, a cautionary com- ment about other Poltys taxa is required. One . of the criteria Scharff & Coddington 1997 used when selecting taxa to include in their analysis, was that the species which were scored should be typical for the genus, or at least an accepted part of the genus. Through- out the genus Poltys there is considerable var- iation in eye arrangements, in presence or ab- sence of a scape on the female epigynum and in some endemic Australian species, presence or absence of a terminal apophysis in the male pedipalp (Smith unpub. data). These are all used as generic characters in this data matrix, yet vary within this genus. Consequently, it is possible that the genera which appear as po- tential relatives in the scenario above might be different if one of the more aberrant Poltys species were used instead. Here, P, illepidus, in addition to being the type species, was judged to be the most useful exemplar as it seems to exemplify the ‘basic’ Poltys body plan, and lacks some of the apparently more I 480 THE JOURNAL OF ARACHNOLOGY derived characters seen elsewhere in the ge- nus. The second aim of this study was to test whether the taxa formerly included in the Pol- tyini would appear as a group when included with the taxa analysed by Scharff & Codding- ton. Even ignoring Homalopoltys, which ap- pears to be a tetragnathid (Smith unpub. data), it is extremely unlikely that the remaining taxa form a monophyletic grouping, although they may all occur scattered among a broader group of araneines. Cyphalonotus is the most consistently placed of these taxa, close to Ar- aneus, and Ideocaira may also belong in the same area of the araneines (Scharff & Cod- dington 1997 clade 57). Micropoltys may be- long in the sister clade to these two (which would be clade 62 in Scharff & Coddington 1997, fig. 82), and, as already discussed, Pol- tys may belong in or near the "Hypsosinga cladek However, given the limitations of this study noted above, these preliminary findings should be subjected to further analysis when the opportunity becomes available. ACKNOWLEDGMENTS I am greatly indebted to Herb Levi & Ma- tjaz Kuntner for making available various drawings and notes which enabled me to rec- ognize several of these genera. Many thanks are also due to the various individuals and in- stitutions from whose collections the relevant specimens have come (not all of these have found their way into this paper, but all were instrumental to the recognition of taxa): Cesar Duran-Barron (UNAM), Christa Deeleman- Reinhold, Ansie Dippenaar-Schoeman & Eliz- abeth Kassimatis (NCAP), Jason Dunlop (ZMB), Rudy Jocque (MRAC), Torbjorn BCro- nestedt (NHRM), Graham Milledge (AM), Erik van Nieukerken & Kees van den Berg (RMNH), Christine Rollard (MNHNP). Thanks too to the many people who have free- ly given time and advice: Nikolaj Scharff who helped me Translate’ the Poltys pedipalp sclerites and provided useful advice, particu- larly at the review stage; also referee Volker Framenau, who was most helpful; Shane Ahyong and Max Moulds for assistance with phylogenetic programs, and Greg Edgecombe for useful search strategies and advice; Sue Lindsay for SEM; Joseph Koh for information on Cyphalonotus; Graham Milledge for sup- port and proofreading. Also to my supervisors Mike Gray and Harley Rose for unstinting en- couragement, sound advice and helpful com- ments on the draft manuscript. I am also in- \ debted to the Australian Museum for use of ; facilities, to The Linnean Society of New South Wales from the Joyce W Vickery Sci- entific Fund for grant assistance for equipment i and to the Organizing Committee of the 16‘^ ICA for assistance towards attending the con- i gress. LITERATURE CITED ; Adams, E.N.III. 1972. Consensus techniques and : the comparison of taxonomic trees. Systematic Zoology 21:390-397. Archer, A.E 1951. Studies in the orbweaving spi- , ders (Argiopidae). 2. American Museum Novi- tates 1502:1-34. Archer, A.E 1965. A new species of Cyphalonotus from Central Africa (Araneae, Argiopidae). Re- vue de Zoologie et de Botanique Africaines 72: 79-82. Coddington, J.A. & N. Scharff. 1996. Problems with “soft polytomies”. Cladistics 12:139-145. Dippenaar-Schoeman, A.S. & A. Leroy. 1996. Notes on the biology of Pycnacantha tribulus, another araneid without an orbweb (Araneae: Ar- aneidae). Revue Suisse de Zoologie, voL hors serie: 165-171. Eberhard, W.G. 1982. Behavioral characters for the higher classification of orb-weaving spiders. Evolution 36:1067-1095. Goloboff, P.A. 1993. NONA. Noname (a bastard son of Pee- Wee), version 2.0 (for Windows). Program and documentation. Computer program distributed by J. M. Carpenter, Dept, of Ento- mology, American Museum of Natural History, New York Koch, C.L. 1843. Die Arachniden. Band X. C.H. Zeh’sche Buchhandlung, Niimberg. Kulczynski, W. 1911. Spinnen aus Nord-Neu-Guin- ; ea. In Resultats de U Expedition Scientifique ^ Neerlandaise a la Nouvelle-Guinee en 1903 sous | les Auspices d’ Arthur Wichmann. Leiden, Zool. | 3(4):423-518. Levi, H.W. 1993. The orb- weaver genus Kaira (Ar- aneae: Araneidae), Journal of Arachnology 21: 209-225. Miller, J. 2000. NONA for Hennig users, version 2.0. George Washington University, online at http://www.gwu.edu/~clade/spiders/nona.htm (accessed 23 April 2005). Nixon, K.C. 1999-2002. WinClada, Version 1.0000. Published by the author, Ithaca, New York, USA. Platnick, N.I. 2005. The World Spider Catalog, Ver- sion 5.5. American Museum of Natural History, online at http://research, amnh.org/entomology/ SMITH— THE POLTYINI 481 spiders/catalog/index. html (accessed 29 June 2005). Scharff, N. & J.A. Coddington. 1997. A phyloge- netic analysis of the orb-weaving spider family Araneidae (Arachnida, Araneae). Zoological Journal of the Linnean Society 120:355-434. Simon, E. 1895. Histoire Naturelle des Araignees I. Pp. 761-1084. Librairie Enclyclopedique de Ror- et, Paris. Simon, E. 1903. Descriptions d'arachnides nou- veaux. Annales de la Societe Entomologique de Belgique 47:21-39. Stowe, M.K. 1986. Prey specialization in the Ara- neidae. In Spiders: Webs, Behavior and Evolu- tion (W.A. Shear, ed.). Stanford University Press, Stanford, California. Swofford, D.L. 1991. When are phylogeny esti- mates from molecular and morphological data in- congruent? In Phylogenetic analysis of DNA se- quences (M.M. Miyamoto & J. Cracraft, eds). Oxford University Press, New York and Oxford. Swofford, D.L. 2001. PAUP*. Phylogenetic Anal- ysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Mas- sachusetts. Manuscript received 1 7 September 2004, revised 5 August 2005. 2005. The Journal of Arachnology 33:482^89 A FOSSIL HARVESTMAN (ARACHNIDA, OPILIONES) FROM THE MISSISSIPPIAN OF EAST KIRKTON, SCOTLAND Jason A* Dunlop: Institut fur Systematische Zoologie, Museum fur Naturkunde der Humboldt-Universitat zu Berlin, InvalidenstraBe 43, D^lOllS Berlin, Germany. E-=mail: jason.dunlop @ museum.hu-berlin.de Lyall I. Anderson: Department of Natural Sciences, National Museums of Scotland, Chambers Street, Edinburgh, EHl IJF, EFnited Kingdom ABSTRACT. A fossil harvestman (Arachnida, Opiliones) from the Mississippian (Visean: Brigantian) of East Kirkton, Scotland is described as Brigantibunum Ustoni new genus and species. At ca. 340 Ma, it represents the second oldest record of Opiliones. Although some details are lacking, this long-legged, small-bodied and rather gracile harvestman is surprisingly modern-looking and appears to show the im- pression of an annulate ovipositor. Its leg anatomy closely matches that of some living Eupnoi and it is tentatively referred to this clade. Like the newly discovered Rhynie chert harvestmen, it reinforces the idea that modern, crown-group Opiliones can be traced back to at least the mid-Paleozoic. Keywords: Taxonomy, new species, Visean, Paleozoic, Eupnoi Fossil harvestmen (Opiliones) are very rare. Although they are best known from Tertiary ambers (e.g., Cokendolpher & Poinar 1998; Stargga 2002; Dunlop & Giribet 2003), there are also two Mesozoic records (Roger 1946; Jell & Duncan 1986) neither of which have been formally named and the first of which is dubious. Until recently the earliest harvestmen were Pennsylvanian (ca. 300 Ma) fossils from the Coal Measures of Commentry in France (Thevenin 1901) and Mazon Creek in the USA (Petrunkevitch 1913). Restudy of Pe- trunkevitch's (1913) putative extinct order Kustarachnida has shown that these Mazon Creek arachnids are misidentified harvestmen (Beall 1986, 1997; Dunlop 2004a). Additional Pennsylvanian-aged harvestmen have been collected from Missouri, USA (Dunlop 2004b) and Poland (Maciek Kania, pers. comm., 2004). The oldest recorded harvestmen are excep- tionally preserved, three-dimensional fossils from the Early Devonian (ca. 400 Ma) Rhynie cherts of Scotland (Dunlop et al. 2003, 2004). Provisionally assigned to the Eupnoi clade as Eophalangium sheari Dunlop et al. 2004, this exquisite, silicified material displays details of internal structures such as genitalia and tra- chea, all of which point towards these ancient harvestmen having a very similar gross mor- phology to living animals. The second oldest harvestman is also rather modern looking and comes from the Mississippian (ca. 340 Ma) of East Kirkton in Scotland (Wood et al. 1985). The East Kirkton harvestman (Figs. 1-3) is preserved in a more typical fashion as a flat- tened impression and superficially resembles living 'daddy long-legs’ forms. In general, Mississippian arachnid fossils are far less common than either Devonian or Pennsylva- nian examples (see Dunlop & RoBler 2003 for a review), thus any record from this time pe- riod is significant. The East Kirkton harvestman was briefly mentioned, with a figure, in the original sum- mary paper dealing with the locality (Wood et al. 1985), It has also been noted or listed in some further publications (Smithson 1989; Selden 1993a, b; Clack 1998; Jerarn 2001; Dunlop & RoBler 2003; Dunlop 2004a). Here we formally describe and name this important specimen and discuss its significance in the light of other recent fossil harvestman discov- eries. METHODS The East Kirkton harvestman was borrowed on research loan from the Hunterian Museum, 482 DUNLOP & ANDERSON— MISSISSIPPIAN HARVESTMAN 483 i I ! Glasgow, United Kingdom (GLAHM A2854), where it is customarily on display. The spec- imen was digitally photographed using a Lei- ca DC 100 digital camera mounted on a Leica MZFLIII microscope and drawn using a cam- era iucida attachment. Adobe Photoshop Lim- ited Edition 5.0 was used to manipulate the digital images. For comparative purposes the type material of Nemastomoides longipes (Pe- trunkevitch 1913) from the Peabody Museum, Yale (YPM 171), M depressus (Petrunkevitch 1913) from the United States National Mu- seum (USNM 37974) and Kustarachne ten- uipes Scudder 1890 (USNM 37967) was ex- amined, along with Eophaiangium shear! from Rhynie (held in the University of Mun- ster, Germany) and extant species preserved in the Museum fur Naturkunde Berlin. Geological setting. — The East Kirkton Limestone is a fossil Konservat-Lagerstatte located near Bathgate, West Lothian, Scot- land; about 27 km west of Edinburgh. The limestone is the lowest of five which occur within the Bathgate Hills Volcanic Formation. This in turn can be correlated with the lower part of the Brigantian Stage of the Visean Se- ries of the Mississippian (= Lower Carbon- iferous in European terminology). The Bri- gantian Stage spans the time interval 336.0- 339.4 Ma. Further details can be found in Rolfe et al. (1994). The East Kirkton locality has also yielded scorpions (Jeram 1994, 2001), myriapods (Shear 1994) and a variety of terrestrial tetrapods including the anthra- cosaurid Silvanerpeton miripedes and the fa- mous stem-ameiote Westlothiana Uzziae; however early insects and arachnid groups like the extinct order Trigoeotarbida, which are usually fairly common in terrestrial de- posits of this age, are so far absent. Preservation.— GLAHM A2854 is pre- served as a compression fossil on the surface of a thin, grey bed of calcareous tuff. The light red-brown coloration of the harvestman is analogous to that of the scorpion fossils etched from the East Kirkton limestone (Jer- am 1994) and suggests that some constituent of the original cuticle still remains, rather than it being a wholly carbonized cuticle. This in turn suggests that chitieoclastic bacteria were excluded from the preservatioeaJ environment during the formation of the deposit and the arthropod fossils contained therein. The fossil originated from bed 82 (S.R Wood pers. comm. 2002). Rolfe et al. (1994) identified the combined thickness of bed 81 and bed 82 (with a lateral variation in thickness between 80 and 140 mm) as the same unit which Gei- kie (1861) named ‘bed bk This unit contains dense accumulations of ostracods on some bedding surfaces as well as charcoalified plant material. This is also the likely source bed of the holotype of Westlothiana Uzziae. Durant (1994) summarized the likely paleoeeviron- ment in which the sediments exposed in the East Kirkton Quarry were deposited in a shal- low lake close to the flanks of an active vol- cano. Volcanic eruption products including ash and tuffs, were eroded and washed down into the lake. Set against a backdrop of active volcanism in the area in which it formed, it is no surprise that ashy bands, pyroclastic frag- ments and chemical sediments influenced by hot spring activity dominate the sequence hosting both terrestrial and aquatic plant and animal fossils. Widespread development of limestone formed from stromatolitic algae also point to an unusual physio-chemical en- vironment for the formation of this deposit. In the fossil’s plane of compression at least four elongate and articulated legs are pre- served. A thin calcite vein cross-cuts the legs of the fossil (Figs. 1, 2). Close study of the harvestman body suggests that other legs may have broken off prior to preservation. How- ever, it is obvious that such delicate structures, which easily break off in extant animals even while still alive, indicate only a short period of transport into the preservational environ- ment. Perhaps this animal was rafted out onto the open lake on floating vegetation before dropping into the water? MORPHOLOGICAL INTERPRETATION Body . — The body (Figs. 1-3) is small and rounded. We suspect it is essentially a dorso- lateral to ventro-lateral compression in which the two sides of the body have become su- perimposed. As is typical for harvestmen, the prosoma and opisthosoma are broadly joined. Features such as eyes or mouthparts are equiv- ocal, even under higher magnification, but the body does come to a slight, angular point on the dorsal side where an eye tubercle might be expected. Low-angle lighting reveals lines on the harvestman body which might corre- spond to the original opisthosomal segmen- tation and/or elements of the coxae which 484 THE JOURNAL OF ARACHNOLOGY Figure 1. — Brigantibunum listoni new genus and species, from the Mississippian (Visean: Brigantian) of East Kirkton, Scotland. This modern-looking specimen is here tentatively assigned to the Eupnoi clade. Scale bar = 5 mm. come up the side of the body in many living harvestmen. These lines do not form a clear pattern running the length of the body and in similar-looking extant harvestmen seg- mentation is generally poorly dehned (e.g.. Shultz 2000) and may only be betrayed by folds or color patterns on the body surface. The fossil also reveals fine tuberculation on the body where cuticle has been preserved and, while there is some degree of locali- DUNLOP & ANDERSON— MISSISSIPPIAN HARVESTMAN 485 zation, the tubercles do not define identifi- able segments. OTipositor.- — ^Oee intriguing feature is an apparently annulate structure overlaying the left side of the body, which is only clearly visible under high magnification and low an- gle lighting (Fig. 3). It is associated with orig- inal cuticle which tends to suggest it is not an artefact, and at least towards the 'dorsaF end there are rows of tiny circular structures across each aenulation which might be tuber- cles or setal sockets. We tentatively interpret this structure as an ovipositor since its annu- late morphology and proportions relative to the body are suggestive of that in a modern eupnoid harvestman (e.g., Shultz 2000). The East Kirktoe scorpions preserve respiratory organs (Jeram 1994), which shows that there is the potential at this locality to recover in- ternal features. If this is an ovipositor, it is the second oldest record of internal genitalia, after the one recovered in Eophalangium sheari from Rhynie (Dunlop et al. 2003, 2004). It implies that the East Kirkton fossil is a female which probably laid its eggs in the substrate. In living harvestmen the ovipositor is appar- ently extended through hemolymph pressure (Shultz 2000). Perhaps the ovipositor in the fossil was squeezed out of the body during compression and came to lie across the opis- thosoma? Legs.— -Four almost complete legs are pre- served, all of which are very long and gracile; up to ca. twelve times the length of the body. A small fragment of either a fifth leg or, per- haps, a pedipalp is also preserved. It is diffi- cult to assign legs unequivocally to their se- quence in the body, but the longest leg (which is also the most gracile) is probably leg 2. This leg is longest in most living (non-cy- phophthalmid) harvestman and has a more tactile function. Indeed all 4 preserved legs express slightly different femur lengths (see Systematics) and this might indicate that the fossil is essentially a lateral view preserving legs 1--4 on one side of the body; with the corresponding legs from the other side either missing or still within the matrix. A tentative numbering scheme is proposed based on this assumption (Fig. 2). Discrete podomeres can be recognized, and the basic leg anatomy is a precise match for living eupnoid harvestmen (cf. Shultz 2000), The femur is elongate and slender. It is fol- lowed by a very short patella which is slightly thicker than the adjacent podomeres. It forms a distinct and bulbous 'kneek The tibia is again elongate and slender, widening distally to form a disjunct articulation with the next podomere,. the basitarsus. This basitarsus is also long and slender, although the transition to the telotarsus is indistinct. In many living harvestmen the telotarsus is composed of many short tarsomeres. These cannot be re- solved in the fossil, but the distal curvature of at least one of the legs (leg 1 in our scheme, see Figs. 1 & 2) implies that it, too, was formed from numerous articulating elements. Claws at the ends of the legs (the apotele) are equivocal. SYSTEMATIC PALAEONTOLOGY ?Eupeoi Hansen & Sprensee 1904 Remarks, — As noted by Selden (1993a), the East Kirkton fossil is clearly a harvestman, but explicit diagnostic characters of higher taxa within Opiliones are not clearly pre- served. Nevertheless, its overall morphology with a 4 mm globular body and long, essen- tially homogeneous legs is wholly inconsis- tent with Cyphophthalmi, which are tiny (typ- ically 1“2 mm) with short, stubby legs. Nor does it resemble the more robust Laniatores in which leg 4 is often enlarged and spiny (although not, for example, in oecopodids) and in which prominent, raptorial pedipalps would be expected. This leaves the Palpatores group, the monophyly of which is currently in dispute (compare Shultz 1998 and Giribet et al. 2002). The older Rhynie chert harvestmen preserve convincing eupnoid characters (Dun- lop et al. 2003, 2004a), thus both Eupnoi and its putative sister taxon lineage, Dyspnoi sen- su Shultz (1998) or (Dyspnoi + Laniatores) sensu Giribet et al. (2002), can be predicted from the Mississippiae. Both the Eupnoi and Dyspnoi clades, which make up the traditional Palpatores group, in- clude long-legged taxa. The extremely long and gracile legs in GLAHM A2854, which are up to about twelve times body length, tend to favor affinities with phalaegiid or scleroso- matid harvestmen (Eupnoi), for example members of common eupnoid genera like Leiobunum C.L. Koch 1839, Opilio Herbst 1798 and Phalangium Linnaeus 1758. Among the Recent Dyspnoi, common genera such as 486 THE JOURNAL OF ARACHNOLOGY Figure 2. — Interpretative drawing of the specimen shown in Fig. 1. Abbreviations: bs = basitarsus, fe = femur, pp = possibile pedipalp, pt = patella, ti = tibia, tt = telotarsus. Legs tentatively numbered from 1 to 4, with leg 2 longest. Scale bar = 5 mm. I Nemastoma C.L. Koch 1836 and Dicranolas- ma S0rensen 1873 typically have legs which are somewhat shorter in relation to body length. Members of the dyspnoid genus Mi- tostoma Roewer 1951 are closer to GLAHM A2854 in terms of leg lengths. Data from the fairly widespread M. chrysomelas (Hermann 1804) in Martens (1978, p. 143) suggests that leg length (13.7 mm) is, at best, about eight times body length (1.7 mm), although in a highly-specialized Alpine cave species like M. anopthalmum (Page 1946) leg length (22.8 mm) can be over fourteen times body length (1.6 mm); data from Martens (1978, p. 149). Clearly leg length is not an ideal character and as noted by Martens, these parameters can vary even within a species and between males and females. The putative ovipositor also hints at a eup- noid. If our interpretation is correct, this an- nulate morphology only occurs in Cypho- phthalmi (rejected for the reasons outlined above) and Eupnoi (Shultz 1998, 2000; Giri- bet et al. 2002). However, explicit autapo- morphies of Eupnoi cannot be resolved un- equivocally in the fossil. The lakeside paleoenvironment is unlikely to have trapped either high mountain and/or cave animals. The long and gracile legs in the fossil are more characteristic for certain phalangiid and scle- rosomatid eupnoids, as opposed to the usual range in non-specialist dyspnoids. On these grounds we tentatively assign the East Kirkton fossil to Eupnoi. Brigantibunum new genus Type and only species. — Brigantibunum listoni new genus and species. Etymology. — From the Brigantian age of the fossil and the suffix “bunum” used in modern genera of small bodied, long-legged harvestmen such as Leiobunum. DUNLOP & ANDERSON— MISSISSIPPIAN HARVESTMAN 487 Figure 3. — Detail of the body region under low angle lighting and immersion in alcohol. Abbreviations: ov? = possible annulate ovipositor, tb = tuberculation of cuticle. Leg sequence as in Fig. 2. Scale bar = 1 mm. i Diagnosis. — Extremely gracile Paleozoic harvestman with long, slender legs up to twelve times the length of the small, ovate body. Femora at least two and a half to three times the length of the body. Remarks. — The Eophalangium sheari ma- terial from the Devonian of Rhynie, Aber- deenshire, Scotland includes a long-legged specimen. Given the very different modes of preservation, direct comparisons with the East Kirkton fossil are problematic. The three-di- mensional Rhynie material yields many char- acters not testable in the East Kirkton speci- men. Furthermore, the (male) Rhynie fossil associated with the long legs is incomplete and the full extent of both the legs and body remains equivocal. Further discoveries may change this interpretation, but we can find no explicit autapomorphies or even reliable ratios of body proportions to argue that the East Kirkton fossil belongs to Eophalangium Dun- lop et al., 2004. Among the Pennsylvanian opilionids, our fossil is clearly not congeneric with the pu- tative Commentry troguloid Eotrogulus fayoli Thevenin 1901, which is a robust animal with an elongate body and comparatively short legs. This leaves three species of Kustarachne Scudder 1 890, two of which are rather incom- plete and doubtful, and three species of Ne- matostomoides Thevenin 1901; of which N. depressus from Mazon Creek is a misidenti- fied phalangiotarbid (Beall 1997; pers. obs.). The East Kirkton fossil appears longer-legged than all these Pennsylvanian forms, although the Mazon Creek specimens are in nodules and the full extent of the legs is, of course, not preserved. In detail, the length of the fe- mur offers a potential diagnostic character and the femora in the East Kirkton fossil are pro- 488 THE JOURNAL OF ARACHNOLOGY portionately longer (ca. 3 times body length) than the femora in Nemastomoides and Kus- tarachne (ca. 1-2 times body length). Overall, GLAHM A2854 is a unique find and the only record of a Mississippian harvestman. It most closely resembles the Commentry species N. elaveris Thevenin 1901, but based on its ex- treme leg length and gracile appearance we assign it to a new genus diagnosed on the body-femur ratio. Brigantibunum listoni new species Figs. 1, 2. ?Earliest known harvestman (Arachnida, Opili- ones): Wood et al. 1985: 355-356, fig. 1. Opilionid or harvestman: Smithson 1989: 676-678; Selden 1993a: 392-393; Selden 1993b: 305-306; Clack 1998: 66-69; Jeram 2001: 374, tabs. 16.1, 16.2; Dunlop & RoBler 2003: 389; Dunlop & Gi- ribet 2003: 371; Dunlop 2004a: 24; 2004b: 67. Type. — Holotype, from the East Kirkton Quarry, near Bathgate, (27 km west of Edin- burgh), West Lothian, Scotland (Grid refer- ence NS 991690), collected by Mr. Stan P. Wood, derived from Unit 82 of the East Kirk- ton Limestone, West Lothian Oil-Shale For- mation, Strathclyde Group, Upper Visean (Brigantian), Mississipian (= Lower Carbon- iferous in European stratigraphy) (GLAHM A2854). Etymology. — For Jeff Liston (University of Glasgow & Hunterian Museum). Diagnosis. — As for the genus. Description. — Body small, rounded, ca. 4 mm in diameter. Red-brown cuticle includes fine tuberculation and possible segment/coxal boundaries. Elongate, annulate structure (?ovipositor), length 2.5 mm, curves across body on left side. Small, incomplete limb el- ement (?pedipalp), length 2 mm, projects from body on right side. Four legs preserved, all long, slender and extremely gracile. All legs with different lengths and podomere propor- tions, tentatively numbered in sequence (see also Morphological interpretation) from lon- gest to shortest: 2 4 13. All legs slightly curved, one leg (probably leg 2) distinctly longer; maximum approximate preserved leg lengths as follows. Leg 1: 38 mm, leg 2: 51 mm, leg 3: 34 mm, leg 4: 40 mm. Femora long, lengths as follows. Leg 1: 12 mm, leg 2: 21 mm, leg 3: 11 mm, leg 4: 14 mm. All patellae short, ca. 1.5 mm. Tibiae longer, lengths as follows. Leg 1: 8 mm, leg 2: 14 mm, leg 3: 6 mm, leg 4: 9 mm. Basitarsi j lengths as follows. Leg 1: 8 mm, leg 2: un- ! clear, leg 3: 8 mm, leg 4: 9 mm. Telotarsi i incomplete, but long and slender. | DISCUSSION Brigantibunum listoni fits into a developing pattern (e.g., Dunlop et al. 2003, 2004; Dun- i lop 2004a, b) in which harvestmen appear to have evolved relatively early into recogniz- able crown-group forms (i.e. animals assign- able to clades with Recent representatives) and exhibit a high degree of stasis, with little fundamental change over hundreds of millions of years. Harvestmen with the same basic shape as the East Kirkton fossil, remain com- mon and abundant today; particularly in the northern hemisphere. To put this into context, the oldest crown-group spiders (Selden 1996) are mesotheles (the most basal living spider clade) and are first recorded from the end of the Pennsylvanian. This is some 100 million years after the oldest crown-group harvestmen from Rhynie, which can be assigned to the eupnoids; a somewhat derived clade. It is also worth mentioning that in recent arachnid phy- , logenies (e.g. Giribet et al. 2002) harvestmen seem to resolve in a fairly basal position (compared to spiders) as part of the so-called ' Dromopoda clade along with scorpions, pseu- doscoipions and solifuges. ACKNOWLEDGMENTS ! We thank Neil Clark and Jeff Liston (GLAHM), Tim White (YPM) and Mark Flor- ence (USNM) for access to material in their care, Stan Wood for information on the type locality and Maciek Kania for information on new Polish material. Gonzalo Giribet, Bill Shear and the editors provided helpful com- i ments on the manuscript during review. LITERATURE CITED Beall, B.S. 1986. Reinterpretation of the Kustar- ! achnida. American Arachnology 34:4. Beall, B.S. 1997. Arachnida. Pp. 140-154. In Rich- ardson’s Guide to the Fossil Fauna of Mazon Creek. (W. Shabica & A. A. Hay, eds). North- eastern Illinois University, Chicago. Clack, J.A. 1998. A new Early Carboniferous tet- rapod with a melange of crown-group characters. Nature 394:66-69. Cokendolpher, J.C. & G.O. Poinar, Jr. 1998. A new fossil harvestman from Dominican Republic am- ber (Opiliones, Samoidae, Hummelinckiolus). Journal of Arachnology 26:9-13. DUNLOP & ANDERSON— MISSISSIPPIAN HARVESTMAN 489 Dunlop, J.A. 2004a. The enigmatic fossil arachnid Kustarachne tenuipes Scudder, 1890 is a har- vestman. Pp, 17-25. In European Arachnology 2002 (F. Samu & Cs. Szinetar, eds). Proceedings of the 20* European Colloquium of Arachnolo- gy, Szombathely 22-26 July 2002. Plant Protec- tion Institute & Berzsenyi College, Budapest. Dunlop, J.A. 2004b. A spiny harvestman from the Upper Carboniferous of Missouri, USA. Pp. 67- 74, In European Arachnology 2003 (D.V. Logu- nov & D. Penney, eds). Arthropoda Selecta (Spe- cial Issue No. 1). Dunlop, J.A., L.I. Anderson, H, Kerp & H. Hass. 2003. Preserved organs of Devonian harvestmen. Nature 425:916. Dunlop, J.A., L.I. Anderson, H. Kerp & H. Hass. 2004, A harvestman (Arachnida: Opiliones) from the Early Devonian Rhynie cherts, Aberdeen- shire, Scotland. Transactions of the Royal Soci- ety of Edinburgh: Earth Sciences 94:341-354 Dunlop, J.A. & G. Giribet. 2003. The first fossil cyphophthalmid (Arachnida, Opiliones) from Bitterfeld amber, Germany. Journal of Arachnol- ogy 31:371-378. Dunlop, J.A. & R. RoBler. 2003. An enigmatic, so- lifuge-like fossil arachnid from the Lower Car- boniferous of Kamienna Gora (Intra-Sudetic Ba- sin), Poland. Palaontologische Zeitschrift 77: 389-400. Durant, G.P. 1994. Volcanogenic sediments of the East Kirkton Limestone (Visean) of West Loth- ian, Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences 84:203-207. Geikie, A. 1861. In The geology of the neighbour- hood of Edinburgh (H.H. Howell & A. Geikie, eds). Memoir of the Geological Survey of the United Kingdom. Giribet, G., G.D. Edgecombe, W.C. Wheeler & C. Babbit. 2002. Phylogeny and systematic position of Opiliones: a combined analysis of chelicerate relationships using morphological and molecular data. Cladistics 18:5-70. Hansen, H.J. & W S0rensen. 1904. On Two Orders of Arachnida. Cambridge University Press, Cam- bridge. Jell, PA. & PM. Duncan. 1986. Invertebrates, mainly insects, from the freshwater Lower Cre- taceous Koonwarra Fossil Bed (Korumburra Group), South Gippsland, Victoria. Memoirs of the Association of Australasian Palaeontologists 3:111-205. Jeram, A.J. 1994. Scorpions from the Visean of East Kirkton, West Lothian, Scotland, with a re- vision of the infraorder Mesoscorpionina. Trans- actions of the Royal Society of Edinburgh, Earth Sciences 84:283-299. Jeram, A.J. 2001. Palaeontology. Pp. 370-392. In Scorpion Biology and Research (R Brownell & G. A. Polls, eds). Oxford University Press, Ox- ford. Martens, J. 1978. Spinnentiere, Arachnida. Weber- knechte, Opiliones. Die Tierwelt Deutschlands 64:1-464. Petrunkevitch, A.I. 1913. A monograph of the ter- restrial Palaeozoic Arachnida of North America. Transactions of the Connecticut Academy of Arts and Sciences 18:1-137. Roger, J. 1946. Resultates scientifiques de la mis- sion C. Arambourg en Syrie et en Iran (1938- 39). Les Invertebres des Couches a Poissons du Cretace superieur du Liban — etude Palebiolo- gique des gisements. Memoires de la Societe Geologique de France. Paleontologie 51:60-64. Rolfe, W. D., G.P. Durant, W.J, Baird, C. Chaplin, R.L. Baton & R.J. Reekie. 1994, The East Kirk- ton Limestone, Visean, of West Lothian, Scot- land: introduction and stratigraphy. Transactions of the Royal Society of Edinburgh: Earth Sci- ences 84:177-188. Scudder, S.H. 1890. Illustrations of the Carbonifer- ous Arachnida of North America. Memoirs of the Boston Society of Natural History 4:443-456. Selden, PA. 1993a. Fossil arachnids — recent ad- vances and future prospects. Memoirs of the Queensland Museum 33:389-400. Selden, P.A. 1993b. Arthropoda (Aglaspidida, Pyc- nogonida and Chelicerata). Pp. 297-320. In The Fossil Record 2 (M.J. Benton, ed.). Chapman and Hall, London. Selden, P.A. 1996. First fossil mesothele spider, from the Carboniferous of France. Revue suisse de Zoologie hors serie:585-596. Shear, W.A. 1994. Myriapodous arthropods from the Visean of East Kirkton, West Lothian, Scot- land. Transactions of the Royal Society of Ed- inburgh: Earth Sciences 84:309-316. Shultz, J.W. 1998. Phylogeny of Opiliones (Arach- nida): an assessment of the “Cyphopalpatores” concept. Journal of Arachnology 26:257-272. Shultz, J.W. 2000. Skeletomuscular anatomy of the harvestman Leiobunum aldrichi (Weed, 1893) (Arachnida: Opiliones: Palpatores) and its evo- lutionary significance. Zoological Journal of the Linnean Society 128:401-438. Smithson, TR. 1989. The earliest known reptile. Nature 342:676-678. Stargga, W. 2002. Baltic amber harvestmen (Opili- ones) from Polish collections. Annales Zoologici 52:601-604. Thevenin, A. 1901. Sur la decouverte d'arachnides dans le Terrain Houiller de Commentry. Bulletin de la Societe Geologique de France, Quatrieme Serie 1:605-611. Wood, S.R, A.L. Panchen & TR. Smithson. 1985. A terrestrial fauna from the Scottish Lower Car- boniferous. Nature 314:355-356. Manuscript received 7 October 2004, revised 5 April 2005. 2005. The Journal of Arachnology 33:490-500 A REVISION OF THE SPIDER GENUS TAURONGIA (ARANEAE, STIPHIDIOIDEA) FROM SOUTH-EASTERN AUSTRALIA Michael R. Gray: Australian Museum, 6 College Street, Sydney, New South Wales 2010, Australia. E-mail: mikeg@austmus.gov.au ABSTRACT. The spider genus Taurongia Hogg 1901 and the species T. punctata (Hogg 1900) are redescribed. Taurongia punctata is shown to be a rather variable species with a widespread distribution across the eastern central Victorian highlands. Taurongia punctata is a robust spider, contrasting with a more gracile new species, T. ambigua, described from the western Victorian highlands. The placement of the latter in Taurongia is provisional and may change once other undescribed "Taurongia group’ genera from eastern Australia have been examined. The Taurongia species dealt with here differ from the latter taxa in having an increased number of cylindrical spigots and a large palpal median apophysis. Keywords: Taxonomy, cribellate, new species Hogg (1900) described two ‘dictynid’ spi- ders from central Victoria under the name Hy- lohius. This name (preoccupied in Coleoptera) was subsequently replaced by Taurongia (Hogg 1901). Lehtinen (1967) characterized the genus, figuring T. punctata (Hogg 1900), and placed it in his Desidae, Desinae. Forster (1970) noted that the available data were in- sufficient for accurate placement of Taurongia in his concept of the Desidae. Taurongia has long been confused with related "Taurongia group’ taxa that are widely distributed in east- ern Australian forests from Tasmania to Queensland. These taxa comprise several un- described genera that are morphologically di- verse but united by characteristics of the gen- italia, notably the palpal tegular structure and the placement of the median apophysis (re- duced to a slender, spine-like process in most taxa, except Taurongia). Here, the genus Tau- rongia is reviewed as a step toward charac- terizing this group of spiders and clarifying their relationships. METHODS Specimen examinations, measurements and drawings were made using a Wild M5 or Lei- ca M12 microscope with graticule and draw- ing attachment. Epigynal preparations were cleared in lactic acid, before mounting in glycerol for microscopic examination. The left side male palp is illustrated. Specimen prep- arations for scanning electron microscopy were taken through 80-100% alcohol stages, 100% acetone and then air dried. Abbreviations and definitions. — “Tegular window” refers to the gap between the prox- imal embolus and the basal part of the con- ductor. BE = body length; CL = carapace length; CW = carapace width; CapW = caput width; LL = labium length; LW = labium width; SL = sternum length; SW = sternum width; EGW = eye group width; AME = an- terior median eyes; ALE = anterior lateral eyes; PME = posterior median eyes; PLE = posterior lateral eyes; MOQ = median ocular quadrangle; RTA = retrolateral tibial apoph- ysis; RVTA = retroventral tibial apophysis; MA = median apophysis; ALS = anterior lat- eral spinneret; PMS = posterior median spin- neret; PLS = posterior lateral spinneret; MAP = major ampullate spigot; mAP = minor am- pullate spigot; Cyl = cylindrical spigots; Pc = paracribellar spigots; mPLS = modified PLS spigot; n = nubbin. The material exam- ined in this study is lodged in the following repositories: AM = Australian Museum, Syd- ney; NHM = Natural History Museum, Lon- don; NMV = Museum of Victoria, Mel- bourne; WAM = Western Australian Museum, Perth. TAXONOMY Superfamily Amaurobioidea Thorell 1870 Taurongia Hogg 1901 Hylobius Hogg 1900: 82 (preoccupied by Hylobius Germar 1817). 490 gray— GENUS TAURONGIA IN AUSTRALIA 491 Figures 1-9. — Taurongia species: 1-4, 8. T. punctata (Hogg); 5-7, 9. T. ambigua new species. 1, 5. Carapace; 2, 6. Sternum and mouthparts; 3, 4, 7. Abdomen, 3, 7. dorsal; 4. ventral; 8, 9. Eyes, anterodorsal. Scale bars: 1 mm (Figs. 1-4, 5-7); 0.5 mm (Figs. 8, 9). Taurongia Hogg 1901: 278 (replacement name); Lehtinen 1967: 267, 326; Platnick 2004. Hylobihoggia Strand 1935: 304 (superfluous re- placement name); Lehtinen 1967: 267; Platnick 2004. Type species. — Hylobius divergens Hogg 1900 by original designation, currently a ju- nior synonym of Hylobius punctatus Hogg 1900. Comment on synonomy of type spe- cies.— Hogg (1900) described two species in his genus Hylobius and designated H. diver- gens as the type. In 1901 he replaced the pre- occupied generic name with Taurongia. Hogg’s material came from the Macedon Dis- trict in Victoria. It comprised the female ho- lotype of Taurongia divergens and the male and female syntypes of T. punctata, although Lehtinen (1967) noted that the female syntype was a juvenile specimen. Lehtinen (1967) placed T. divergens, the type species, in syn- onymy with T. punctata, presumably because the male T. punctata specimen provided the better character set. Subsequent collecting by the author has not revealed a second species of Taurongia in the Macedon District. Diagnosis. — Cribellate spiders with a ro- bust or gracile body form. Carapace dark brown without obvious patterning. Male palp: cymbium digitiform, spinous; median apoph- ysis large, spatulate; RTA large, ventrad. Epi- gynum divided by a median septum, or sep- tum indistinct. PMS with 4-6 cylindrical spigots. Redescription. — Medium-large robust or gracile cribellate spiders. Carapace, jaws and legs dark reddish brown, anterior caput and jaws darkest; lateral carapace with dark radial streaks from fovea; legs not banded. Dorsal and lateral abdomen dark brownish-grey, with more or less distinct pallid chevrons dorsally, the 3 anterior pairs with small sigillae; venter grey, bounded laterally by pallid, dotted lines. 492 THE JOURNAL OF ARACHNOLOGY Figure 10. — Distribution of Taurongia punctata (closed squares) and T. ambigua (open circle). with two lines of paired spots medially (Figs. 1-4, 7). Body and leg hairs plumose, feathery hairs absent. Carapace with prominent caput; pro- file moderately arched, highest in mid-caput region; foveal slit moderately long and deep, curving down onto concave rear slope of car- apace (Figs. 1, 5). Clypeus wide, ca. 3 X width of an AME, anterior margin strongly convex. Chilum an undivided, median plate. Eyes eight, AME or PME smallest (Figs. 8, 9). Eye group moderately narrow, EGW ca. 0.50-0.60 X width of caput; eyes in two rows, from above AER recurved, PER procurved- straight; MOQ longer than wide, slightly nar- rower anteriorly. PME, PLE and ALE with ca- noe-shaped tapeta. Cheliceral paturon robust, proximally kneed, with large boss; fangs strong, short; fang groove short (Fig. 2, T. punctata); paturon and fangs more gracile and longer in T. ambigua (Fig. 6). Two adjacent retromarginal teeth (sometimes set in paler area of cuticle); and 3 adjacent promarginal teeth, last tooth extended as a strong carina; retromargin with one long modified seta near base of fang, several modified setae above promargin. Maxillae broad, longer than wide, lateral margins convex with a strong antero- lateral linear serrula. Labium longer than wide, widest anterior to baso-lateral excava- tions, narrowing to a weakly concave apex. Sternum cordate, longer than wide, shortly to strongly pointed between coxae 4 (Figs, 2, 6). Legs 1423, with inclined and vertical hairs. Trochanters slightly to strongly notched. Ret- rocoxal hymen absent. Three tarsal claws, su- perior 9-11 teeth, inferior 2-3 teeth; claw tufts and scopulae absent but ventral tarsi and metatarsi 1-4 strongly hirsute in T. punctata, much less hirsute in T. ambigua. Female pal- pal tarsi spinose; palpal claw with 11-12 teeth. Trichobothria increasing in length dis- tally, in single row on tarsi (6-7) and meta- tarsi (5-6); two rows on tibia; present on male and female palpal tarsus and tibia. Bothria collariform, proximal plate with weak to well defined longitudinal ridges (Figs. 31, 35). Tar- sal organ placed distal to trichobothria, cap- sulate with an ovoid, more or less key-hole shaped pore (Figs. 32, 36). Calamistrum: ca, 0,4 X length of metatarsus, subcentrally to proximally placed, with a dorsally contiguous field of recumbent setae; delimited at each end by a retrodorsal spine. Male palp (Figs. 11, 12, 37, 38): Cymbium with a digitiform apex with several bristles and spines. Bulb subcircular to ovoid. Tegu- lum with a narrow prolateral-basal sclerotised region within which the sperm duct runs in an ovoid loop, and from which the conductor gray— GENUS TAURONGIA IN AUSTRALIA 493 Figures 11-17. — Taurongia punctata. 11 — 15. Male palp. 11, 12. Palp, ventral, retrolateral; 13. Bulb, ventral (Mt Donna Buang); 14. Median apophysis (Mt Buller); 15. Tibia, retrolateral (Woodend). 16, 17. Epigynum, ventral, dorsal. Scale bar: 0.25 mm ! arises anteriorly; and a large retrolateral-basal membranous region from which the MA aris- I es basally. MA usually large, membranous and hyaline, often ‘spatula- shaped’; less com- monly reduced in size (Fig. 14). Embolus spi- I niform, curving in a semicircle from its pro- ! lateral tegular origin around the conductor i margin. Small tegular window present at em- I bolus/conductor base. Conductor weakly T- I shaped, with a short, membranous stalk sup- porting a semicircular-falciform head with a marginal embolic groove, narrowing retro- distally as a reflected or elongate tip (Figs. 11, 37). Tibia about as long as wide, with two distal apophyses, the RVTA and a ventrad placed RTA. Patella about as long as wide with a dorsal bristle. Epigynum: Fossa divided by a distinct me- diae septum expanding posteriorly into a pos- terior lobe (Fig. 16), or fossa open and septum indistinct (Fig. 39). Lateral teeth absent. Cop- ulatory ducts narrow, very short or simply 494 THE JOURNAL OF ARACHNOLOGY coiled, opening postero-laterally (Figs. 17, 41). Paired spermathecae ovoid, well separat- ed, placed lateral to copulatory duct openings at posterior end of fossa. Tracheal system simple, with four un- branched tracheal tubes confined to the ab- domen. Spiracle just anterior to cribellum, about 0.4 X as wide as cribellum plate (Fig. 18). Spinnerets: PMS with 4-6 cylindrical spigots (Fig. 20). Included species. — Taurongia punctata (Hogg), T. ambigua new species. Comments. — The new species described here, T. ambigua, is attributed to Taurongia largely on the basis of its similarities to the type species in genitalic and spinneret char- acters (similar cymbial, tegular and MA struc- ture; relatively simple copulatory ducts; in- creased numbers (4-6) of cylindrical spigots on PMS (only 1 or 2 spigots present in related taxa (pers. obs.)). However, there are also sig- nificant differences in body build, eye sizes, trochanteral notches and cuticular sculpturing, which make the placement of T. ambigua in Taurongia provisional. Taurongia punctata (Hogg 1900) Figs. 1-4, 8, 10, 11-23, 34-36 Hylobius punctatus Hogg 1900: 84, plate XII, fig. 3. Hylobius divergens Hogg 1900: 82, plate XII, fig. 2. Taurongia punctata (Hogg): Hogg 1901: 279; Leh- tinen 1967: 267, figs. 122, 123, 127; Platnick 2004. Taurongia divergens (Hogg): Hogg 1901: 279; Lehtinen 1967: 267 (placed in synonymy with T. punctata). Hylobihoggia divergens (Hogg): Strand 1935: 304. Hylobihoggia punctata (Hogg): Strand 1935: 304. Type material. — Hylobius punctatus: lec- totype (present designation) male (examined), 1 paralectotype juvenile female [not exam- ined, noted by Lehtinen (1967)], Macedon District, Victoria, Australia, H.R. Hogg (NHM). Hylobius divergens: Holotype female (ex- amined), Macedon District, Victoria, Austra- lia, H.R. Hogg (NHM). Other material. — AUSTRALIA: Victoria: 1 d, 1 $, Sanitorium Picnic Ground Lake, Mt Macedon, 37°23'S 144°35'E, irregular sheet web on rotting log with egg sac, 23 February 1996, M. & A. Gray (AM KS45401); 1 $ same data as KS45401 (AM KS45403); 1 $, same data as KS45401 (no egg sac) (AM KS45402); 1 $, same data as KS45402 (AM KS45404); 1 $, Mt Disappointment area, 37°26'S 145°08'E, 26 August 1973, M. Gray (AM KS34511); $, Warburton, 37°45'S 145°42'E, 8 September 1959, A. Neboiss (AM KS34512); 2 $ , Mt Macedon, 37°23'S 144°35'E, rainforest, in rotting log, 14 March 1970, M. Gray (AM KS34513-14); 1 9, Blue , Range Rd, 13 km S. of Thornton, 37°19'S 145°5LE, 9 April 1978, M.R. Gray (AM I KS88187); 1 d, 1 9, 1 juvenile, Rubicon State Forest, 13 km S. of Thornton on Roys- ton Rd via Rubicon, 37°19'S 1455 LE, 7 April | 1978, M.R. Gray (AM KS88180); 2 d, data f as for AM KS88180, collected as juveniles, : matured July and August 1978 (AM j KS88181-2); 1 d, 1 9, 7.5 km SE. of Wood- : end on Mt Macedon Road, 37°28'S 144°37'E, in log, 4 April 1978, M.R. Gray (abdomen of AM KS88176 used for SEM) (AM KS88176- I 77); 1 9, data as for AM KS88176 except 6 April 1978 (AM KS88179); 2 9, Omeo High- f way 52 km N. of Omeo between Glen Wills i and Sunnyside, 36°50'S 147°3LE, 13 April ; 1978, M.R. Gray (AM KS88 188-89); 2 9,3 I km E. of Mirimbah on Mt Stirling Rd, 37°06'S 146°27'E, 8 April 1978, M.R. Gray i (AM KS88 183-84); 1 9, 7 km E. of Mirim- ; bah on Mt Stirling Rd, 37°09'S 146°29'E, 920 j m, irregular sheet web leading to retreat in crevice in bank, 8 April 1978, M.R. Gray (AM KS88185); 1 d. Box Corner, 4.5 km N. | of Mt Buller Village, 37°07'S 146°26'E, 8 i April 1978, 1000 m, M.R. Gray (AM | KS88186); 1 d. Central Highlands, Forestry , Rd 26, 0.2 km WNW. of Donna Buang Rd , junction, 37°43WS 145°39"30"E, flight inter- cept trap, Eucalyptus forest, 21 January-7 April 1995, G. Milledge (NMV K6557); 1 d, ^ Central Highlands, 0.7 km N of Acheron Gap, ! 7 km N. of Mt Donna Buang, 37°40'17"S 145°44'20"E, pitfall trap, Eucalyptus forest, 28 December 1995-21 February 1996, G. Mil- j ledge (NMV K6558); 1 9, The Beeches, ' 37°28'S, 145°49'E, 25 May 1991, M.S. Har- t vey, M.E. Blosfelds (WAM 98/2049); 1 9, ' Cumberland Falls, 37°34'S, 145°53'E, under ' log, 27 May 1991, M.S. Harvey, M.E. Bios- felds (WAM 98/1995). ' Diagnosis. — Differs from T. ambigua by its more robust build, smaller eye size, absence of deep trochanteral notches, male palp with conductor apex strongly curved and epigynal fossa divided by a median septum. gray— GENUS TAURONGIA IN AUSTRALIA 495 Figures 18-23. — Taurongia punctata, spinnerets (female). 18. Spinneret field; 19. ALS (LHS); 20. PMS (RHS); 21. PMS, anterior area; 23. PLS (LHS), 24. mPLS and Pc spigots on apical PLS. 496 THE JOURNAL OF ARACHNOLOGY Description. — Male (Mt Macedon, AM KS45401): BL 9.54, CL 5.67 (range 5.20- 5.67), CW 3.87, CapW 2.98, EGW 1.66, LL 0.97, LW 0.79, SL 2.94, SW 2.29. Body ro- bust, caput wide (Fig.l). Eyes: smaller and eye group narrower (ca. 0.5 caput width) than in T. ambigua; PME smallest, ALE > AME > PLE > PME; AME weakly protuberant on a low common tubercle (Fig. 8). Sternum cor- date, moderately long and shortly pointed pos- teriorly (Fig. 2). Legs: robust, relatively short, 1423 (I: 17.00; II: 15.07; III: 12.87; IV: 15.87); ratio tibia I length:CW = 1:0.93. Ven- tral tarsi and metatarsi moderately hirsute. Trochanters 1, 2 unnotched, 3, 4 slightly notched. Spination: I: femur dl- 1-0-2, pO-0-2- 0; patella 0; tibia v2-2-2, pl-1-1-0, rl- 1-1-0; metatarsus dO-0-2, v2-2-l, pO- 1-0-1, rO- 1-0-1; tarsus 0; II: femur dl -2-0-2, pO- 1-0-1; patella 0; tibia v2-2-2, pl-1-1-0, rl -1-1-0; metatarsus dO-1-2, V2-2-1, pl-1-0-1, rl-1-0-1; tarsus 0; III: femur dl -2-0-2, pO- 1-0-1; patella 0; tibia dO-0- 1-0-0, V 1-2-2, pi- 1-0- 1-0, rl-1-0-1-0; metatarsus dO-1-2, v2-2-l, pl-1-1, rl-1-1; tar- sus 0; IV: femur dl- 1-0-2, pO- 1-0-1; patella 0; tibia vl-1-2, pi -1-0- 1-0, rl-1-0-1-0; metatar- sus dO-1-2, v2-2-l, pl-1-1, rl-1-1; tarsus 0. Male palp (Figs. 11-15): distal conductor strongly tapered and curved to a short, spine- like apex; MA usually large and prominent; RTA a large ventrad, rectangular plate; RVTA thick, peg-like. Female (Mt Macedon, AM KS45402): BL 14.47, CL 7.27 (range 5.36-7.27), CW 4.87, CapW 4.00, EGW 2.07, LL 1.17, LW 0.97, SL 3.49, SW 2.76. Body, eyes and legs similar to male. Legs: 1423 (I: 17.73; II: 15.73; III: 13.40; IV: 16.53); ratio tibia I length to CW = 1:0.90. Cuticle surface smooth to weakly ridged. TO capsule smooth, pore ovoid with a short, narrow slit proximally (Fig. 36). Spi- nation: I: femur dl -2-0-2, pO-0-2-0; patella 0; tibia v2-2-2, pO-1- 1-1-0, r 1-0- 1-0- 1-0-0; meta- tarsus dO-0-2, V2-2-1, pO- 1-1 -0-1, rO-0- 1-0-1; tarsus 0; II: femur d 1-2-2, pO-1-1; patella 0; tibia v2-2-2, pl-1-1-0, rl- 1-1-0; metatarsus dO-O-2, V2-2-1, pl-1-0-1, rl-1-0-1; tarsus 0; III: femur d 1-2-2, pO-1-1; patella 0; tibia dl- O-l-O-O, vl-1-2, pl-1-0-1-0, rl-1-0-1-0; meta- tarsus dO-1-2, v2-2-l, pl-1-1, rl-1-1; tarsus 0; IV: femur dl- 1-0-2, pO- 1-0-1; patella 0; tibia vl-1-2, pl-1-1-0, rl-1-1-0; metatarsus dl-2-2, v2-2-l, pl-1-1, rO-0-1; tarsus 0. Epigynum (Figs. 16, 17): fossa divided by a moderately wide and arched median septum, becoming wider and lobe-like posteriorly; copulatory duct openings postero-lateral, ducts very short I' and narrow, entering the spermathecae antero- ? medially; spermathecae ovoid, well separated. Spinning organs (Figs. 18-23): cribellar plate bipartite, each field about a quarter as wide as long and separated by a narrow seam (about i 0.1 X of a field length); seam and posterior plate margin sclerotized (in male, cribellum almost as wide as in female but with non- functional fields). Spinnerets short, ALS = PLS, PMS shortest; ALS broad, very short ' apical segment with wide margins; PLS slen- : der with longer, conical apical segment. Spig- ots: ALS: 2 MAP spigots, mesal, adjacent, un- j equal; ca. 100 piriform spigots; PMS: 1 mAP with 4 fused paracribellar bases antero-ectally adjacent [5, 5, 2, 2 spigots respectively (2, 2 spigots basally fused only)]; 6 aciniform spig- ots (1 anterior, rest distributed); 6 cylindrical spigots; PLS: ca. 16 aciniform spigots, dis- tributed; 1 subapical “modified PLS” spigot with 1 paracribellar spigot, and 1 nubbin al- most entirely fused to side of mPLS; 7-8 cy- ‘ lindrical spigots (basal to subapical). Variation. — Given the distribution of this species across disected highland forest terrain, it is not surprising that considerable morpho- logical variation is encountered. In females, the epigynal septum may be moderately wide (Fig. 16) or much narrower. The loops of the sperm duct on the tegulum may be open or closed and vary in size. The MA is usually large and obvious but its size and shape vary (Figs. 11, 13); some reduction is evident in a male specimen from the Mt Buller region (Fig. 14), but the specimen is badly damaged and more material is needed to check its spe- cific status. Distribution. — Central Highlands of Vic- toria (Southern Great Dividing Range) from the Mt Buller region to Warburton and west to the Macedon region. Biology. — Cribellate sheet webs associated with logs, rocks and soil banks, guyed out with coarse retreat threads; sheet tapers to a variably defined retreat funnel ending inside a log or rock cavity or in a shallow soil burrow. Spiders run underneath sheet. Two egg sacs made of fine white flocculent silk (AM KS45401 and KS45403) were observed at Mt Macedon in February 1996. They were both found within cavities in rotting logs near the gray— GENUS TAURONGIA IN AUSTRALIA 497 Figures 24-36. — Taurongia species: 24 — 30. T. ambigua new species, spinnerets (female). 24. Cribel- lum; 25. MAP spigots, ALS (RHS); 26. ALS (LHS); 27. PMS (LHS); 28. Paracribellar spigots, PMS; 29. PLS (LHS); 30. mPLS spigot and nubbins (n) on apical PLS. 31-36. Sensilla (tarsus 1). 31-33, T. ambigua new species: 31. Trichobothrium base; 32. Tarsal organ; 33. Cuticular patterning and ovoid sensillae. 34- 36, T. punctata. 34. ventral tarsal hairs; 35. Trichobothrium base; 36. Tarsal organ. 498 THE JOURNAL OF ARACHNOLOGY base of their respective retreat funnels. Each sac was ca. 1cm in diameter, circular in plan, curved above but flatter below and suspended within a network of strong threads attached to the retreat walls. Taurongia ambigua new species Figs. 5-7, 9, 10, 24-33, 37-41 Type material. — AUSTRALIA: Victoria: Holotype d, 12 km from Halls Gap on Vic- toria Valley Road, Grampians Range, 3708 'S 142°31'E, under log in small, irregular sheet web, tall open forest, 26 March 1974, M.R. Gray (AM KS45501). Paratypes: 1 $, same data as holotype (AM KS5292); 3 9, same data as holotype except 27 March 1974 (AM KS5290, KS88171-72); 1 9, same data as KS5290 except web under rock in gully (AM KS45502). Other material. — AUSTRALIA: Victoria: 1 9, same data as AM KS5290, abdomen used for SEM (AM KS88173). Etymology. — A reference to the uncertain generic placement of this species. Diagnosis. — Differs from T. punctata by its gracile build (narrow caput, slender legs), rel- atively larger eyes with AME smallest, deeply notched trochanters, undivided epigynal fossa, elongate retrolateral conductor limb and strong cuticular sculpturing. Description. — Male (holotype): BE 6.77, CL 3.75, CW 2.65, CapW 1.45, EGW 0.88, EL 0.59, LW 0.54, SL 1.78, SW 1.56. Gracile cribellate spiders. Carapace amber-brown, darkest at caput; patterning restricted to radi- ating darker lines from fovea. Abdomen dark grey-brown with dark anterodorsal stripe flanked by 5 pallid patches smallest posteri- orly, and 3 pairs of pallid lateral stripes (Fig. 7). Carapace weakly arched. EGW ca. 0.6 X width of caput. Eyes normal size (relatively larger than in T. punctata), AME smallest: ALE > PEE > PME > AME (Fig. 9). Jaws vertical, boss small. Sternum cordate, extend- ing posteriorly between coxae 4. Legs: rela- tively long, slender, 1423 (I: 18.73; II: 14.55; III: 12.47; IV: 15.60); ratio tibia I length:CW = 1 :0.54. All trochanters deeply notched. Hairs plumose, most inclined, a few vertical; feathery hairs absent; metatarsi and tarsi not ventrally hirsute; U' and 2"^* metatarsi and tarsi with many long, curled hairs. Spination: I: fe- mur dl -2-0-2, pO- 1-1-1; patella 0; tibia v2-2- 2, pi -1-0- 1-0, rO- 1-0- 1-0; metatarsus dO-0-2, v2-2-l, pO- 1-0-1, rO-1-1; tarsus 0; II: femur dl -2-0-2, pO- 1-1-1; patella 0; tibia dO-0-1-0, v2-2-2, pO- 1-0- 1-0, rO- 1-0- 1-0; metatarsus d2- 2-2, V2-2-1, pO-0-1, rO-0-1; tarsus 0; III: femur d 1-2-2, pO-1-1; patella d 1-0-1; tibia d 1-0- 1-0, v2-2-2, pO- 1-0- 1-0, rO- 1-0- 1-0; metatarsus d2- 0-2, v2-2-l, pO- 1-0-1, rO- 1-0-1; tarsus 0; IV: femur dl-1-2, pO-0-1; patella dO-0-1; tibia dl- 0-0- 1-0, V2-2-2, pO- 1-0- 1-0, rO- 1-0- 1-0; meta- tarsus d2-l-2, v2-2-l, pO- 1-0-1, rO-0-1; tarsus 0. Male palp (Figs. 37, 38): conductor T- shaped with a slender, elongate retrodistal spine. TW present, larger than in T. punctata. Large ‘spatulate’ MA, membranous, partly hyaline. Large ventrad RTA with pointed apex, with a smaller, blunt RVTA arising at its base. Female (Grampians Range, AM KS5292): BE 7.35, CL 3.64 (range 2.85-3.68), CW 2.47, CapW 1.56, EGW 0.86, LL 0.61, LW 0. 54, SL 1.82, SW 1.55. Legs lacking numer- ous curled hairs of male and ventral metatarsi III, IV more hirsute distally. Otherwise body, eyes and legs similar to male. Legs: 1423 (I: 13.71; II: 11.56; III: 10.55; IV: 12.80); ratio tibia I length:CW = 1:0.71. Tarsal claws: su- perior, 5-7, inferior, 0-2; with 2-3 curved sustentacLilar hairs. Cuticle strongly sculpted with a closely ridged surface pattern, inter- rupted by numerous ovoid, plaque-like puta- tive sensilla with small distal pores (Fig. 33). TO capsule finely ridged with a narrow, ovoid, keyhole-like pore (Fig. 32). Spination: I: femur d 1-2-2, pO-1-1-1; patella 0; tibia v2- 2-2, pi- 1-0- 1-0, rO- 1-0- 1-0; metatarsus do-0- 2, v2-2-l, pO- 1-0-1, rO- 1-0-1; tarsus 0; II: fe- mur d 1-2- 1-2, pO-1-1-1; patella 0; tibia v2-2-2, pO- 1-0- 1-0, rO- 1-0- 1-0; metatarsus dO- 0-2, v2-2-l, pi -1-0-1, rO- 1-0-1; tarsus 0; III: femur dl -2-0-2, pO- 1-1-1; patella dO-0-1; tibia dl-0-1-0, V2-2-2, pO-1-0-1-0, rO-1-0-1-0; metatarsus d2-l-2, v2-2-l, pO- 1-0-1, rO- 1-0-1; tarsus 0; IV: femur d 1-0- 1-0-2, pO- 1-0-1; pa- tella 0; tibia d 1-0-0- 1-0, v2-2-2, pO- 1-0- 1-0, rO- 1-0- 1-0; metatarsus d2-l-2, v2-2-l, pO-1-0- 1, rO-0-1; tarsus 0. Calamistrum short, less than 0.25 X length of metatarsus, in 2"^^ prox- imal quarter of metatarsus IV. Epigynum (Figs. 39-41): epigynal fossa open, without a distinct septum but with a median seam or low ridge. Copulatory duct openings postero-lat- eral, ducts longer and wider than in T. punc- tata and curved around spermathecae. Spin- ning organs: (Figs. 24—30). PLS very slender gray— GENUS TAURONGIA IN AUSTRALIA 499 Figures 37-41. — Taurongia ambigua new species: 37, 38. Male palp, ventral and retrolateral; 39, 40, 41. Epigynum, ventral, lateral and dorsal. Scale bar: 0.25 mm. and shorter than ALS. Cribellar plate bipartite, each field about a third as wide as long and separated by a wide seam (about 0.3 X a field j length); Spigots: ALS: 2 MAP spigots, mesal, ' adjacent, unequal; 50-60 piriform spigots; ! PMS: 1 mAP; 1 fused paracribellar base (5-7 spigots); 4 aciniform spigots (1 anterior); 4 cylindrical spigots; PLS: 12 aciniform spigots, distributed; 1 subapical “modified PLS” spig- I ot; paracribellar spigots absent but at least 2 ! nubbins present flanking mPLS; 3 cylindrical j spigots. Distribution. — Recorded only from the type locality. , RELATIONSHIPS OF TAURONGIA i An analysis of the relationships of amau- I robioid spiders with grate- shaped eye tapeta I [exemplified by the genera Borrala Gray & Smith 2004 and Pillara Gray & Smith 2004 (Gray & Smith 2004)] suggests that Tauron- gia lies at the base of a group of genera in- cluding Stiphidion Simon 1902 and Wabua Davies 2000, which in turn is basal to the “grate-shaped tapetum” genera. This data (in preparation) suggests that Taurongia may be associated with the Stiphidioidea of Griswold et al. (1999). ACKNOWLEDGMENTS I am grateful to Peter Lilywhite (Museum of Victoria) and Mark Harvey (Western Aus- tralian Museum) for the loan of material from their respective collections. Paul Hillyard (Natural History Museum) kindly provided me with access to the type material of Tau- rongia. The genitalic illustrations were made by Helen Smith. LITERATURE CITED Forster, R.R. 1970. The Spiders of New Zealand. Part III. Otago Museum Bulletin 3:1-184. Gray, M.R. & H.M. Smith. 2004. The “striped” group of stiphidiid spiders: Two new genera from northeastern New South Wales, Australia (Araneae: Amaurobioidea: Stiphidiidae). Rec- ords of the Australian Museum 56:123-138. 500 THE JOURNAL OF ARACHNOLOGY Griswold, C.E., J.A. Coddington, N.I. Platnick & R.R. Forster. 1999. Towards a phylogeny of en- telegyne spiders (Araneae, Araneomorphae, En- telegynae). Journal of Arachnology 27:53-63 Hogg, H.R. 1900. A contribution to our knowledge of the spiders of Victoria: including some new species and genera. Proceedings of the Royal So- ciety of Victoria (N.S.) 13:68-123. Hogg, H.R. 1901. On Australian and New Zealand spiders of the suborder Mygalomorphae. Pro- ceedings of the Zoological Society of London 1901:218-279. Lehtinen, RT. 1967. Classification of the cribellate spiders and some allied families, with notes on the evolution of the suborder Araneomorpha. Annales Zoologici Fennici 4:199-468. Platnick, N.I. 2004. The world spider catalog, ver- sion 5.0. American Museum of Natural History, online at http://research.amnh.org/entomology/ spiders/catalog/index. html. Strand, E. 1935. Miscellanea nomenclatorica zool- ogica et palaeontolgica, VII. Folia Zoologica et Hydrobiologica 7:300-306. Manuscript received 31 December 2004, revised 3 June 2005. 2005. The Journal of Arachnology 33:501-508 REVISION OF SPIDER TAXA DESCRIBED BY KYUKICHI KISHIDA: PART 1. PERSONAL HISTORY AND A LIST OF HIS WORKS ON SPIDERS Hirotsugu Ono: Department of Zoology, National Science Museum, Tokyo, 3=23-1 Hyakunin-cho, Shinjuku-ku, Tokyo, 169-0073 Japan. E-mail: ono@kahaku.go.jp ABSTRACT. The personal history of forgotten Japanese arachnologist, Kyukichi Kishida (1888-1968) is described for the first time based on information collected from the literature and through interviews with the late Prof. Seikichi Kishida (1931-2002), the fourth son of K. Kishida. A complete list of Kishida’s works on spiders is provided. Much confusion resulted from the species and higher taxa descriptions or species designations made by Kishida. In many cases he first proposed a new name for an undescribed species found but left its description to his followers. Therefore, some species were really described by another person, while many nomina nuda were produced. A revision of each taxon with systematical and nomenclatural problems will be given in forthcoming parts of this serial (in preparation). Keywords: Bibliography, arachnology, Kyukichi Kishida, Japan Kyukichi Kishida (1888-1968) was a Jap- anese zoologist who studied morphology and systematics of various groups of animals in- cluding spiders, mites, pseudoscorpions and other arachnids, myriapods and insects, as well as sipueculids, birds and mammals. He described from Japan not only small animals : such as spiders, mites and beetles, but also some mammals such as a bat, a vole and even a wolf. He was a pioneer in the history of Japanese I arachnology. Because nobody presented lec- ' tures on arachnology in Japanese universities at that time, he taught himself with European literature and founded some zoological soci- eties in Japan. His students included: Seiji Yu- i hara (1906-1929), Toshio Uyemura (1909- j 1988), Makoto Yoshikura (1911-2003), Koji Nakatsuji (1911-1945), Toshihiro Komatsu [ (1911-1982), Izumi Kayashima (191 1-), Ya- j sunosuke Chikuni (1911-2005), and Takeo Yaginuma (1916-1995). Most present-day Japanese arachnologists including the present author were influenced intellectually by T. Ya- ginuma who made an effort to popularize ar- achnology in Japan with his book, ‘Spiders of Japan in Colour’ (Yaginuma 1960). I In 1929, Kishida established Lanzan-kai, i The Society of Arachnology and Zoology, in I Tokyo and published the journal, Lansania (Fig. 1). The figure on its cover indicates Lan- I zan Ono (1729-1810) to whom Kishida paid respect. Lanzan Ono was an active herbalist in the Edo Era (1603-1867), who published a series of books on Japanese flora and fauna. The society of Lanzan-kai was, however, not always successful and became inactive after only a few years. The Arachnological Society of East Asia was established in 1936 under Kishida and took the place of the Lanzan-kai, and the organ Acta Arachnologica has been continuously published for about 70 years. Despite these accomplishments, Kishida’s legacy is poorly known and some of his spider taxonomy has created considerable confusion. For instance, of more than 100 publications by K, Kishida (see the following pages), only four are listed in Roewer (1942), three in Bon- net (1945), 24 in Brignoli (1983), and only a few are included in the newest international database (Platnick 2005). His works were for- gotten even by Japanese arachnologists. Many of the taxa named by Kishida were not always described correctly and the depos- itory of his collection was unknown. Conse- quently, these were left as nomina nuda. Only a few cases have been solved, for instance, Prodidomus imaidzumii Kishida 1914 was re- described by Platnick (1976), the salticid Chi- rothecia insulana Kishida 1914 was revised and transferred to Harmochirus by Logunov et al. (1997), the corinnid genus Utivarachna Kishida 1940 was recognized by Deeleman- Reinhold (2001), and a small theraphosid 501 502 THE JOURNAL OF ARACHNOLOGY Figure 1. — Front cover of Lansania, the first ar- achnological journal in the world (commenced in 1929) published by Kyukichi Kishida. from Taiwan, Yamia watasei Kishida 1920 was recently redescribed by Haupt & Schmidt (2004). The purpose of this study is to bring the whole aspect on problematical names of spi= ders caused by Kishida’s treatment to light by providing: 1) his personal history and a char- acter sketch, 2) a list of his publications on spiders, 3) a list of spider taxa named by him, 4) a list of valid names extracted from these, 5) a list of nomina nuda, and 6) information on type specimens. This contribution deals with parts 1 and 2. The remaining sections will be provided in forthcoming publications. METHODS Information about Kishida's personal his- tory was acquired through interviews with the late Prof. Seikichi Kishida (1931-2002), the fourth son of Kishida. Publications by Kyu- kichi Kishida were found by searching the li- brary complexes of universities in Japan, and a complete list of his works was made. The missing depositories of his spider collection were followed up. All the Latin names of spi- ders made by K. Kishida were listed from his papers as well as those of other Japanese ar- achnologists and their originality and author- ship were determined according to the past and present rules of the International Code of Zoological Nomenclature. The systematic po- sition of species with valid names was judged based on comparison with specimens in the arachnid collection of the Department of Zo- ology, National Science Museum in Tokyo. Some new synonymies are determined. Valid names as well as remaining nomina nuda are herein listed. RESULTS Brief Personal History of Kyukichi Kish- ida.— Kyukichi Kishida was born in 1888 at Maizuru in Kyoto Prefecture, in central Japan. He grew up during the middle of the Meiji Era (1868-1912), during which Japan became very quickly westernized. Between 1603 and 1867 (Edo Era) this country was closed and isolated from European sciences. Since Lud- wig Koch (1878) first reported on Japanese spiders with Latin names, only European peo- ple led this field. Bosenberg & Strand (1906) described about 400 species and recorded al- most all the common species in the Japanese spider fauna. After graduating from the Teachers' Col- lege of Kyoto Prefecture in 1908, Kishida be- gan his career as a teacher in a primary school. At the same time he learned zoology from the European literature and published his first report (1907) on a spider. This paper was the first by a Japanese researcher to describe a spider species in Latin. Between 1913 and 1914, he published a monograph of Japanese spiders serially in 12 parts in the Scientific World. In 1915, he moved to a junior high school in Odate, northern Japan, and gave lectures on biology, geology and even music. However, after three years he resigned and entered the Department of Zoology of Tokyo University to study zoology. In 1921, he was employed at the Ministry of Agriculture as a scientist. Some of his most important papers were writ- ten at that time, for instance on Yamia Kishida 1920 (Kishida 1920) and Heptathela Kishida 1923 (Kishida 1923). He always considered it more important to place species in a system- atic context within the Araneae rather than to record and describe each species. In 1940, he was employed at Waseda Uni- versity as a lecturer but had to be evacuated from Tokyo in 1944 due to the situation ere- ONO— PERSONAL HISTORY AND WORKS OF K. KISHIDA 503 ated in the city from the events of World War IL He moved to his home in Kyoto to escape the bombings by the American Air Force. Af= ter the war he returned to Tokyo in 1948 and was employed at the Forestry Agency. Unfor- tunately, the great confusion in social condi- tions that prevailed in Japan for about ten years during and after the war decreased his activities in arachnology and his interest tend- ed mainly toward mammalogy and ornithol- ogy during this period. Late in life in 1961, he received the Doctor of Science degree at Hiroshima University with a study in osteology of the Japanese Se- row Capricornulus (an artiodactyle) and in the next year he received the Doctor of Agricul- ture at Tokyo University of Agriculture with a study of Lagomorpha. He died at the age of 80 in 1968 from Par- kinson’s disease. Many unpublished manu- scripts found at his home after death suggest- ed that his erudition with extensive knowledge in zoology may not be shown in full. In 1969, the Arachnological Society of East Asia pub- lished Nos. 49/50 of Atypus as a memorial issue for K. Kishida. M. Yoshikura, T Ko- matsu, 1. Kayashima, K. Morikawa, T. Yagin- uma and T. Uyemura wrote memoirs of him. It is both a strong and a weak point of his character that he had such a wide range of knowledge and interests in zoology. At the time, he was the only specialist in Japan who knew the names of spiders. This led him to assign a new name first without providing a formal description, particularly when he ob- tained undescribed species collected during zoological expeditions and was asked to iden- tify the specimens. The formal descriptions he left to his followers and sometimes he re- turned the specimens to the collectors. It de- pended on his followers whether this new spe- cies would be really described or only cited with Latin names probably assigned by Kish- ida. Therefore, many nomina nuda exist, while some were described by other research- ers. For example, an araneid, Suzumia orien- talis named by Kishida was described three times by Yuhara (1931), Nakatsudi (1943) and Kayashima (1943) from different type locali- ties in Japan and Taiwan. Although the species was regarded as a junior synonym of Cyrto- phora moluccensis (Doleschall 1857) sensu lato, the authorship needs to be confirmed for the future phylogenetic analysis on this group Figure 2. — Portrait of Kyukichi Kishida in 1 964. [Photograph by Seikichi Kishida.] (Ooo 1994). Explanation of each systematic and nomenclatural problem will be given in the coming parts of this subject (in prepara- tion). Works on spiders (Araeeae) of KyukicM Kishida, — The titles were translated from original Japanese to English by Ono, except- ing those with asterisks which were original; t: published after death. Kishida, K. 1907. Notes on the spider’s name “Joro-gumo.” Hakubutsu-no-tomo (a jour- nal of natural history) 47:358-360. Kishida, K. 1908a. Real and common Japa- nese names of Argiope amoena. Hakubut- su-no-tomo 48:27. Kishida, K. 1908 b. Instructions to publish books on spiders. Hakubutsu-no-tomo 56; 284-286. Kishida, K. 1908c. Topics on the spiders (part 1) . The Magazine of Natural History, Tokyo (Hakubutsugaku-zasshi) 8(90):2 1-28. Kishida, K. 1909a. Some specimens of spi- ders. The Magazine of Natural History, To- kyo 104:19-22. Kishida, K. 1909b. Topics on the spiders (part 2) . Studies on Japanese names of spiders 504 THE JOURNAL OF ARACHNOLOGY (continued from volume 8, number 90). The Magazine of Natural History, Tokyo 108:7- 12. Kishida, K. 1909c. Topics on the spiders (part 3) . Collecting and preservation. The Mag- azine of Natural History, Tokyo 1 10:10-16. Kishida, K. 1909d. Topics on the spiders (part 4) . Studies on their classification. The Mag- azine of Natural History, Tokyo 111: 14-23. Kishida, K. 1909e. Topics on the spiders (part 5) . Methods and discussions. The Magazine of Natural History, Tokyo 113:6-11. Kishida, K. 1909f. Supplementary notes of ‘Topics on the spiders.” The Magazine of Natural History, Tokyo 114:11-13. Kishida, K. 1909g. On spiders used for edu- cational material in primary schools. Con- tinued. Kyoto-fu Kyoiku-kai Zasshi (Bul- letin of the Educational Association in Kyoto) 208:16-19. Kishida, K. 1909h. On spiders used for edu- cational material in primary schools. Kyo- to-fu Kyoiku-kai Zasshi 206:19-23. Kishida, K. 1910a. Some specimens of spi- ders. Second report. The Magazine of Nat- ural History, Tokyo 115:13-15. Kishida, K. 1910b. Some specimens of spi- ders. Third report. The Magazine of Natural History, Tokyo 117:1-9. Kishida, K. 1910c. Supplementary notes on spiders. The Magazine of Natural History, Tokyo 118:1-9. Kishida, K. 1910d. Notes on lycosid spiders of Japan, Part 1. Hakubutsu-no-tomo 74: 99-101. Kishida, K. 1911. Notes on a jumping spider, Icidella interrogations. Hakubutsu-no- tomo 80:38-40. Kishida, K. 1912. Examples of mimicry in spiders. The Scientific World (Kagaku- se- kai) 5(10):76-78. Kishida, K. 1913a. Mating of thomisid spi- ders. Science, Kyoto 3(8):369-374. Kishida, K. 1913b. Notes on Joro-gumo {Ne- phila clavata). The Scientific World 7(3): 27-31. Kishida, K. 1913c. Spiders of Japan, Part 1. The Scientific World 7(4): 19-22, 1 pL Kishida, K. 1914a. Spiders of Japan, Part 2. The Scientific World 7(5):31-34. Kishida, K. 1914b. Spiders of Japan, Part 3. The Scientific World 7(6):30-33. Kishida, K. 1914c. Spiders of Japan, Part 4. The Scientific World 7(7):39-43. Kishida, K. 1914d. Spiders of Japan, Part 5. The Scientific World 7(9):40-43, 1 pL Kishida, K. 1914e. Spiders of Japan, Part 6. ! The Scientific World 7(ll):36-40. ;■ Kishida, K. 1914f. Spiders of Japan, Part 7. The Scientific World 7(12):37-42. Kishida, K. 1914g. Spiders of Japan, Part 8. The Scientific World 7(13):35-38, 1 pi. Kishida, K. 1914h. Spiders of Japan, Part 9. The Scientific World 8(l):44-47, 1 pi. Kishida, K. 1914i. Spiders of Japan, Part 10. The Scientific World 8(2):28-32, 1 pi. Kishida, K. 1914j. Spiders of Japan, Part 11. ’ The Scientific World 8(3):31-34. Kishida, K. 1914k. Spiders of Japan, Part 12. The Scientific World 8(4):32-36. Kishida, K. 1915. Studies on egg sacs of spi- ders. The Scientific World 9(4):33-35. Kishida, K. 1920a. Spider fossils from Japan. Zoological Magazine, Tokyo 32:261. ; Kishida, K. 1920b. Notes on Yamia watasei, a new spider of the family Aviculariidae*. ^ Zoological Magazine, Tokyo 32:299-301 , pL 3. Kishida, K. 1920c. Occurrence of a liphistiid : spider in Japan. Zoological Magazine, To- kyo 32:360-363. Kishida, K. 1921a. Retreats of Araneae ther- aphosae, Part 1. Zoological Magazine, To- kyo 33:60-67. Kishida, K. 1921b. Miscellaneous notes on ar- achnology, part 1, (1) — (2). The Scientific World 14(7):20-25. Kishida, K. 1921c. Miscellaneous notes on ar- i achnology, part 1, (3) — (4). The Scientific World 14(8):32-35. | Kishida, K. 192 Id. Retreats of Araneae ther- j aphosae. Part 2. Zoological Magazine, To- kyo 33:109-118. , Kishida, K. 1921e. Exihibition of a marine J spider and Heterothele Kirnurai. In the sec- ’ retary's reports on the regular meeting (26* March) of the Zoological Society of Japan. [ Zoological Magazine, Tokyo 33:135. Kishida, K. 192 If. Miscellaneous notes on ar- achnology, part 2, (5) — (9). The Scientific World 15(l):31-35. Kishida, K. 1921g. Miscellaneous notes on ar- | achnology, part 3, (10) — (14). The Scientif- ic World 15(2):32-37. Kishida, K. 1921h. Miscellaneous notes on ar- achnology, part 4, (15) — (16). The Scientif- ic World 15(3):28-31. Kishida, K. 1923a. Liphistiid spiders, as an ONO— PERSONAL HISTORY AND WORKS OF K. KISHIDA 505 example of primitive animal. Zoological Magazine, Tokyo 35:134-135. Kishida, K. 1923b. Translation of selected passages from Nils Holmgren, 1920, Zur Ontogenie der Stomodealbriicke bei den Spinnentieren. Ark. f. Zook, Stockholm, Bd. 13, Hafte 1-2, No. 1, p. 1-9. Zoological Magazine, Tokyo 35:230-231. Kishida, K. 1923c. Heptathela, a new genus of liphistiid spiders*. Annotationes Zoolo- gicae Japonenses 10:235-242. (Written in English.) Kishida, K. 1923d. Translation of selected passages from Nebel Catherine Elizabeth, 1918, The amount of food eaten by the spi- der, Aranea sericata. Transact. Wisconsin Acad. Sci., Arts & Lett, Vol. 19, pt. 1, pp. 524-530, with 4 tables. Zoological Maga- zine, Tokyo 35:505. Kishida, K. 1924. Spiders from northern Sak- halien, collected by Mr. T. Uchida, Bachelor of Science, with description of a new spe- cies of the genus Dolomedes from Okinawa Prefecture, Zoological Magazine, Tokyo 36: 510-520. Kishida, K. 1926. Spiders. Pp. 303-343, 1 pi. In A Fundamental Study of Animals as Teaching Materials of the State Textbook of Sciences for the Fourth Year (Okazaki, J. et ak). Bunyo-sha, Tokyo. Kishida, K. 1927. Araneae. Pp. 956-970 In Figuraro de Japanaj Bestoj (Uchida, S. et ak). Hokuryukan, Tokyo. Kishida, K. 1928a. Notes on the spiders. Part 1. The Monthly Journal of Science, Tokyo (Ri-gakkai) 26(10):28-33. I Kishida, K. 1928b. Notes on the spiders. Part 2. The Monthly Journal of Science, Tokyo I 26(11):27-31. I Kishida, K. 1928c. Arachnida. Pp.446-491. In I Fuji-no-kenkyu (Studies of Mt. Fuji), Vol. i 6, Fuji-no-dobutsu, Fuji-no-shokubutsu i (Animals and Plants of Mt. Fuji) (Kishida, ' K., & Y. Yabe). Kokin-shoin, Tokyo. Kishida, K. 1928d. Trapdoor spiders of Japan ' and their bearing on zoogeography*. An- notationes Zoologicae Japonenses 11:385- I 387. (Written in English.) I Kishida, K. 1929a, Trap-door spiders of Japan I and their bearing on zoo-geography [ab- i stract of an oral presentation]. Pp.l054- j 1055. In Proceedings of the Third Pan- Pa- cific Science Congress, Tokyo, October 30th-November 11th, 1926, Vol. 1 (for 1928). The National Research Council of Japan, Tokyo. (Written in English.) Kishida, K. 1929b. A Japanese translation of “Cocoon-making by the tarantula” written by Baerg, WJ. (1929). Lansania, Tokyo l(5):65-67. Kishida, K. 1929c. On the oviposition of a clubionid spider, Chirac anthiuni ruhicun- dulum. Lansania, Tokyo l(5):73-74. Kishida, K. 1929d. Book review: Savory, T.H., 1928, The Biology of Spiders. Lan- sania, Tokyo 1(7): 103. Kishida, K. 1930a. A new scheme of classi- fication of spider families and genera*. Lansania, Tokyo 2(1 3): 33-43. Kishida, K. 1930b. Geographical distribution of the spider families*. Lansania, Tokyo 2(15):65-68. Kishida, K. 1930c. On the systematic position of a Japanese spider, Talanites dorsilineatus Doenitz et Strand, 1906*. Lansania, Tokyo 2(16):81-87. Kishida, K. 1930d. A key to the spider fami- lies*. Lansania, Tokyo 2(18): 1 15-123. Kishida, K. 1930e. A new Formosan oxyopid spider, Peucetia formosensis n. sp.* Lan- sania, Tokyo 2(20): 145-1 50. Kishida, K. 1931a. Book review: Esaki, T, 1930, Myriapods and Arachnids, Iwanami- shoten, Tokyo, 128 pp. Lansania, Tokyo 3(21):4. Kishida, K. 1931b. A key to the subfamilies, tribes and genera of the oxyopid spiders*. Lansania, Tokyo 3(21):5. Kishida, K. 1931c. On spiders from the island of Idzu-Ohshima, Tokyo-fu, Japan*. Lan- sania, Tokyo 3(24):59-61. Kishida, K. 193 Id. Postscript. Pp.1-3. In A Study of Spiders (Yuhara, S.). Sogo-ka- gaku-shuppan-kyokai, Tokyo. Kishida, K. 193 le. Life of the spiders. Part 1. The Monthly Journal of Science, Tokyo 29(9):33-36. Kishida, K. 193 If. Life of the spiders. Part 2. The Monthly Journal of Science, Tokyo 29(10):33-36, 1 pi. Kishida, K. 193 Ig. Life of the spiders. Part 3. The Monthly Journal of Science, Tokyo 29(ll):31-34. Kishida, K. 193 Ih. Life of the spiders. Part 4. The Monthly Journal of Science, Tokyo 29(12):31-34, 1 pi. Kishida, K. 1932a. Book review: Savory, T.H., 1926, British Spiders, Their Haunts 506 THE JOURNAL OF ARACHNOLOGY and Habits, Oxford, 180 pp. Lansania, To- kyo 4(31):2. Kishida, K. 1932b. Synopsis of the spider family Gnaphosidae*. Lansania, Tokyo 4(31):3-14. Kishida, K. 1932c. Spiders. Pp. 141-150. In Science Pictorial Series, Vol. 4, Konchu-no- Kyoui (Wonders of Insects) (Nakama, T. ed.). Shinko-sha, Tokyo. Kishida, K. 1932d. Spiders. Pp. 220-222. In: Science Pictorial Series, Vol. 5, Kenbikyo- ka-no Kyoui (Wonders of the Microscopic World) (Nakama, T, ed.). Shinko-sha, To- kyo. Kishida, K. 1933a. Ordgarius hobsoni, new to the Japanese fauna. Zoological Magazine, Tokyo 45:30. Kishida, K. 1933b. Idiobiologia Aranearum*. 82 pp. Ars, Tokyo. Kishida, K. 1934. Spiders from Xingan-ling. Zoological Magazine, Tokyo 46:513. Kishida, K. 1935. Notes on two species of Japanese zodariid spiders. Journal of Zool- ogy and Botany, Wakayama (Kishu-Do- shokubutsu) 2(2): 1-5. Kishida, K. 1936a, Notes on Glenognatha nip- ponica, a Japanese Tetragnathine spider*. Lansania, Tokyo 8(75):65-67. Kishida, K. 1936b. Argiope amoena, female and male. Acta Arachnologica 1(1): cover photograph. Kishida, K. 1936c. A synopsis of the Japanese spiders of the genus Argiope in broad sense*. Acta Arachnologica 1(1): 14-27, pi. 3. Kishida, K. 1936d. Sheet web of Linyphia marginata. Acta Arachnologica 1(2): cover photograph. Kishida, K. 1936e. Funnel web and egg sack of Agelena limbata. Acta Arachnologica l(2):pl. 5. Kishida, K. 1936f. Notes on two spider genera Chiracanthium and Clubiona^ . Acta Arach- nologica 1(2):34-4L Kishida, K. 1936g. Heteropoda venatoria, fe- male. Acta Arachnologica 1(3): cover pho- tograph. Kishida, K. 1936h. A synopsis of the Japanese spiders of the genus Dolomedes *. Acta Ar- achnologica 1 (4): 1 14-127, pL13. Kishida, K. 1936i. An arachnologist from Czechoslovakia, Dr. Baum visited Japan with his wife. Acta Arachnologica 1(4): 151-153, pi. 12. Kishida, K. 1936j. Spiders from Korea. In Uyemura, T, the oral presentation by Mr. Kishida at the First General Meeting of the Arachnological Society of East Asia. Acta j Arachnologica 1(4): 156. ! Kishida, K. 1936k. Spiders and harvestmen of Nikko. Pp.489-494. In Nikko-no Dobutsu to Shokubutsu (Fauna and Flora of Nikko) ! (Tosho-gu ed.). Yoken-do, Tokyo. ; Kishida, K. 19361. Preface; notes on the Chi- ! nese character (Kanji) of Spider; notes on the Chinese character of trapdoor spider; ; notes on the Japanese name of Doosia spi- der; notes on the Japanese name of Argiope amoena; notes on the Japanese name of Ar- ' aneus ventricosus; notes on the Japanese name of a eresid spider; notes on the Jap- anese name of Storena hoosi; information ^ of the Arachnological Society of East Asia. Pp. 7, 24, 46, 140, 144, 148, 152, 156, 178, In Iconographia Colorata Vivida Aranear- ^ um Japonicarum, Vol. 1 (Komatsu, S.). : Ranzan-kai, Tokyo. ' Kishida, K. 1937a. A secret of spiders. Tokyo [ Asahi-Shinbun Newspaper 18256 (Febru- ary 8, 1937):4. ; Kishida, K. 1937b. Sheet web of Linyphia marginata. Acta Arachnologica 2(1): cover photograph. Kishida, K. 1937c. Notes on some spider-egg predators of the dipterous family Chlorop- idae*. Acta Arachnologica 2(3):90-94, pis. 4-5. Kishida, K. 1937d. A synopsis of the Japanese spinous spiders of the genus Gasteracantha in broad sense*. Acta Arachnologica 2(4): 138-149. Kishida, K. 1938. Collecting of spiders and the way of making specimens. The Monthly Journal of Science, Tokyo 36(7):36-41, 1 pi. Kishida, K. 1939a. Diversity of spiders. Sho- gakusei-no kagaku (a science magazine for school children) 2(6):21 (plate). Kishida, K. 1939b. How to collect and study spiders. Shogakusei-no-kagaku 2(8):1130- 1133, 5 figs. Kishida, K. 1939c. An essay on collecting. Aspirator, mites and others. La Scienca Grafikajo, Science Pictorial 28(7):84-89. Kishida, K. 1939d. A general view of the fau- na of northern China. Kagaku-pen (a sci- ence journal) 4(ll):60-75. Kishida, K. 1939e. Widow spiders. Tokyo ONO— PERSONAL HISTORY AND WORKS OF K. KISHIDA 507 Asahi-Shinbue Newspaper 19291 (Decem- ber 17, 1939):7. Kishida, K. 1940a, A biography of Haruo Fu- kasawa, Acta Arachnologica 5(2):46-58. || Kishida, K. 1940b. Notes on two species of spiders, Doosia japonica and Utivarachna fukasawana. Acta Arachnologica 5(2): OS- MS. Kishida, K. 1943. Preface. Pp.i-iii. In Spiders of Taiwan (L Kayashima). Toto-shoseki, Tokyo. Kishida, K. 1954. New records of trapdoor spiders. Atypus, Osaka 7:28. Kishida, K. 1955. A synopsis of spider family Ageleeidae*. Acta Arachnologica 14(1): 1- 13. Kishida, K. 1956. Occurrence of Heptathela kimurai on Amami-oshima Island. Atypus, Osaka 10:33. Kishida, K. 1959. Araeeae. Pp. 367-375. In An Annoted List of Animals of Okinawa Island (Okada, Y. ed.). The Society for Bi- ological Education in Okinawa, Naha. Kishida, K. 1962. To the memory of Mr. Har- uo Takashima, Atypus 26/27:6. Kishida, K. 1966. Personal record of Mr. Koji Kaneko. Acta Arachnologica 20(1): 8. Kishida, K. 1966a. On 15 orders of the class Arachnida; a key to the 15 orders of the class Arachnida*. 6 pp. (Based on the un- published material made by the author in 1915.) Kishida, K. 1966b. On 68 families of the order Araneida; a key to the spider families*. 23 pp. (Based on an unpublished material made by the author in 1938.) , Kishida, K.f 1969a. Common names of spi- ders. Kishidaia, Tokyo 1:1-2. Kishida, K.f 1969b. Common names of spi- ders, 2. Kishidaia, Tokyo 2:1-2. Kishida, K.f 1969c. Common names of spi- ders, 3. Kishidaia, Tokyo 4:1-4. Kishida, K.f 1969d. Notes on Cibunea fron- talis. Kishidaia, Tokyo 8:1-5. ' Kishida, K.f 1969e. Notes on a trapdoor spi- der “Shinaeo-totategumo.” Kishidaia, To- kyo 10:1-3. ' Kishida, K.f 1969f. Notes on primitive spi- ders of the Japanese Empire. Kishidaia, To- kyo 10:3-6. Kishida, K.f 1971. On the occurrence of er- esid spiders in the eastern Asia. Atypus, Osaka 57:1-3. Kishida, K.f 1989. Common names of spi- ders; common names of spiders, 2; common names of spiders, 3; notes on Cibunea fron- talis; notes on a trapdoor spider “Shinano- totategumo;” notes on primitive spiders of the Japanese Empire. Pp.1-3, 9-10, 23-27, 47-51, 61-63, 63-64. In Reprint of Kishi- daia Nos. 1-10 (Kumada, K. ed.). Tokyo Spider Study Group, Tokyo. ACKNOWLEDGMENTS The author would like to express his cordial thanks to the late Professor Seikichi Kishida and his family for offering valuable informa- tion and materials concerning Kyukichi Kish- ida, and to Dr. Paula Cushing for critically reading the manuscript of this paper, to Ms. Yoshie Yamazaki for searching old literature, and to Dr Mark Harvey, Dr Norman L Plat- nick, Dr. Shojiro Asahina, Prof. Iwao Obara, Mr. Eiichi Shinkai and late Mr. Nobom Tak- ahashi for kind advice. This study is partly supported by the Grant-in-aid No. 16540431 for Scientific Research by the Ministry of Ed- ucation, Science, Sports and Culture, Japan. LITERATURE CITED Bonnet, P. 1945. Bibliographia Araneoram. VoL 1. P. Bonnet, Toulouse. Bosenberg, W. & E. Strand. 1906. Japanische Spin- nen. Abhandlungen herausgegeben von der Senckenbergischen Naturforschenden Gesell- schaft, Frankfurt am Main 30:93-422. Brignoli, P.M. 1983. A catalogue of the Araneae described between 1940 and 1981. Manchester University Press, Manchester. Deeleman-Reinhold, C. 2001. Forest spiders of South East Asia. With a revision of the sac and ground spiders (Araneae: Clubionidae, Corinni- dae, Liocranidae, Gnaphosidae, Prodidomidae and Trochanterriidae [sic]). Brill, Leiden, Boston and Koln. Haupt, J. & G. Schmidt. 2004. Description of the male and illustration of the female receptacula of Yamia watasei Kishida, 1920. Spixiana 27:199- 204. Kayashima, 1. 1943. Spiders of Taiwan. Toto-sho- seki, Tokyo. Kishida, K, 1907. Notes on the spider's name “Joro-gumo.” Hakubutsu-no-tomo (a journal of natural history) 47:358-360. Kishida, K. 1920. Notes on Yamia watasei, a new spider of the family Aviculariidae. Zoological Magazine, Tokyo 32:299-307. Kishida, K. 1923. Heptathela, a new genus of li- phistiid spiders. Annotationes Zoologicae Japo- nenses 10:235-242. Koch, L. 1878. Japanesische Arachniden und My- 508 THE JOURNAL OF ARACHNOLOGY riapoden. Verhandlunden der kaiserlich-konig- lichee zoologisch-botanischen Gesellschaft in Wien 27:735-798. Logunov, D.V., H. Ikeda & H. Ono. 1997. Jumping spiders of the genera Harmochirus, Bianor and Stertinius (Araneae, Salticidae) from Japan. Bul- letin of the National Science Museum, Tokyo, Series A (Zoology) 23:1-16. Nakatsudi, K. 1943. Some Arachnida from Is. Oki- nawa and Is. Amami-Osima. Journal of Agricul- tural Science, Tokyo Nogyo Daigaku (Tokyo Ag- ricultural University) 2:181-194. Ono, H. 1994. Spiders described by Koji Nakatsudi. Acta Arachnologica 43: 108-1 11. Platnick, N.L 1976. On Asian Prodidomus (Ara- neae, Gnaphosidae). Acta Arachnologica 27:37- 42. Platnick, N.L 2003. The World Spider Catalog, Ver- sion 5.5. American Museum of Natural History, New York at http://research.amnh.org/entomolo- gy/spiders/catalog/index .html . Roewer, C.F. 1942. Katalog der Araneae. VoL 1. R. Friedlander und Sohne, Bremen. Yaginuma, T, 1960. Spiders of Japan in Colour. Hoiku-sha, Osaka. Yuhara, S. 1931. Study of Spiders. Sogo-kagaku- shuppan-kyokai, Tokyo. Manuscript received 25 January 2005, revised 12 July 2005. 2005. The Journal of Arachnology 33:509-515 FORAGING STRATEGIES OF ERIOPHORA EDAX (ARANEAE, ARANEIDAE): A NOCTURNAL ORB^WEAVING SPIDER Leonor Ceballos: ECOSUR, AP 36, Tapachula, Chiapas, Mexico Yann Henauth ECOSUR, AP 36, Tapachula, Chiapas, Mexico. E-mail: yhenaut@ tap-ecosur.edu.mx Luc Legal: LADYBIO-CNRS/UPS,l 18, route de Narbonne-bat. 4R3, 31062 Toulouse cedex 4”-Erance ABSTRACT. Studies on the ecology of orb spiders have focused on diurnal spiders, especially field studies. Nocturnal spiders, however, face different conditions due to the type of prey found at night. A field study was conducted to observe the activity of adult females of Eriophora edax in their natural environment, and to analyze their predation efficiency and web retention properties. Most of the spiders were observed around sunset, which suggests that E. edax tends to build webs in the early evening. In order to evaluate the predation efficiency of E. edax we compared its behavior and web retention properties with the behavior of a diurnal orb-weaving spider, Verrucosa arenata. Two prey types, a diurnal Hyme- noptera and a nocturnal Lepidoptera, were selected and presented to the spiders, to record approach time and prey capture time. The results showed that E. edax spent more time to capture Hymenoptera than to capture Lepidoptera. During the experiments of web prey retention time, Hymenoptera consistently showed greater tumbling than Lepidoptera, but the total retention time was the same for both prey types. Our results showed that E. edax forages strictly at night and, in terms of prey capture and web retention, was more efficient when preying on Lepidoptera. Keywords* Eriophora edax, web-building spider, nocturnal activity, prey selection. Web-building spiders present a unique case of “sit-and-wait” predation (Heiling 1999), so they are not expected to exhibit prey special- ization (Uetz 1990). However, recent studies have shown that many web-building spiders exhibit considerable dietary specialization (Riechert & Luczak 1982; Stowe 1986; Nen- twig 1987). For example, Tetragnatha mon- tana Simon 1874, an orb weaver found in Eastern Europe, feeds mainly on mosquitoes (Dabrowska-Port & Luczak 1968; Dabrows- ka-Port et al. 1968; Luczak 1980). Habitat choice and activity pattern of the species are closely tied to the occurrence and activity of the preferred prey (Uetz 1990). It has been suggested that nocturnal web- building, particularly in the tropics, is an ad- aptation to avoid the visibility of webs in day- time (Rypstra 1979, 1982). The optical properties of some orb webs tend to reduce its visibility, especially in low-light and varying ’ Corresponding author. background conditions (Craig et al. 1985; Craig 1986). Several species of orb weaving spiders ingest their previous web and replace it with a new one (Breed et al. 1964; Eberhard 1971; Carico 1986). The renewal of the web is critical, because a web’s ability to capture food decreases over time as a result of contact with prey and non-prey items that destroy both threads and glue (Chacon & Eberhard 1980). In a study on the predatory capacity of four sympatric species of web-building spiders that inhabit coffee plantations in Southern Mexico, Henaut et al. (2001) found that the consump- tion of prey was related to the predatory strat- egy of each spider species. For example, Gas- teracantha canciformis (Linnaeus 1785), a diurnal orb weaving spider, built a new web every morning and prey storage was never ob- served. In contrast, Cyclosa caroU (Hentz 1850), another diurnal orb web spider, built a “permanent” web (only renewed when dam- aged) and stored prey on a stablimentum, 509 510 THE JOURNAL OF ARACHNOLOGY which may explain the very low incidence of immediate prey consumption observed in this species (Henaut et al. 2001). However, a cen- sus of the prey captured by C caroli and G.canciformis showed that both speeies have a marked positive electivity for Diptera and Hymenoptera (Ibarra-Nunez et al. 2001). There are numerous reports concerning pre- dation by web-building spiders (Heiling 1999; Henaut et al. 2001; Ibarra-Nunez et al. 2001) although the vast majority involves diurnal species. In contrast, the present study inves- tigated the foraging activity of a nocturnal orb web spider, Eriophora edax (Blackwell 1896 (Araneidae)). This Pan-american species with a body length ranging from 12-16 mm (Levi 1970) was selected due to its nocturnal activ- ity and its abundance. The web of E. edax is vertical, and the spider stays at the hub of the web with its head facing down. The study was divided in two parts. First, we examined in situ the activity and the prey captured by adult females of E, edax. Second, we compared the prey capture behavior of E. edax with the prey capture behavior of a di- urnal orb weaving spider. METHODS Study site. — The study was conducted in July and August 2002 in a coffee plantation at the agricultural experimental station “Ro- sario Izapa” of the INIFAP (Institute Nacional de Investigaciones Forestales, Agricolas y Pe- cuarias), situated at 400 m above sea level in the state of Chiapas, southern Mexico (14° 58' N, 92° 09' W). The climate is tropical, warm and humid. Heavy rainfall (3000 mm per month) occurs from May-October. During the course of the study, temperature fell to ap- proximately 23 °C at night, and rose to about 33 °C during the day. The relative humidity was around 85%, day and night. Spider activity. — We observed the spiders’ activity for three nights without rain (when spiders are active and observers can stay the entire night in the field). Observations were done from 1800 to 0700. At this time of the year sunset oecurred between 1900 and 1930 and sunrise between 0630 and 0700. We walked hourly along a 200 m transect (using a chronometer to check the time), to check for E. edax spiders and their webs. It took from 30-45 min to record all the spiders of a tran- sect. Flashlights with dark red plastic cover facilitated observation while neither attraeting insect prey, nor disturbing the spiders’ natural photoperiod (Herberstein & Elgar 1994; Heil- ing 1999). On each transect walk we recorded the spi- ders present in the bush with or without a web and the absence of individuals previously re- corded. We marked spiders’ positions individ- ually with a numbered piece of white plastic located on the nearest twig. Spider activities were reeorded as: building the web, catching a prey (when a spider was wrapping a prey with silk), and eating a prey (when a spider was actually bitting a prey or was handling it in its chelicerae). All voucher specimens are deposited in the Collection of the Laboratory of Arthropod Ecoethology (Laboratorio de Ecoetologia de Artropodos) in Ecosur, Tapachula, Mexico. Spiders’ prey. — Prey items captured in the webs were visually identified to the level of order. These prey items were not removed from the webs. Prey identification to lower levels, although desirable, would have result- ed in substantial disturbance of the webs. We compared the hourly numbers of each order of prey captured by E. edax web with a Chi- square test (SPSS 10.00 for Windows). Predation efficiency and web retention properties. — In order to evaluate the preda- tion efficiency and the web retention proper- ties of E. edax, we conducted two field ex- periments during the same months but on different nights than the activity observations. We selected Verrucosa arenata (Walckenaer, 1841) (body length: 8-15 mm) as a model of diurnal orb weaving spiders. Like E. edax, it is an araneid, builds its web every day, and dismantles it at the end of its daily activity period. However, it is as strictly diurnal as E. edax is nocturnal. Finally, V. arenata is pre- sent in the same habitats as E. edax. Two ex- perimental prey types were selected, because they are abundant in the coffee plantation (Ibarra-Nunez 1990). Adults of the moth Si- totroga cereallela (Olivier 1819) (Lepidop- tera, Gelechidae) were selected as represen- tatives of a noeturnal prey, while the stingless bee Scaptotrigona mexicana Guerin (Hyme- noptera, Apidae) was chosen as an example of a diurnal prey. Prey specimens were obtained from laboratory cultures. For both prey types, field experiments were performed during three days for V. arenata and during three nights 1 CEBALLOS ET AL.— FORAGING STRATEGIES OF E. EDAX 511 CZD appearance of webs CZG disappearance of webs -•-Nunn, of spiders Time of night ■ Figure 1. — Number of Eriophora edax individuals with or without a web (number of spiders), appear- ; ance of webs and disappearance of webs in the study site during a 12 hour observation period. Sunset occurred between 1900 hrs and 1930 hrs; sunrise between 0630 hrs and 0700 hrs. I for E. edax. For each type of prey and for each spider species, 20 individuals were tested for ! the predation and web retention studies. For ! each prey type, observations were made in the i same 24 hour period for both spider species. For the predation efficiency experiments, webs were selected based on the following criteria: no signs of remains of prey, spider I was an adult female located at the center of the web. Each prey item was gently blown into the web with the aid of an inverted as- pirator from a distance of 10 cm. All prey were alive and visually undamaged before and after introduction into the web. Once the prey made contact with the web, the behavior of the spider was registered in terms of approach and prey capture (measured : in seconds). The prey capture event started at 1 the moment the spider bit the prey, continued i with its manipulation and finished when the I spider took it to the center of the web. Ob- servations were conducted for a 5 min period, which was enough for recording the complete capture event. We compared the predation ef- ficiency of both spider species with both types of prey with an ANOVA (Statistica 6.0). For the web retention experiments, webs were selected based on the same criteria as above. Spiders were carefully removed from their web, and prey items were blowed the same way as mentioned before. Once the prey made contact with the web, a small piece of paper was set at the impact point to measure the distance the prey tumbled. The prey was observed for a 5 min period, after which the tumbling distance was measured in centime- ters. If the prey remained on the web for more than 5 min, a second tumbling distance of the same prey was measured after 30 min. Once the experiment ended, the spider was returned to its web. The data obtained from both ob- servation periods (5 min and 30 min) were contrasted for spider and prey types (nocturnal vs. diurnal) with an ANOVA (Statistica 6.0). RESULTS Spider activity. — E. edax was not ob- served before 1900. Most of the spiders ap- peared on coffee bushes or were building their webs between 1900-1930, around sunset time {n = 48 of a total of 74 spiders observed for the three nights). Around 68% of the spiders were present between 1900 and 2000, and we 512 THE JOURNAL OF ARACHNOLOGY Time of night Figure 2. — Frequency of web building, prey catching and prey eating by Eriophora edax in a coffee plantation during the 12 hrs observation period. Sunset occurred between 1900 hrs and 1930 hrs; sunrise t between 0630 hrs and 0700 hrs. observed no spider after 0700, when the sun rose (Fig. 1). Although spiders were able to build their web all night, this activity was more intense between 1900 and 2200. Other smaller peaks of this activity occurred around 0100 and 0500. E. edax requires less than one hour to build its web, as the web was completed be- tween two subsequent data recordings, and most often the spiders had already caught a prey when its web was observed for the first time (Fig. 2). Catching prey was most intense at the be- ginning of the night, between 1900 and 2300. Then the catching activity decreased through the night, although this activity increased again slightly between 0300 and 0400 just af- ter the second peak of building (Fig. 2). Spiders began to eat prey at 1900, but this activity peaked at 2100, right after the peak of catching activity. Other smaller peaks of eating activity occurred at 0000, and between 0400 and 0500. We also observed that spiders stopped eating before sunrise (0700), even if they had caught a prey (Fig. 2). E. edax is more active at the beginning of the night than at the end of the night (Fig. 2). Of the 74 spiders observed during the three nights, 55.4% caught only one prey, 9.5% ' caught two prey and 35.1% did not capture I any prey. Spiders’ prey. — The main prey items caught by E. edax {n = 55) were Lepidoptera (67.7%), Coleoptera (21.5%), Diptera (9.2%), and Hymenoptera (1.5%). Minor taxa includ- ed Orthoptera and Hemiptera (< 0.1%). The number of prey items of different insect orders differed statistically (x^ = 150.7, d.f. = 8, F = 0.001). Lepidoptera were mostly caught at the beginning of the night (1900-2100) with a second, smaller peak of capture between i 0300 and 0400. Coleoptera were caught by the ' spiders between 1900 and 2200. * Predation efficiency and web retention properties. — The time spent to reach and cap- ture a prey as well as the tumbling of the prey into the webs varied according to the spider species (Table 1). E. edax spent less time to reach Lepidopterans but more time to capture Hymenopterans than V. arenata and the turn- [ bling is more important with the web of E. edax (Table 1). However, the comparison be- tween prey show that the time spent to reach CEBALLOS ET AL.— FORAGING STRATEGIES OF E, EDAX 513 Table 1. — Time (in seconds ± SE) of each spider species (Eriophora eclax and Verrucosa arenata) to reach and to capture the offered prey (Lepidoptera and Hymenoptera), as well as each prey’s tumbling distance (in cm) in the web after 5 min and after 30 min of observation. ***, P < 0.001; *, P < 0.05; ns, Not significant P > 0.05, ANOVA. E. edax V. arenata Comparison Time to reach prey (sec) Lep. 3.3 ± 0.7 46.3 ± 18.3 * Hym. 12.8 ± 8.6 17.2 ± 7.5 ns Prey capture time (sec) Lep. 18.1 ± 1.9 20.6 ± 5.2 ns Hym. 71.3 ± 7.5 20.7 ± 2.6 Tumbling after 5 min (cm) Lep. 1.1 ± 0.8 1.9 ± 1 ns Hym. 5.2 ± 1.02 4.8 ± 1.3 ns Tumbling after 30 min (cm) Lep. 8.6 ± 1.2 0.8 ± 0.4 Hym. 8.5 ± 1.2 5.8 ± 1.6 ns a prey was not significantly different between the two prey types = 2; P = 0.16). I We also found significant differences in the prey capture time between the two spider spe- cies. E. edax spent more time to capture Hy- menoptera than to capture Lepidoptera (Fj^g I = 46; P = 0.000). On the other hand, the time V. arenata spent to capture both Hymenoptera and Lepidoptera did not differ significantly {Fi^3s = 0.000; P = 0.9). Prey tumbled differently according their type (diurnal or nocturnal), and to the spider : species to which the web belonged, for both : observation periods (5 min and 30 min). Dur- ing the 5 min observation period, Hymenop- tera tumbled a longer distance than Lepidop- I tera in E. edax and it tended to be the same ; in V. arenata webs {E. edax: F,jg = 10; P = I 0.002; V. arenata: Fj jg = 3.2; P = 0.08). Dur- ; ing the 30 min observation period Hymenop- ' tera tumbled a longer distance than Lepidop- ! tera in V. arenata but the tumbling was not j different in E. edax {E. edax: Fj jg = 0.005; P i = 0.9; V. arenata: Fjjg = 15.6; P = 0.000). ' During the 5 min observation period, both ! E. edax and V. arenata webs retained 100% I of the blown prey. After 30 min, E. edax web I retained 95% of the Lepidoptera and 90% of ! the Hymenoptera, and V. arenata webs re- tained 80% of Lepidoptera and 50% of Hy- ; menoptera (x^ = 0.5, d.f. = 1, P = 0.5). DISCUSSION I Our results confirm that Eriophora edax j forages strictly at night, spins a new web ev- I ery night and dismantles it at dawn. Most spi- I ders started to build their web just after sunset, j and all spiders had disappeared at sunrise. Even if E. edax caught prey during the 12 h observation period, it seems to have a strategy to “build, catch and eat” in a short period of time. Most prey is caught and eaten within a two hour period after the web is built. In com- parison with other orb-weaving spiders (Hen- aut et al. 2001) this spider captures a low number of prey (generally just one per night) and never makes prey caches. We did not ob- serve E. edax relocate its web after capture and consumption of a prey. Thus, whether new arrivals during the night are new spiders or spiders building a second web in a new place, remains to be tested. Capture rates were higher at the beginning of the night, probably due to the level of prey activity at this time (unpubl. data). As other Eriophora species, E. edax preyed mainly on Lepidoptera. For example, Herberstein & El- gar (1994) found that E. transmarina (Key- serling 1865) captured mostly Lepidoptera, which were also more abundant at night. Another evidence of the strategy of E. edax to maximize capture time is that most spiders waited on their web almost all night, even when they had already caught some prey. Also, spiders did not dismantle their web until just before dawn, even if they had not caught or eaten a prey. Although E. edax's main prey was Lepi- doptera it did vary its diet by eating other or- ders of insects, such as Coleoptera, Diptera, Hymenoptera, Orthoptera and Hemiptera. Nyffeler (1999) found that overall fewer than 10 arthropod orders (Diptera, Homoptera, Hy- menoptera, Heteroptera, Collembola, Coleop- tera, Lepidoptera, and Araneae) make up the bulk of the prey of common agroecosystem spiders. Dietary mixing seems to be advanta- 514 THE JOURNAL OF ARACHNOLOGY geous by optimizing a balanced nutrient com- position needed for survival and reproduction (Greenstone 1979; Uetz et al. 1992; Toft 1995). However, in comparison with the di- urnal spider E. edax is more efficient at reach- ing and capturing the moth than the bee, and its web offers a better prey retention for Lep- idoptera than V. arenata's web. The predatory behavior of the nocturnal spider seems to be more specialized towards moths, though the results on web retention might be influenced by differences in web properties caused by the difference in temperature between day and night (around 10 °C). ACKNOWLEDGMENTS We are grateful to the staff of INIFAP ex- perimental station at Rosario Izapa who kind- ly granted access to the field site. We thank Sophie Calme, Jean-Paul Lachaud and Julio Rojas Leon for their helpful comments and suggestions on a previous version of the man- uscript; Jorge Merida for his help in the field; and Javier Valle Mora for statistical advice. Sophie Calme revised English spelling of the last version. Financial support was provided by Consejo Nacional de Ciencia y Tecnologfa (CONACYT, Mexico). LITERATURE CITED Breed, A.L., V.D. Levine, D.B. Peakall & P.N. Witt. 1964. The fate of the intact orb-web of the spider Arcmeiis diadematiis Clerk. Behaviour 23:43-60. Carico, J.E. 1986. Web removal patterns in orb- weaving spiders. Pp. 306-318. In Spiders: webs, behavior and evolution. (W. A. Shear, ed.). Stan- ford University Press, Stanford, California. Chacon, P. & W.G. Eberhard. 1980. Factors affect- ing numbers and kinds of prey caught in artificial spider webs, with consideration of how orb webs trap prey. Bulletin of the British Arachnological Society 5:29-38. Craig, C.L., A. Okuba, & V. Andreasen. 1985. Ef- fect of spider orb web and insect oscillations on prey interception. Journal of Theoretical Biology 115:201-211. Craig, C.L. 1986. Orb- web visibility: the influence of insect flight behaviour and visual physiology on the evolution of web designs within the Ar- aneoidea. Animal Behaviour 34:54-68. Dabrowska-Port, E. & J. Luczak. 1968. Spiders and mosquitoes of the ecotone of alder forest (Carici elongate- Alnetum) and oakpine forest (Pino- Quercetum). Ekologia Polska 16:461-483. Dabrowska-Port, J. Luczak, & K. Tarwid. 1968. The predation of spiders on forest mosquitoes in field experiments. Journal of Medical Entomol- ogy 5:252-256. Eberhard, W.G. 1971. The ecology of the web of Uloborus diversus (Araneae: Uloboridae). Oec- ologia 6:328-342. Greenstone, M.H. 1979. Spider feeding behavior optimizes dietary essencial amino acid compo- sition. Nature 282:501-503. Heiling, A. 1999. Why do nocturnal orb-web spi- ders (Araneidae) search for light?. Behavioral Ecology and Sociobiology 46:43-49. Henaut, Y., J. Pablo, G. Ibarra-Nunez & T Williams. 2001. Retention, capture and consumption of ex- perimental prey by orb-web weaving spiders in coffee plantations of Southern Mexico. Entomo- logia Experimentalis et Applicata 98:1-8. Herberstein, M.E. & M.A. Elgar. 1994. Foraging strategies of Eriophora transmarina and Nephila plumipes (Araneae: Araneoidea): nocturnal and diurnal orb-weaving spiders. Australian Journal of Ecology 19:451-457. Ibarra-Nunez, G. 1990. Los artropodos asociados a cafetos en un cafetal mixto del Soconusco, Chia- pas, Mexico. I. Variedad y abundancia. Folia En- tomologica Mexicana 79:207-231. Ibarra-Nunez, G., J.A. Garcia; J.A. Lopez & J.P. Lachaud. 2001. Prey analysis in the diet of some ponerine ants (Hymenoptera: Formicidae) and web-building spiders (Araneae) in coffee plan- tations in Chiapas, Mexico. Sociobiology 37: 723-756. Levi, H.W. 1970. The ravilla group of the orb weaver genus Eriophora in North America (Ar- aneae: Araneidae). Psyche 3:280-302. Luczak, J. 1980. Behaviour of spider populations in the presence of mosquitoes. Ekologia Polska 31: 625-634. Nentwig, W. 1987. The prey of spider. Pp 249-263. In Ecophysiology of spiders. (W. Nentwig, ed.). Springer- Verlag, Berlin, New York. Nyffeler, M. 1999. Prey selection of spiders in the field. Journal of Arachnology 27:317-324. Riechert, S.E. & J. Luczak. 1982. Spider foraging: behavioral responses to prey. Pp. 353-385. In Spider Communication: mechanisms and ecolog- ical significance. (P. N. Witt & J. S. Rovner, eds.). Princeton University Press. Princeton, New Jersey. Rypstra, A.L. 1979. Foraging folks of spiders, a study of aggregate behavior in Cryptophora ci- tricola Forskal (Araneae: Araneidae) in west Af- rica. Behavioral Ecology and Sociobiology 5: 291-300. Rypstra, A.L. 1982. Building a better insect trap: an experimental investigation of prey capture in a variety of spider webs. Oecologia 59:312-319. Stowe, M.K. 1986. Prey specialization in the Ara- CEBALLOS ET AL„— FORAGING STRATEGIES OF E. EDAX 515 neidae. Pp. 101-131. In Spiders: Webs, Behavior and Evolution. (W. A. Shear, ed.). Stanford Uni= versity Press, Stanford, California. Toft, S. 1995. Value of the aphid Rhopalosiphum padi as food for cereal spiders. Journal of Ap- plied Ecology 32:552-560. Uetz, G.W. 1990. Prey selection in web-building spiders and evolution of prey defenses. Pp. 93- 128. In Insect defenses. Adaptive Mechanisms and strategies of prey and predators. (D. L. Evans & J. O. Schmidt, eds.). State University of New York Press, Albany, New York, Uetz, G.W., J. Bischoff & J, Raver. 1992. Survi- vorship of wolf spiders (Lycosidae) reared on different diets. Journal of Arachnology 20:207- 211. Manuscript received 16 September 2004, revised 13 September 2005. 2005. The Journal of Arachnology 33:516-522 THE WASP SPIDER ARGIOPE BRUENNICHI (ARACHNIDA, ARANEIDAE): BALLOONING IS NOT AN OBLIGATE LIFE HISTORY PHASE Andre Walter, Peter Bliss^and Robin F.A. Moritz: Institut fur Zoologie, Martin- Luther-Universitat Halle-Wittenberg, Hoher Weg 4, D-06120 Halle (Saale), Germany. E-mail: bliss@zoologie.uni-halle.de ABSTRACT. Aerial dispersal (“ballooning”) of Argiope bruennichi spiderlings has been claimed to be an obligate life history trait and a prerequisite for spinning prey-capture webs. If this were true, a bal- looning phase would be essential for any laboratory rearing of A. bruennichi making rearing protocols particularly elaborate. We tested the significance of ballooning for second-instar spiderlings in the labo- ratory and showed that the ballooning behavior is not essential for building prey-capture orb webs. Our results also give no evidence for the hypothesis that recent natural selection has changed ballooning behavior in newly founded field populations. Keywords: Araneae, ballooning experiment, laboratory rearing, web-building behavior. Ballooning is a common dispersal mecha- nism for many modern spiders (Coyle 1983; Dean & Sterling 1985; Weyman 1993), and this behavior is particularly important for maintaining genetic cohesion among Argiope populations (Ramirez & Haakonsen 1999). The life history of Argiope is characterized by ballooning, the aerial transport on wind-blown silk threads. A good example for the impor- tance of ballooning for range expansion is the Palearctic wasp spider Argiope bruennichi (Scopoli 1772). The spider is an r-strategist (Guttmann 1979), characterized by high aerial dispersal capability and an ongoing postgla- cial expansion of its geographical range in Eu- rope (van Helsdingen 1982). Females of A. bruennichi produce up to five cocoons in the field, often containing several hundred eggs (Crome Si Crome 1961; Kohler & Schaller 1987). The expansion of the species has ac- celerated in the second half of the last century probably due to factors favoring dispersal by ballooning (Guttmann 1979; Levi 1983; Sach- er & Bliss 1990; Scharff & Langemark 1997; Jonsson & Wilander 1999; Smithers 2000). The wasp spider prefers grassy or herbaceous vegetation in open, ephemeral or shrubby sites (Wiehle 1931; Pasquet 1984; Malt 1996) in coarse-grained (patchy) landscapes (Gillandt ' Corresponding author. & Martens 1980; Sacher & Bliss 1989) and has regionally benefited from an extension of farming production and urbanization (Loh- meyer & Pretscher 1979; Arnold 1986; Nyf- feler & Benz 1987). River valleys have been identified as favored dispersal corridors fur- ther supporting the importance of ballooning for dispersion (Gauckler 1967; Puts 1988). Follner & Klarenberg (1995) claimed bal- looning to be an obligate phase in the devel- opment of A. bruennichi. These authors mon- itored the pre-ballooning and ballooning behavior of spiderlings in a grassland study site near Munich (Germany). Since they never found aggregations of orb webs in the neigh- borhood of the cocoons from which the over- wintering second instar spiderlings eclosed and they only observed the construction of first prey-capture orb webs after a ballooning trip, they concluded “that aeronautic behav- iour in Bavarian populations of A. bruennichi is obligatory”. Moreover, these authors sug- gested that spiderlings, which have hatched from the cocoon, will starve to death, unless they perform a ballooning trip. Ballooning should thus be an obligate phase to switch from a non-predatory, passive phase to one of active predation by spinning prey-capture orbs. Follner & Klarenberg (1995) argued that the obligatory aerial dispersal might be a re- sult of recent natural selection and be the rea- 516 T j WALTER ET AL.— BALLOONING IN Ai?G/OP£ 317 1 Figures 1-4, — Design and course of the ballooning experiment. The spiderlings were placed on a spatula (sp) and exposed to a light air current by a fan (ve) and heat source (hs), which were placed at the left edge of a lab bench (240 cm). After cutting the drag line the spiderlings became airborne to land on the lab bench, which served as a landing strip (Is), 1. Pre-ballooning behavior: sp = spatula; ve = ventilator (light breeze); hs == heat source (25 Watt lamp, distance to spatula = 20 cm); sh = spiderling hanging from a dragline; bt = ballooning thread. 2. Initial ballooning phase. 3. Airborne spiderling: Is = “landing strip” (lab bench of 240 cm length). 4. Landing phase. son behind the swift expansion of the species. New populations v/hich are established during ■ a period of expansion are always founded by individuals, which have ballooned. If ballooning were a truly obligate phase, it would not only be important for natural selec™ tion but also be important for any rearing pro^ tocol for A. bruennichi. Allowing for balloon- ing in a rearing procedure might easily render ; laboratory breeding unfeasible as it could : prove to be too time-consuming and laborious. I However, an obligate ballooning phase has i never been observed before, neither in other I Argiope nor in the generally well studied A. ! bruennichi. Tolbert (1976, 1977) studied baL I loonieg behavioral elements of A. trifasciata i (ForskM 1775) and A. aurantia Lucas 1833. He concluded from field and laboratory ob- servations that “it is unnecessary for spider- lings of either Argiope species to engage in aerial dispersal before building an orb web” (Tolbert 1977), which is an obvious discrep- ancy to Folleer’s and Klareeberg’s (1995) I claims. We here test the significance of bal- ^ looning for the construction of the first prey- S capture web in the laboratory by comparing i spiderlings reared under two experimental ( conditions, one with and one without balloon- j ing. I We collected cocoons of A. bruennichi {n = 6) in dry and semi-dry grasslands north- east of Halle (Saale) in late April 2002 (Ger- many, 160 m a.s.L, 5r3331" N, 0ir52'49" E). They were maintained in the lab in indi- vidual glass vials (9 cm diameter, 13 cm height, coated with fine gauze) at 23 ± 2 °C and mist-sprayed with water every two days to avoid desiccation. The vial bottom was covered with initially wet cellulose wadding (1 cm). Second-instar spiderlings hatched from the cocoons in early May. One day after hatching we simulated indi- vidual ballooning for 60 spiderlings (10 from each cocoon) by exposing the spiderling on a spatula to an air stream generated by a heat source and a fan (see Figs. 1-4 for details of the experimental design). We observed behav- ioral elements in the pre-ballooning phase in detail and noticed its mode. When the spider- ling became airborne, we tracked it and re- trieved it at the “landing strip” (Figs, 3, 4). The ballooning experiment was repeated im- mediately (re-ballooning) for each individual to satisfy a possible “ballooning drive” (see Tolbert 1977). The spiderlings had to actively participate in this experiment by showing the entire sequence of pre-ballooning and bal- looning behavior (Figs. 1-4). Following the experiments, the “balloon- ers” were kept in the same unheated indoor room with windows admitting indirect natural light. They were housed in groups {n “ 20) in three gauze covered glass terraria (50 X 30 X 31 cm; 25 ± 3 °C; 65 ± 10% RH) and fed ad libitum 45-50 live Drosophila me lanog as- ter once a day. Every two days we sprinkled the inside surfaces of the terraria with water. This prevented desiccation and allowed for 518 THE JOURNAL OF ARACHNOLOGY Web-building activity Days — — — Ballooner — — Non-Ballooner Figure 5. — Web-building activity of the A. bruennichi spiderlings during laboratory rearing for both ballooners and non-ballooners. normal drinking behavior of the spiderlings. The bottoms of the terraria were covered with a layer of commerciak pasteurized potting soil (3 cm) with grass tufts, some dry twigs and wooden skewers to enhance the number of po- tential attachment points for web building. A control group of spiderlings (n = 60) was treated in the same way, but without the bal- looning procedure (“non-ballooners’'). In both groups (ballooners vs. non-ballooners) spiderlings and orb webs were noted three times daily at 6 a.m., 12 p.m. and 6 p.m. to ensure individual based data sets. The rearing period was cut off after 19 days when all the surviving individuals had spun their first prey- capture orb- webs. Voucher specimens are deposited in the En- tomological Collection of the Martin-Luther- University Halle- Wittenberg (Zoological In- stitute), Germany (identification number 2568). The web-building activity of the spiderlings increased in both the ballooners and the non- ballooners over time and reached 90 ± 5% for ballooners (n = 54, three terraria) and 95 ± 5% for non-ballooners (n = 57, three terraria) within a period of 19d (Fig. 5). The differ- ences in the web-building activity (Fig. 5) were not statistically significant between the I two groups of spiderlings (Kruskal-Wallis test, | P = 0.7515; tested for daily built-first webs). i The mean latency time for web-building (time | from hatching from the cocoon to the con- ' struction of the first prey-capture web) was ! 8.61 ± 4.28 days and 8.18 ± 3.60 days for ballooners (n = 54) and non-ballooners (n = I 57) respectively. This difference was not sta- tistically significant (t-test, P = 0.56). Although mortality increased in the second half of the observation period (Fig. 6), it did ^ not exceed 22% at the end of the experiment (ballooners: 21.7 ± 2.89%, n = 13, non-bal- looners: 20.0 ± 8.66%, « == 12, difference not i significant, t-test, P = 0.77). The surviving animals caught prey in their orb webs and showed normal development with up to four molts within the experimental time. Using our protocol, we could initiate the j full sequence of ballooning behavior promptly in every experiment. The A. bruennichi spi- derlings always showed an identical sequence ; of pre-ballooning and ballooning behavior (Fig. 1-4). When exposed to the heat from the j lamp, they displayed the “ballooning drive” i WALTER ET AL.— BALLOONING IN ARGIOPE BRUENNICHI 519 Mortality Days “ — ^Ballooner “ “ Non-Balfooner Figure 6. — Mortality of the A. bruennichi spiderlings during laboratory rearing. I behavior. Individuals walked to the margin of the spatula, spooled out a dragline and dropped down hanging from the line. While ' suspended and holding on to the drag line, they let out an additional line of 50-100 cm : ballooning silk (Fig. 1). When this was lifted I by the breeze generated by the fan and the I heat source, the spiderlings cut the dragline I and became airborne (Figs. 2, 3). After land- I ing (Fig. 4) they hauled in the ballooning line, j formed it with the legs into a silk blob and finally ate the silk, bringing the ballooning be- havioral sequence to completion. Tolbert (1977) observed two modes of preparation for ballooning in sympatric field populations of A. trifasciata and A. aurantia. I A spiderling attempting to become airborne : climbed to the top of some blade of grass or other structures and adopted the typical “tip- I toe” posture by depressing the cephalothorax 1 and elevating the opisthosoma. Multiple silk ! lines were thee exuded from the spinnerets. I When moving air generated sufficient silk, the I spiderling became a “ballooner” (Nielsen I 1932; Richter 1970; Eberhard 1987). Alter- j natively, the spiderling could become airborne by dropping and hanging from a dragline, spinning a ballooning thread, which then grad- ually lifted and lengthened in the breeze. The ballooner then cut the dragline and floated off into the air (Nielsen 1932; Bristowe 1939). Argiope bruennichi can display both pre- ballooning modes. However, the drop and dragline mediated ballooning seems to be more frequent (Follner & Klareeberg 1995). In the field, second-instar spiderlings usually attach the draglines to tips of grass blades or they use silk threads which connect the tips of grass haulms as attaching points (Follner & Klarenberg 1995). In our experiments, we of- fered individual spiderlings optimal starting conditions, and we never observed the tip- toe ballooning mode. Follner (1994) suggested that “tip-toe” might be a tactical alternative for individuals in unfavorable starting points (e.g., overcrowded tips of grass blades). Our results show that it is not necessary for spiderlings of A. bruennichi to engage in ae- rial dispersal before building a prey-capture web. While ballooning is frequent in the field (Follner & Klareeberg 1995), it is clearly not an obligate part in the development of this species. In spite of the rapid expansion of the species over the past decades and the potential importance of aerial dispersal for colonizing new habitats, the role of ballooning in A. 520 THE JOURNAL OF ARACHNOLOGY bruennichi does not differ from A. trifasciata and A. aurantia where this phase in life his- tory is also not obligate (Tolbert 1977). The mortality of about 20% after 19 days in both experimental groups (difference statis- tically not significant) suggests that rearing of A. bruennichi spiderlings to adulthood may be challenging. Our rearing method based on a diet with Drosophila melanogaster, similar to Muller & Westheide (1993), worked well for our purpose, where we only tested the effects of ballooning in second-instar spiderlings on their ability to make their first web. On average, more then eight days elapsed before A. bruennichi spiderlings began to build their first prey-capture web. This ap- pears to be a surprisingly long period, because the animals can only feed once the first web is built. We cannot exclude that this is a lab- oratory artifact, for example due to unattrac- tive sites for web construction. However, the long latency did not interfere with the rearing regime. The animals appeared to be well adapted to temporary starvation because the mortality was low in this phase (Fig. 6). Also in the field, the spiderlings do not immediately start with prey-capture web construction (Foll- ner & Klarenberg 1995) and endure extended periods of starvation. Argiope spiderlings eas- ily survive several days nearby their cocoons, sometimes with communal meshworks of in- terlocking dragline threads (“communal tan- gles”) (Tolbert 1976, 1977; Follner & Klar- enberg 1995) where they find shelter until favorable weather or microclimate conditions allow for ballooning (Tolbert 1977; Follner & Klarenberg 1995; see also Suter 1999 for physics of ballooning). Argiope spiderlings actively select suitable web sites by ballooning, re-ballooning or walking (Enders 1973; Tolbert 1977; Follner & Klarenberg 1995). Also in this nonpreda- tory phase the spiderlings must avoid starva- tion. Tolbert (1976) kept A. aurantia spider- lings in the laboratory without food and water. Mortality remained moderate in these experi- ments for several days and only increased dis- tinctly about two weeks after hatching. The behavioral ballooning sequence could be easily triggered under artificial conditions in our study, suggesting that it will also occur in the field whenever environmental condi- tions allow. Therefore dispersal and popula- tion structure will be primarily driven by mi- croclimatic conditions in the local habitats. The local persistence of non-emigrants (non- ballooners and short-distance ballooners) in A. bruennichi populations might facilitate aggre- gated dispersion patterns, just as in weather phases which are unfavorable for aerial dis- persal. Given ballooning is a less effective means of long distance dispersal than previ- ously thought (Roff 1981; Decae 1987; Wise 1993; Bonte et al. 2003), this could also ex- plain the genetic differentiation among habitat patches in other Argiope species (Ramirez & Haakonsen 1999). The role of natural selection in range ex- pansion has recently been discussed for in- sects in the context of global warming (e.g., Pimm 2001; Thomas et al. 2001). However, improving environmental conditions at range margins can initiate range extensions purely on the basis of ecological, physiological and population-dynamic processes not requiring any evolutionary change (Thomas et al. 2001; see also Coope 1995; Williamson 1996). Our results are in line with these views and reject the hypothesis of Follner & Klarenberg (1995) that evolutionary processes have changed bal- looning behavior in newly founded popula- tions. ACKNOWLEDGMENTS We are grateful to Peter Neumann, Gail E. Stratton and two anonymous reviewers for helpful comments on previous drafts of the manuscript. We thank Christian WW. Pirk for statistical advice, Vlastimil Ruzicka and Theo Blick for providing literature. This study was supported by a research fellowship to A.W. from the State of Saxony-Anhalt and by the Deutsche Eorschungsgemeinschaft (DFG) grant to P.B. (BL 776/1-1). LITERATURE CITED Arnold, A. 1986. Die Wespenspinne Argyope bruennichi (Scopoli) im Stadtgebiet von Leipzig (Arachnida, Araneae). Entomologische Nachri- chten und Berichte 30:268-269. Bonte, D., N. Vandenbroecke, L. Lens & J.-P. Mael- fait. 2003. Low propensity for aerial dispersal in specialist spiders from fragmented landscapes. Proceedings of the Royal Society of London (B) 270:1601-1607. Bristowe, W.S. 1939. The Comity of Spiders. Vol. I. Ray Society, London. Coope, G.R. 1995. Insect faunas in ice age envi- ronments: why so little extinction? Pp. 55-74. In WALTER ET AL.— BALLOONING IN ARGIOPE BRUENNICHI 521 Extinction Rates (J.H. Lawton & R.M. May, eds.). Oxford University Press, Oxford. Coyle, EA. 1983. Aerial dispersal by mygalomorph spiderlings (Araneae, Mygalomorphae), Journal of Arachnology 1 1 :283-286. Crome, W. & I. Crome. 1961. Paarung und Eiablage bei Argyope bruennichi (Scopoli) auf Grund von Freilandbeobachtungen an zwei Populationen im Spreewald/Mark Brandenburg (Araneae: Aranei- dae), Mitteilungen aus dem Zoologischen Mu- seum in Berlin 37:189-252. Dean, D.A. & W.L. Sterling. 1985. Size and phe- nology of ballooning spiders at two locations in eastern Texas. Journal of Arachnology 13:111- 120. Decae, A.E. 1987. Dispersal: ballooning and other mechanisms. Pp. 348-356. In Ecophysiology of Spiders (W. Nentwig, ed.). Springer- Verlag, Ber- lin Heidelberg New York. Eberhard, W.G. 1987. How spiders initiate airborne lines. Journal of Arachnology 15:1-9. Enders, F. 1973. Selection of habitat by the spider Argiope aurantia Lucas (Araneidae). American Midland Naturalist 90:47-55. Follner, K. 1994. Ballooning und Mikrometeorolo- gie bei Argiope bruennichi (Scopoli, 1772) (Ar- aneae, Argiopidae). Diploma Thesis, Ludwig- Maximilians-Universitat Mlinchen, 76 pp. Follner, K. & A. Klarenberg. 1995. Aeronautic be- haviour in the wasp-like spider, Argiope bruen- nichi (Scopoli) (Araneae, Argiopidae). Pp. 66- 72. In Proceedings of the 15'^ European Colloquium of Arachnology, Ceske Budejovice 1994 (V. Ruzicka, ed.). Czech Academy of Sci- ences, Institute of Entomology, Ceske Budejo- vice. Gauckler, K. 1968. Argyope bruennichi, die schone Wespenspinne, durchwandert Franken, erreicht die Oberpfalz und das Bayerische Alpenvorland. Mitteilungen der Naturhistorischen Gesellschaft Numberg 2:1-5. Gillandt, L. & J.M. Martens. 1980. Verbreitung und Biotopbindung der Wespenspinne Argiope bruennichi im Landkreis Liichow-Dannenberg (Arach.: Araneae). Verhandlungen des naturwis- senschaftlichen Vereins Hamburg (NF) 23:309- 318. Guttmann, R. 1979. Zur Arealentwicklung und Okologie der Wespenspinne {Argiope bruenni- chi) in der Bundesrepublik Deutschland und den angrenzenden Landern. Bonner zoologische Bei- trage 30:454-486. Helsdingen, P. J. van. 1982. Postglaciale uitbreiding van Argiope bruennichi Scop, nu ook tot in Ned- erland. De Levende Natuur 84:121-123. Jonsson, LJ. & P. Wilander. 1999. Ar getingspin- deln, Argiope bruennichi, etablerad i Sverige? Entomologisk Tidskrift 120:17-21. Kohler, G. & G. Schaller. 1987. Untersuchungen zur Phanologie und Dormanz der Wespenspinne Argiope bruennichi (Scopoli) (Araneae: Aranei- dae). Zoologisches Jahrbuch fur Systematik 1 14: 65-82. Levi, H.W. 1983. The orb- weaver genera Argiope, Gea, and Neogea from the Western Pacific re- gion (Araneae: Araneidae, Argiopinae). Bulletin of the Museum of Comparative Zoology 150: 247-338. Lohmeyer, W. & P. Pretscher. 1979. Uber das Zus- tandekommen halbruderaler Wildstauden-Queck- en-Fluren auf Brachland in Bonn und ihre Be- deutung als Lebensraum fiir die Wespenspinne. Natur und Landschaft 54:253-259. Malt, S. 1996. Untersuchungen zur Rolle ausge- wahlter netzbauender Spinnen (Araneida) im tro- phischen Beziehungsgefiige von Halbtrockenra- sen. Dissertation, Friedrich-Schiller-Universitat Jena, 134 pp. Muller, M.C. & W. Westheide. 1993. Comparative morphology of the sexually dimorphic orb-weav- ing spider Argiope bruennichi (Araneae: Aranei- dae). Proceedings of the XII International Con- gress of Arachnology, Brisbane 1992. Memoirs of the Queensland Museum 33:615-620. Nielsen, E. 1932. The Biology of Spiders. With es- pecial Reference to the Danish Eauna. Vol. I & II, Levin & Munksgaard, Copenhagen. Nyffeler, M. & G. Benz. 1987. The foliage-dwelling spider community of an abandoned grassland ecosystem in eastern Switzerland assessed by sweep sampling. Mitteilungen der Schweizer- ischen Entomologischen Gesellschaft 60:383- 389. Pasquet, A. 1984. Predatory-site selection and ad- aptation of the trap in four species of orb-weav- ing spiders. Biology of Behaviour 9:3-19. Pimm, S.L. 2001. Entrepreneurial insects. Nature 411:531-532. Puts, C. 1988. lIArgiope fasciee en Belgique: ra- dioscopie d’une conquete. Bulletin des Reserves Naturelle et Omithologiques de Belgique 4:116- 118. Ramirez, M.G. & K.E. Haakonsen. 1999. Gene flow among habitat patches on a fragmented land- scape in the spider Argiope trifasciata (Araneae: Araneidae). Heredity 83:580-585. Richter, C.J.J. 1970. Aerial dispersal in relation to habitat in eight wolf spider species {Pardosa, Ar- aneae, Lycosidae). Oecologia 5:200-214. Roff, D.A. 1981. Dispersal in dipterans: its costs and consequences. Journal of Animal Ecology 46:443-456. Sacher, P. & P. Bliss. 1989. Zum Vorkommen der Wespenspinne {Argiope bruennichi) im Bezirk Halle (Arachnida: Araneae). Hercynia (NF) 26: 400-408. Sacher, P. & P. Bliss. 1990. Ausbreitung und Bes- tandssituation der Wespenspinne {Argiope bruen- 522 THE JOURNAL OF ARACHNOLOGY nichi) in der DDR — ein Aufruf zur Mitarbeit. Entomologische Nachrichten und Berichte 34: 101-106. Scharff, N. & S. Langemark. 1997. Hvepseedder- koppen Argiope bruennichi (Scopoli) i Danmark (Araneae; Araneidae). Entomologiske Meddelel- ser 65:179-182. Smithers, R 2000. Argiope bruennichi (Scopoli, 1772): a review of recent British records. Newslet- ter of the British arachnological Society 87:2-3. Suter, R.B. 1999. An aerial lottery: the physics of ballooning in a chaotic atmosphere. Journal of Arachnology 27:281-291. Thomas, C.D., E.J, Bodsworth, R.J, Wilson, A.D. Simmons, Z.G. Davies, M. Musche & L. Con- radt. 2001. Ecological and evolutionary process- es at expanding range margins. Nature 41 1:577- 581. Tolbert, W.W. 1976. Population dynamics of the orb weaving spiders Argiope trifasciata and Argiope aurantia (Araneae, Araneidae): density changes associated with mortality, natality and migra- tions. Ph.D. Thesis, University of Tennessee, Knoxville, 172 pp. Tolbert, W.W. 1977. Aerial dispersal behavior of two orb weaving spiders. Psyche 84:13-27. Weyman, G.S. 1993. A review of the possible caus- ative factors and significance of ballooning in spiders. Ethology Ecology & Evolution 5:279- 291. Wiehle, H. 1931. Araneidae. Pp. 1-136. In Die Tierwelt Deutschlands und der angrenzenden Meeresteile, 23. Teil. Spinnentiere oder Arach- noidea, VI. Agelenidae — Araneidae (F. Dahl, ed.). Gustav Fischer Verlag, Jena. Williamson, M, 1996. Biological Invasions. Chap- man & Hall, London. Wise, D.H. 1993. Spiders in Ecological Webs. In Cambridge studies in ecology (H.J.B. Birks & J.A. Wiens, eds.). Cambridge University Press, 328 pp. Manuscript received 17 September 2004, revised 10 August 2005. 2005. The Journal of Arachnology 33:523-532 CAN SIMPLE EXPERIMENTAL ELECTRONICS SIMULATE THE DISPERSAL PHASE OF SPIDER BALLOONERS? James R* Bell: Warwick HRI, Wellesbourne, Warwickshire CV35 9EF. E-mail: j .nbell @ warwick.ac.uk David A« Bohan and Richard Le Fevre: Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ Gabriel S« Weyman: Syngenta, Jealott’s Hill International Research Centre, Bracknell, Berkshire RG42 6EY ABSTRACT. Here we describe the structure of a fall speed chamber designed to measure, with low experimental error, the terminal velocities (fall speeds) of spiders of known weight and a given length of silk. We also describe the construction of a simulated individual (SI) which could later be used to estimate the distance travelled by ballooning spiders in the field. Our data and analysis suggest that Oedothorax spp. (Linyphiidae) and Pachygnatha degeeri (Tetragnathidae) individuals have fall speeds that can be described by their silk length and mass. Of the observed deviance in the fall speeds, 73.7% could be explained by a GLM model common to both species groups. Overlaying the SI fall speed data on this GLM surface suggests that the Sis have similar fall speed behaviors to spiders. However, further estimation is necessary before Sis could be considered valid models for evaluating spider ballooning distances. Keywords? Dispersal, ballooning, schottky diodes, silk, fall speed chamber Ballooning research has faced a seemingly intractable question for over 300 years: how far do ballooning spiders disperse once air- borne? While there have been attempts to ob- serve ballooning distances visually, which suggest that spiders move no more than a few hundred metres in any one attempt (e.g. MacCook, 1877; Folleer & Klareeberg 1995; Schneider et al. 2001), it also may be inferred from anecdotal evidence that spiders also make journeys of several hundred kilometres (Yoshimoto & Gressitt 1960; Okuma & Kisi- moto 1981). However, such visual and anec- dotal data are rare and have yet to yield any significant data for the great majority of bai- looners, including the linyphiids (Bell et al, 2005). Although models of ballooning dis- tance have been constructed (e.g. Thomas et al 2003), the predicted distances have yet to be verified. The lack of progress is perhaps surprising given recent advances in radar technology (Chapman et ai. 2003). Although Rothamsted Research’s vertical looking radar (VLR) can measure the horizontal speed, displacement direction, body alignment, mass and shape of flying insects up to 1 km above ground level (Chapman et al. 2003), as yet ballooning spi- ders cannot be uniquely identified. The VLR fails to resolve ballooeers because spiders lack distinctive allometric ratios and tend to have masses near or below the critical thresh- old for the radar (Chapman et al. 2003; Jason Chapman pers. comm.). Recently, indirect molecular genetic tech- niques have been employed as an alternative to measuring airborne spiders directly (Good- acre 2004). This approach was designed to de- tect the effect of dispersal rates on the genetic diversity of a number of key linyphiid popu- lations across the British mainland and its is- lands. The research has shown that popula- tions on islands have lower genetic diversity than those found on the mainland, implying changes in gene flow with isolation distance and island size. It should be noted however, that these findings were not independent of Wolbachia infections which confounded ob- served gene flow measurements. Other molec- ular studies, which indirectly estimate bal- looning distance using gene flow, have been conducted (as reviewed by Bell et al. 2005) 523 524 THE JOURNAL OF ARACHNOLOGY inductive loop to which a schottky diode is attached Figure 1. — A simulated individual, which shows a 16mm dipol antenna attached to an inductive loop. The 0.3mm schottky diode (not visible) is fixed to the inductive loop using gold spatter technology. The complete unit weighs 8mg. i but have not yielded estimates for the distanc- es travelled by individual spiders. We present an alternative approach, based upon synthetic models for spiders that we term simulated individuals (SI). Sis show great promise because they are traceable and biologically inert, thus resolving problems of airborne detection and Wolbachia infection. However, while we have begun to understand the physical properties of spiders and their im- plications for ballooning (Suter 1991, 1992, 1999), the properties of Sis are unknown and their comparative behavior remains untested. In this paper, a description of the technology and data used to compare SI properties against spiders is presented, concluding with a dis- cussion of future research prospects for bal- looning. METHODS Simulated individuals and spiders. — Males and females of Oedothorax spp. (mixed apicatus, fuscus and retusus species: Linyphi- idae) and Pachygnatha degeeri (Tetragnathi- dae) have been recorded ballooning many times (see world catalog in Bell et al. 2005). These species were used as model ballooners for comparison with a simulated individual j (SI). While the properties of an SI are yet to be established, the desirable traits should be that it: i) is structurally similar to a spider, consisting of a body and an associated silk component; ii) is able to generate its own drag to enable it to become airborne; iii) is trace- able, producing an automated signal of its lo- cation; iv) allows manipulation of the silk component to known levels of drag; and, last- ly v) is a 'nuir spider with no behavior which minimizes drag variability (i.e. absence of bit- ing and reeling of the silk line and reduced body posture modification). Schottky diodes, mounted onto an inductive loop with a dipole antenna (referred to as ‘diodes’ hereafter) have the potential to offer these properties, de- spite having none of the physical attributes of spiders (Fig. 1). We used 8 mg diodes in the following experiments. Spiders create drag with single or multiple silk lines that may account for 75% of the total drag of the spider (Humphrey 1987). For BELL ET AL.— SIMULATION OF BALLOONING DISPERSAL PHASE 525 I the diodes, simulated silk was adopted initial- ly as a replacement for natural silk. Titanium- coated fibre glass was identified as a possible solution and responded positively to very light convection currents (i.e. < 1 m/s). However, it proved to be fragile despite being four times the diameter (400 nm) of natural linyphiid silk (e.g. Tenuiphantes tenuis 100 nm). In light of these flaws, we used natural silk. Although linyphiid silk is too fine to manipulate easily, it was possible to attach the drag line silk of ; immature Araneus diadematus (Araneidae) to the diodes. For both spiders and diodes, all individuals were weighed before being intro- duced to the Rothamsted fall speed chamber described below. In total, 38 spiders (Oedoth- orax spp. n ~ 13; P. degeeri n = 25) and 4 diodes were dropped attached to silk lengths I within the range of 0-2.3 m. Rothamsted fall speed chamber. — The physical structure of the 9 m vertical chamber was relatively simple and included three de- ; tector stages and a hotwire (Fig. 2). The hot- wire was used as a silk-shearing mechanism to allow suspended spiders to be dropped j without human intervention. The first detector stage was used to manipulate the silk length, between 0.11 m and 2.3 m, at which a sus- ; pended spider or diode entered free fall and the two remaining stages measured the fall speed of each individual having reached ter- minal velocity. As a precursor to entering the chamber, spiders were first prompted to drop I down on a drag line from an oscillating probe. ' Having produced a dragline of >10 cm, spi- I ders were then fixed to the hotwire and al- ! lowed to pay out more silk until triggering the ' first detector stage (Fig. 3). In separate exper- iments, the diodes were suspended on fixed lengths of silk placed on the hotwire. The silk j was sheared by one of two methods: either a) I the spider broke the first detector stage light ! beam which automatically triggered the ho- ! twire (Figs. 2 & 3); or, b) if shorter lengths of silk (i.e. < 0.11 m), silkless drops or fixed I drops with diodes were required, a PC-oper- ated drop mechanism which manually trig- gered both the three detector stage light beams ' and the hotwire to an ‘on’ position was used. I After either the hotwire or manual drop had i been triggered, the spider or diode entered free fall for at least 5.4 m (i.e. depending on the first detector stage height) until it was mea- sured passing through the second detector stage at terminal velocity when timing started (Fig. 2). Timing was stopped, and the fall speed computed, when the individual passed through the third detector stage. The hardware environment behind the fall speed chamber measurements is based on the simple principle that when an object breaks a light beam, a passive record can be logged at a given point in time. Technically, the cham- ber included its own microprocessor controller based on a PIC16F876 running at 20 MHz and programmed using CCS PICC compiler (Fig. 4). This controller was connected to a PC run- ning dedicated software through a RS232 port, which allowed the user to control the light source and silk release mechanism (i.e. auto- matic/manual release) remotely. All control outputs were by opto-isolated open drain mos- fet drivers. The hardware detected falling ob- jects through the use of a medium area photo diode (41.3 mm^) connected to a two stage high gain amplifier. A first order bandpass fil- ter was used to remove unwanted signals be- low 300 Hz and above 5 KHz. The photo di- ode was mounted in a black box, with one end cut off, to help prevent ambient light interfer- ing with the source light. The initial design for the detection system was to incorporate a laser diode with line generator lens as the light source. However, the tested lasers were found to have a small but significant fluctuation in their output which made it impossible to dis- tinguish the object signal from noise when used in conjunction with the detection circuit. The circuit will need to be redesigned before lasers can be used in this application. As an alternative, high power quartz halo- gen bulbs (60 W) were used in conjunction with two 0.8 mm slits spaced at about 160 mm apart so that a reasonably fine line beam could be produced (Fig. 5). To focus the light onto the photo diode, Fresnel lenses (—300 mm wide, cut from a 280 mm square lens along the diagonal, 50 groves per inch and a focal length of 234 mm) were employed. The signal from the photo diode was then amplified and filtered before being applied to the single in- put channel of the analogue switch driven by a free running 3 KHz quartz clock (Fig, 4). The two output channels of this switch were then applied to the inputs of the voltage com- parator. Any low frequency variation of the input signal due to amplifier drift or ambient light falling on the photo diode was ignored 526 THE JOURNAL OF ARACHNOLOGY air movement due to chimney effect Figure 2. — Side view of the Rothamsted fall speed chamber. by the comparator. However, any object pass- ing through the light beam produced a much faster change in signal level which trigged the comparator. The comparator output was used as the trigger input to the microprocessor con- troller. The inherent precision of the micro- processor quartz clock ensured that the accu- racy of the system fell within at least ± 1 ms BELL ET AL.— SIMULATION OF BALLOONING DISPERSAL PHASE 527 Figure 3. — Top of Rothamsted fall speed chamber showing the hotwire release mechanism to which a spider is suspended on a silken line. (accuracy checked against a calibrated Systron Donner counter timer type 6250 A) and rep- resented a fall time recorder error of the spi- ders sampled between 0.0046-0.055%. Statistical analysis. — Fall speeds were an- alyzed using a Generalized Linear Model (GLM), the Normal distribution and the log- link function in Genstat (version 6, VSN in- ternational, Oxford, UK; McCullagh & Nelder 1989). Logio(Silk Length + 1) was fitted in the model as the explanatory variable, with spider species and logio(Spider Mass) as cov- ariates. The model fit was checked for over- dispersion in the data (McCullagh & Nelder 1989). The model’s standardized residuals were checked for linearity, leverage and ho- mogeneity (McCullagh & Nelder 1989). No attempt was made to fit a GLM to the provi- sional data for the diodes. RESULTS The GLM was found to fit the spider fall speed data extremely well, explaining some 73.7% of the GLM deviance observed (Fig. 6). The data were found to be underdis- persed, suggesting that the data were more regularly distributed than expected for data conforming to the Normal distribution. An empirical scale parameter was used to adjust the model fitted estimates of error to account for this underdispersion (see McCullagh & Nelder 1989). Spider fall speeds were found to decrease with increasing silk length (fi^g = 2.87, P = 0.004) and increase with an increase in spider body mass “ 3.25, P < 0.001). There was no interaction between silk length and spider body mass (tj 35 = 0.28, P = 0.78). No difference in the GLM was found with spider species ~ 0.70, P = 0.49), and no inter- action was found between spider species and silk length (ti 3^ = 0.62, P = 0.53) nor spider weight (^136 = 1.31, P = 0.19). Thus, a com- mon GLM was applicable to both Oedothorax spp. and P. degeeri: 528 THE JOURNAL OF ARACHNOLOGY Figure 4. — Block diagram detailing system electronics. logio(Fall Speed) = 3.83 - 0.951og,o (Silk Length +1) + l.lllogio (Spider Mass) Spider sex was a non-significant model covariate — 0.06, P = 0.95). However, the Oedothorax spp. are sexually dimorphic with respect to weight (females =2.4 ± 0.3 mg; males = 0.8 ± 0.004; n = 50.14, P < 0.001), yielding sex specific fall speeds for a given silk length in this species group. We plotted the fall speeds for diodes over the GLM in Fig. 6. The overlaid data suggests that diodes behave in a manner that is analo- gous to the spiders. BELL ET AL. — SIMULATION OF BALLOONING DISPERSAL PHASE 529 Photo diode and enclosure 0.29m 1.05m Light beam - Second 1mm slit detection zone Fresnel lens - 300 x 50mm (L X H) and FL = 234mm Halogen bulb and first 1 mm slit in metal enclosure \ n Senso' Plan view Figure 5. — Diagramatic view of the filtered light constantly monitored by the photo diode. DISCUSSION Rothamsted fall speed measurements.- — This experiment unequivocally demonstrates that natural spider silk can be attached to di- odes and that drag, and consequently fall speeds, can be systematically manipulated through the length of the silk line. The ob- served positive relationship between drag and silk length for Sis was analogous, but not identical, to spiders in free fall. Despite the limits of the provisional data presented, our results are supportive and imply that these di- odes represent a simple, yet viable paradigm of real spiders. Encouragingly, these diodes source (4=-) producing a large detection zone which is have the potential to develop our understand- ing of spider ballooning far beyond our pre- sent knowledge. Spider ballooning research is limited, al- though scientists are aware of the importance of silk in ballooning (Bell et ah 2005). For example, the effect of silk length on the fall speeds of spiders (Suter 1991), moth larvae (Lepidoptera) (Batzer 1968; Barel 1973; Mitchell 1979; McManus & Mason 1983; Ra- machaedrae 1987) and spider mites (Tetran- ychidae) (Jung & Croft 2001) has already been demonstrated. Of these, Suter’s (1991) seminal research attempted to evaluate fall 530 THE JOURNAL OF ARACHNOLOGY Figure 6. — Observed fall speed data for individuals of Pachygnatha degeeri (•), Oedothorax spp. (o) and diode (▼) against silk length and spider mass. The hatched surface represents the GLM fitted model for fall speeds with silk length and spider mass. times independently of human error. Even so, fall speeds still had to be estimated by extrap- olation because of the short fall distances in Suter’s experimental chamber. The advantages of the Rothamsted fall speed chamber are that measurements may be taken in near- still air conditions and when they have reached their terminal velocity, after individuals have fallen at least 5.4 m. Despite this, the results are sub- ject to error due to spider behaviors when fall- ing. Here no attempt was made to control for postural variation, such as spreading or with- drawing legs, which has been estimated to have up to a 10 fold effect on body drag (Su- ter 1992). Postural control may also be im- portant in mites which manipulate drag in a similar fashion to spiders and may be able to influence where they land (Jung & Croft 2001). Such postural control could account for some of the unexplained variation in our GLM and may be estimated by placing digital cam- eras inside the Rothamsted fall speed cham- ber. The posture of the photographed spiders, once categorized by shape, might then be in- cluded as a third covariate within the GLM. However, this behavior might only be expect- ed to account for a maximum of 25% of the observed variation (deviance) in the fall speed data (Suter 1991; see also Humphrey 1987). Allowing spiders to reach their terminal ve- locities over a comparatively large distance simplifies the mathematics of calculating ter- minal fall speeds, but also has the potential to be biologically erroneous. Purely from obser- BELL ET AL.— SIMULATION OF BALLOONING DISPERSAL PHASE 531 vatioe during handling, both species tested were able to pay-out silk at a rate of > 1 m/s, I particularly when individuals adopted "es- cape’ behaviors. If this payieg^out of silk oc- curred during freefall, thee the lightest indi- viduals could produce several meters of ‘undetected’ silk length during their fall. In practice this is unlikely, given that the residual variation was relatively small. While it is im- portant to highlight posture and silk reeling as sources of error, they are inescapable covar- iates of spider ballooning. Posture variation might be standardized, though not removed, ' by anaesthetising individuals with carbon di- oxide before entering the chamber (Jung & Croft 2001). However, this would have an im- pact on an individual’s ability to produce silk. The solution to evaluating the effects of pos- ture and variation in silk length during freefall can only be to increase the number of obser- vations (replicates). Suter (1991, 1992) recognized the impor- tance of spider mass, which served to increase the fail speeds at a given length of silk. Math- ematical models which seek to determine the : probability of dispersal based on a species by species account, also need to parameterize i mass and consider sex as a covariate where I obvious differences in males and females oc- j cur. As far as we are aware, this has been ig- ' noted to date. ! The future of schottky diodes to simulate the dispersal phase of spider ballooeers.— ■ This research has shown that simulating bal- I looeers has potential. Natural silk attached to diode bodies produces drag in a manner di- rectly analogous to a spider. Scanning har- monic radar has been shown to be effective in tracking diode-tagged bees for up to 900 m from the radar station (Osborne et al. 1999). To follow Sis, the use of a similar scanning radar set-up is planned. Using this technology i we can explore unanswered questions includ- i ing; how far do ballooeers travel; and, what is the pattern of dispersal of ballooners within a 1-2 km range? However, our research is at an early stage. While releasing diodes in the I field is the ultimate objective, several aspects I of SI behavior need further estimation before I SI ballooning data can be captured. Notably, the dependence of fall speeds on diode mass requires evaluation because the 8 mg diodes used do not represent the majority of ballooe- ers, which are under 2 mg (Greenstone et al. 1987); although much heavier spiders can be found ballooning. Reducing schottky diode mass by at least 75% would affect the drag dramatically and could require models for Sis that differ significantly from that estimated here for spiders. Only after completion of this diode model estimation phase of the project could radar-based fieldwork follow. ACKNOWLEDGMENTS Thanks are due to Alan Smith, Jason Chap- man, Kelvin Conrad (RR) and Andrew Mead (Warwick HRI) for their helpful discussions and advice. We also thank Ian Denholm (RR), David Skirvin and Rosemary Collier (War- wick HRI) for proof reading the MS. We are grateful for the assistance of Soeren Toft as ecology editor and thank the two peer review- ers for their comments. This paper is dedicat- ed to the memory of Julian Haughtoe, natu- ralist and friend. LITERATURE CITED Bare!, CJ.A. 1973. Studies on dispersal of Adoxo- phyes orana F.V.R. in relation to the population sterilization technique. Mededelingen Landbhoo- gesch, Wageningen 73:1-107. Batzer, H.O. 1968. Hibernation site and dispersal of spruce budworm larvae as related to damage of sapling balsam fir. Economic Entomology 61: 216-220. Bell, J.R., D.A Bohan., E.M. Shaw & G.S. Wey- man. 2005. Ballooning dispersal using silk: world fauna, phylogenies, genetics and models. Bulletin of Entomological Research 95:69-114. Chapman, J.W., D.R. Reynolds, & A.D. Smith, 2003. Vertical-looking radar: A new tool for monitoring high-altitude insect migration. Bio- science 53:503-511. Follner, K. & A.J. Klarenberg. 1995. Aeronautic be- haviour in the wasp-like spider Argiope bruen- nichi (Scopoli) (Araneae. Argiopidae). In: V. Ruzicka (ed.) Proceedings of the 15th European Colloquium of Arachnology. Ceske, Budejovice. pp. 66-72. Goodacre, S. 2004. Population structure and evolu- tionary relationships among linypMid spiders: a molecular analysis. In: Scientific abstracts of the 16* International Congress of Arachnology, Gent, Belgium 2-7 August 2004. p. 56. http://allserv. rug.ac.be/'^jpmaelfa/Abstracts%20Lezingen% 20(all).pdf Greenstone M.H., C.E. Morgan, A.L. Hultsch, R.A. Farrow & J.E. Dowse. 1987. Ballooning spiders in Missouri, USA, and New South Wales, Aus- tralia: family and mass distributions. Journal of Arachnology 15:163-170. 532 THE JOURNAL OF ARACHNOLOGY Humphrey, J.A.C, 1987. Fluid mechanical con- straints on spider ballooning. Oecologia 73:469- 477. Jung, C. & B.A. Croft. 2001. Aerial dispersal of phytoseiid mites (Acari: Phytoseiidae): estimat- ing falling speed and dispersal distance of adult females. Oikos 94:82-190. MacCook, H.C. 1877. The aeronautic flight of spi- ders. Proceedings of the Academy of Sciences of Philadelphia 1877:308-312. McCullagh, P. & J.A. Nelder. 1989. Generalized Linear Models, 2nd edition ed., Chapman and Hall. McManus, M.L. & C.L. Mason, 1983, Determina- tion of the settling velocity and its significance to larval dispersal of the gypsy moth (Lepidop- tera: Lymantridae). Environmental Entomology 12 pp 270-272. Mitchell, R.G. 1979, Dispersal of early instars of the Douglas-fir tussock moth. Annals of the En- tomological Society of America 72:291-297. Okuma, C, & R, Kisimoto. 1981. Air borne spiders collected over the East China Sea, Japanese Jour- nal of Applied Entomology and Zoology 25: 296-298. Osborne, J.L., S.J. Clark, R.J. Morris, I.H. Wil- liams, J.R. Riley, A.D. Smith, D.R. Reynolds & A.S. Edwards. 1999. A landscape-scale study of bumble bee foraging range and constancy, using harmonic radar. Journal of Applied Ecology 36: 519-533. Ramachandran, R. 1987, Terminal velocity of the first instar Ectropis excursia (Guenee) (Lepidop- tera: Geometridae). Proceedings of the Indian Academy of Sciences (Animal Science) 96:673- 678. Schneider, J.M., J. Roos, Y, Lubin & J.R. Henschel 2001. Dispersal of Stegodyphus dumicola (Ara- neae, Eresidae): they do balloon after all! Journal of Arachnology 29:114-116. Suter, R.B. 1991. Ballooning in spiders: results of wind tunnel experiments. Ethology, Ecology & Evolution 3:13-25. Suter, R.B. 1992. Ballooning: data from spiders in freefall indicate the importance of posture. Jour- nal of Arachnology 20:107-113. Suter, R.B. 1999. An aerial lottery: the physics of ballooning in a chaotic atmosphere. Journal of Arachnology 27:281-293. Thomas, C.EG., P. Brain & PC. Jepson. 2003. Ae- rial activity of linyphiid spiders: modelling dis- persal distances from meteorology and behav- iour. Journal of Applied Ecology 40:912-927. Yoshimoto, C.M. & J.C. Gressitt. 1960. Trapping of air borne insects on ships in the Pacific (part 3). Pacific Insects 2:239-243. Manuscript received 16 November 2004, revised 10 July 2005. 2005. The Journal of Arachnology 33:533-540 NOCTURNAL NAVIGATION IN LEUCORCHESTRIS ARENICOLA (ARANEAE, SPARASSIDAE) Thomas N0rgaard^: Department of Zoology, University of Zuerich, Winterthurerstrasse 190, CH"8057 Zuerich, Switzerland. E-mail: thomasn @ dr f n . org . na ABSTRACT. When the males of the Namib Desert spider Leucorchestris arenicola (Araneae, Sparas- sidae) reach the adult stage they undertake long nocturnal searches for females. From these searches they return to their home burrow often in a straight line only retracing a fraction of their outward path if at all. Distances of 40 m and 13 m are conservative estimates of the mean round trip length and maximum distance from the burrow. Returning to the starting point of a round trip of such length is theoretically only possible if the navigator uses external cues for positional reference. The possible involvement of a range of external cues in the male L. arenicola was investigated. The direction of gravity, the sun, polarized sunlight, olfaction, constant wind direction and vibrational beacons are ruled out or deemed unlikely to be involved in the spiders’ homing. Keywords* Homing, egocentric, geocentric navigation, path integration, dead reckoning Complex long distance navigation by ar- thropods is usually associated with the for- midable navigational capabilities of the eu- social hymenopterans such as bees and ants (e.g., von Frisch 1967; Wehner 1992). In spi- ders long distance traveling is most often done by ballooning involving extrusion of silk threads into the air (Suter 1991). This form of transportation is, however, only used by rela- tively small spiders. In large spiders such as the mygalomorph spider Aphonopelma hentzi Girad 1854 (Araneae, Theraphosidae) travels over long distances are by walking rather than ballooning. However, these spiders only do one-way excursions without returning to the starting point (Janowski-Bell & Horner 1999). Keeping a straight line so as not to end up at the starting point, which may represent an area where resources are overexploited or an area not well suited for a given life stage, might require actual navigation (Dacke et al. 2003). However, returning to the starting point of an excursion, i.e. showing homing behav- ior, is a far more demanding navigational task for an animal than a long walk in a chosen direction. In spiders, studies of homing have so far been reported to occur over distances of less than a meter (Seyfarth & Barth 1972; Sey faith * Current address: Gobabeb Training and Research Centre, RO. Box 953 Walvis Bay, Namibia, et al. 1982; Gomer & Class 1985; Dacke et al. 2001). However, in the central Namib De- sert a spider shows impressive skills of navi- gation. Henschel (1990, 2002) was the first to notice that the adult males of Leucorchestris arenicola Lawrence 1962 (Araneae, Sparas- sidae), like foraging bees or ants, also return to the starting point after excursions over dis- tances of tens of meters on the desert floor. The purpose of the present account is to outline the current state of knowledge about the mechanisms used or not used in the long distance homing navigation of L. arenicola, show new results concerning the role of vi- brational beacons, and finally point out the most promising leads that will be followed in future experiments. LEUCORCHESTRIS ARENICOLA AND ITS MOVEMENT PATTERNS Leucorchestris arenicola is an endemic sparassid (Jager 1999) of the Namib Desert. It is a large spider weighing up to 5 g (Henschel 1990), heavy enough to leave footprints in the sand (per. obs.). Adult males have standing leg spans often exceeding 10 cm (Fig. 1). Adult females have shorter legs but are usually slightly heavier than the males. Adult males comprise up to 12% of the population and oc- cur only in the summer period (September- April) (Henschel 1990). The spiders dig 30- 533 534 THE JOURNAL OF ARACHNOLOGY Figure 1. — Adult male L. arenicola showing protective coloration against the dune sand. Scale bar: 5 cm. 40 cm long burrows in the sand at an angle of ca. 30 degrees (Henschel 1990). This gets the spider to a depth of approximately 25 cm where climatic conditions are far more toler- able than on the desert surface (Henschel 1990). They are strictly nocturnal spiders, most frequently first becoming active an hour after sunset (Fig. 2). This was established us- ing infrared beam sensors and time-event data loggers (TinyTag). The beams were placed so they crossed the entrance of the burrows. Thereby the time a spider left the burrow was recorded. This activity pattern is probably an adaptation to the high temperatures in the de- sert during the day and the relative absence of predators at night (Cloudsley-Thompson 1983; Henschel 1990). Like many nocturnal desert spiders they have a light color render- ing them inconspicuous against the desert sand (Cloudsley-Thompson 1983; Dippenaar- Schoeman & Joque 1997). In the desert, the spider is found at the dune base where the sand is more stable and less stony compared to the slip face of the dunes and the gravel plains found between the dunes (for defini- tions of dune habitats see Robinson & Seely 1980). The spiders are highly territorial and defend an area with a radius of 3-4 m from their burrow (Henschel 1990; Birkhofer 2002). Especially burrow construction by an- other spider triggers strong aggressive behav- ior from a territory owner (Birkhofer 2002). Females and immature spiders mainly restrict their surface activity, e.g., prey capture, to within their territories. The main prey is te- nebrionid beetles. The prey are killed on the desert surface and then dragged into the bur- row (Henschel 1994). At the time the imma- ture spiders disperse from their maternal bur- row or when an adult female leaves her offspring, they may walk beyond their 3-4 m territory boundaries. However, these are one- way trips over distances far shorter than the roundtrip of the adult males. Observing the tracks of the spiders, it quickly becomes clear that adult males truly are the ones that regu- N0RGAARD^NOCTURNAL navigation in a sparassid 535 ,35 Hours after sunset Figure 2. — -Frequency of appearances from the burrows on the desert surface by male L. arenicola (n = 75) in relation to sunset (0 = sunset). Activity recordings were made with infrared beam sensors and time-event data-=loggers. larly wander far. The adult male spiders’ tracks can easily be identified by the size of the leg span and the conspicuous drum and scrape marks often seen on the paths (pers. obs.). When reaching the adult stage the male L. arenicola begins making long excursions searching for mating opportunities. These searches for the burrows of adult females are trips several orders of magnitude larger than I the spiders’ body size and over far longer diS“ I tances than their average territory size and were, therefore, described as long-distance ex- i cursioes (Heeschel 2002). The general layout I of the male spiders’ excursions can be divided into two sections: an outward path and a re- ! turn or homing path. The outward path is : characterized by a meandering and occasion- ally very tortuous searching walk, while the ! return often is a straight line walk heading to- : wards the burrow across ground not covered I on the way out (Fig. 3). By examining the general movement pat- tern of the male L. arenicola and drawing upon information from other navigating ar- thropods, especially spiders, we can list the probable methods male L. arenicola uses for homing. HOMING NAVIGATION Theoretically, a male L. arenicola could navigate to and from his burrow using two principally different methods. The spider could use either a geocentric or an egocentric system of references for determining his po- sition. If navigating by geocentric cues, the male spider must determine his position rela- tive to his burrow using landmarks in the sur- roundings. This requires memorization of a to- pographic map of the surroundings, also known as a cognitive map (Tolman 1948). The use of such a map has been suggested for hon- ey bees (Gould 1986). So far however, the ev- idence for this has not been conclusive and the behavior of navigating arthropods studied has been explained by simpler mechanisms than a memorized topographic map (Wehner & Menzel 1990). In such a eon-map fashion, iaedro.arks in the surroundings and the contour they present against the horizon are used in homing by wood ants {Formica japonica) (Fukushi 2001; Fukushi & Wehner 2004). If doing egocentric navigation the spider should assess his position in relation to Ms burrow by using information collected while he is walking. Therefore, instead of having a 536 THE JOURNAL OF ARACHNOLOGY map, the navigator continuously keeps track of all distances and directions traveled using this information to “calculate” the direction towards the burrow. This form of navigation is called dead reckoning or path integration (Mittelstaedt 1985). The necessary information about distances and directions steered can be obtained either ideothetically or allothetically (Mittelstaedt 1985). These two methods may be employed simultaneously. Ideothetic path integration implies that the spider navigates based entire- ly on internally gained information (Mittel- staedt 1985). This has been shown to be the case in the homing of the ctenid spider Cup- piennius salei Keyserling 1877 which can re- turn to its refuge using only information gath- ered from the lyriform organs (Seyfarth & Barth 1972). Pure ideothetic navigation is, however, susceptible to accumulation of errors ultimately leading to severe loss of precision. It is, therefore, only usable when navigating over shorter distances (Benhamou et al. 1990). When traveling the distances navigated by the male L. arenicola, external cues are, therefore, supposedly necessary. Doing path integration and using external cues is called allothetic navigation (Mittelstaedt 1985). A number of external cues are known to be used by several arthropods when they are navigating by use of path integration. The sun and the moon are well-known sources of directional informa- tion, used directly or indirectly via the polar- ized light patterns and spectral gradients they produce in the sky (Tongiorgi 1969; Rossel & Wehner 1986; Wehner 1994, 1997; Wehner et al. 1996; Dacke et al. 1999; Gal et al. 2001; Dacke et al. 2003). The direction of gravity (Bartels 1929; Hill 1979), constant wind di- rection (Wehner & Duelli 1971) and perhaps magnetism (Ugolini & Pezzani 1995) are also cues used by arthropod navigators. Often more than one of these external cues are used in order to achieve better precision. THE HOMING OF L. ARENICOLA Based on empirical and theoretical grounds several experiments were designed and carried out in search of the external cues used in the navigation of L. arenicola. To begin unravel- ing the mechanisms of homing navigation for male L. arenicola it is important to record and analyze paths in detail. A method to record the paths in all three dimensions was therefore Figure 3. — Trajectory of a single night excursion of a male L. arenicola projected onto a 2 dimen- sional plane viewed from above. Total path length was 810 m. developed (Nprgaard et al. 2003). A marker was placed along the paths each time the di- rection of the walk changed by more than the spiders leg span (approx. 5°). This divided the path into segments. The length of each seg- ment was measured using a tape measure, the direction with a compass and the slope with a digital inclinometer (Bosch DNM 60 L). i These recordings found a path length (mean ± s.e.) of 4092 cm ± 664 cm and a maximum distance to the burrow of 1313 cm ± 223 cm (Nprgaard et al. 2003). The longer the path, ! the more difficult complete tracking becomes. These path measurements were therefore bi- ased towards shorter distances as focus was solely on recording complete round trips. The area in which the recordings took place was densely populated by spiders and naturally bordered by interdune gravel plains and ripar- ian vegetation of the ephemeral Kuiseb river. Recent path recordings in another more open and less densely populated area have found far longer distances traveled by the spiders. An N0RGAARD— NOCTURNAL NAVIGATION IN A SPARASSID 537 approximately 810 m long path is the longest detailed round trip excursion recorded to date of any spider (Fig. 3). The ability to record the paths in all three dimensions allowed for an analysis of the slopes encountered by the spiders during their excursions. A constant slope of the substrate, i.e. direction of gravity, could potentially provide the spider with a us= able compass during its navigation. However, the sand surface of the desert is corrugated by the wind and no even slope existed, and use of the direction of gravity in the spiders' nav- - igatioe was therefore ruled out (N0rgaard et al. 2003) (Figs. 4-6). Long distance homing over ground which was not covered during the outward trip im= mediately excludes the use of pheromone trails. Direct homing by olfactory means is unlikely to function over the long distances the spiders travel; this is corroborated by ob- servatioes of spiders having different homing directions on the same night and the occur- I rence of changing wind directions. Directly using constant wind direction as a compass cue is unlikely for the same reason and be- cause of the turbulence at the surface caused by the sand ripples. Olfactory cues may how- ^ ever still be involved in the final pinpointing I of the burrow. Sand is the major component of the spiders’ habitat and one of the physical properties of :■ this substrate is its ability to conduct vibra- I tions as surface waves in the range between I 300-500 Hz (Brownell 2001). These frequen- i cies have wavelengths of 9-15 cm (Brownell I 2001), and the leg span of L. arenicola falls [ into this range. Some spiders are highly sen- sitive to vibrations detected by the lyriform organs on their legs (Foelix 1996), raising the possibility that the spiders could derive direc- tional information from a vibration source. With a geophone one can hear such sand vibrations. If vegetation hummocks have a distinct sound this might create a “sound landscape” with unique “landmarks” or sound beacons usable in the spiders’ naviga- I tion, in the same way as a visual landmark i possibly could. Therefore, an experiment was I carried out to investigate whether or not the I spiders could be using such sound beacons. I Two speakers were buried in the sand as bea- j cons. Beacon A was placed at a distance of 5 m from a male spider’s burrow and beacon B was placed 10 m away in the same direction. An amplifier (Star sound SSA-2040) and a MP3 player (Loomax 300 M), both powered by a 12 V battery, supplied the audio signal for the beacons. A continuous 300 Hz tone audible in the sand from a distance of at least 20 m was emitted from beacon A starting be- fore sunset. At night when the spider had left his burrow beacon A was switched off and beacon B switched on. In this way the position of the beacon was virtually shifted. In nine experiments, each with different males, no ef- fect of switching the position of a sound bea- con was found. All spiders behaved as if un- disturbed, searching for females, mating, and returning to their burrows as normal. Thus, while these spiders are likely to depend heavi- ly on vibration sensing for prey detection, this sensitivity does not appear to be important for navigation. Of the celestial cues available to the spi- ders, only the moon and the polarized light it produces need be considered here as they are strictly nocturnal. Individual bright stars, star constellations or perhaps the band formed by the Milky Way might also be used by the spi- ders as a compass cue. CONCLUDING REMARKS Many possible external cues are available to the navigating male L. arenicola and, as described above, a number of these have by now been ruled out entirely or must be con- sidered highly unlikely to be involved in the process. Of the possible non-visual external cues, magnetism remains to be tested. Mag- netism used for bipolar positional reference may be used by lobsters to return to a specific area (Boles & Lohmann 2003). This is not sufficiently precise to locate a tiny burrow en- trance in the desert floor. Moreover, the dis- tances over which L. arenicola wanders are probably too short to allow for magnetic nav- igation. Recent experiment has shown that vi- sion plays a role in the navigation done by L. arenicola (unpub. data). Thus, with our cur- rent knowledge, a visually based navigation system appears to be the most promising ex- planation of the remarkable homing behavior of L. arenicola. The necessity of visual cues has been shown in the wolf spider Lycosa ta- rantula (Linneaus 1758) (Ortega-Escobar 2002), even though it is navigating over dis- tances far shorter than what is seen in L. ar- enicola. 538 THE JOURNAL OF ARACHNOLOGY Figure 4-6. — 4. Path of a male L. arenicola spider projected on to a flat plane. The black star marks the burrow and the arrowheads indicate the direction in which the spider had walked. The small dots along the path each represents a marker put down for the path measurement. The numbers denote every fifth marker and thus path segment. 5. Elevation profile of the spider path illustrated in Fig, 4. The burrow is positioned at the zero elevation line. 6. Histogram showing the slope of each segment of the spider path illustrated in Fig, 4. The 0° line is horizontal. (Adapted from Nprgaard et al, 2003). N0RGAARD— NOCTURNAL NAVIGATION IN A SPARASSID 539 The experiments so far have been focused on the compass component of the spiders' navigation mechanism, but path integration also requires an odometer. Many other ques- tions call for investigation. For example, why do the males return to the burrow from which they started out? Is the energy cost of building a new burrow too high because it is necessary to have a deep burrow to survive high day- time temperatures? Or is it simply too risky to build a new burrow because of cannibalism (Henschel 1990; Birkhofer 2002)? Due to the scale of the excursions, most of the experiments with male L. arenicola can only take place in the field. The collection of data is, therefore, subjected to the constraints of the climate of the Namib Desert and the seasonal availability of adult males. These are conditions that may slow down, but not stop, the progress in gaining knowledge about the astounding homing navigation of L. arenicola males. ACKNOWLEDGMENTS I thank Joh R. Henschel, Rudiger Wehner, S0ren Toft and two anonymous referees for their comments on earlier drafts. This work was supported by the Swiss National Science Foundation (grant no. 31^61844.00 to R.W.). Also I thank the Namibian Ministry of Envi- ronment and Tourism and the Gobabeb Train- ing & Research Centre for permission to work in the Namib-Naukluft Park and for the use of facilities. LITERATURE CITED Bartels, M. 1929. Sirmesphysiologische und psy- chologische untersuchungen an der trichterspinne Agelena labyrinthica (CL). Zeitschrift ftir Ver- gleichende Physiologic 10:527-591. Benhamou, S., J.P. Sauve & P. Bovet. 1990. Spatial memory in large scale movements: efficiency and limitation of the egocentric coding process. Journal of Theoretical Biology 145:1-12. Birkhofer, K. 2002. Territorial behaviour of Leu- corchestris arenicola (Araneae: Sparassidae). M.Sc. thesis. Technische Universitat Darmstadt. Boles, L.C. & K.J. Lohmann. 2003. True navigation and magnetic maps in spiny lobsters. Nature 421: 60-63. Brownell, P. 2001. Sensory ecology and orienta- tional behaviors. Pp 159 — 183. In Scorpion Bi- ology and Research (P. Brownell & G. Polls eds.). Oxford University Press, Oxford, New York. Cloudsley-Thompson, J.L. 1983. Desert adaptations in spiders. Journal of Arid Environments 6:307- 317. Dacke, M., T.A. Doan & D.C. O'CarrolL 2001. Po- larized light detection in spiders. Journal of Ex- perimental Biology 204:2481-2490. Dacke, M., D-E, Nilsson, C.H. Scholtz, M. Byrne & E.J. Warrant. 2003. Insect orientation to po- larized moonlight. Nature 424:33. Dacke, M., D-E, Nilsson, E.J. Warrant, A.D. Blest, M.E Land & D.C. O’CarrolL 1999. Built-in po- larizers form part of a compass organ in spiders. Nature 401:470-473. Dippennaar-Schoeman, A. & R. Jocque. 1997. Af- rican spiders. An identification manual. Plant Protection Research Institute Handbook no. 9. Pretoria, South Africa. Foelix, R.E 1996. Biology of Spiders. 2"^* edition. Oxford University Press, Oxford, New York. Fukushi, T. 2001. Homing in wood ants, Formica japonica: use of the skyline panorama. Journal of Experimental Biology 204:2063-2072. Fukushi, T. & R. Wehner. 2004. Navigation in wood ants Formica japonica: context dependent use of landmarks. Journal of Experimental Biology 207:3431-3439. Gal, J., G. Horvath, A. Barta & R. Wehner. 2001. Polarization of the moonlit clear night sky mea- sured by full-sky imaging polarimetry at full moon: comparison of the polarization of moonlit and sunlit skies. Journal of Geophysical Re- search D19 106:22647-22653. Corner, P. & B. Claas. 1985. Homing behaviour and orientation in the funnel-web Spider, Agalena la- byrinthica Cierck. Pp 275 — 297. In Neurobiolo- gy of Arachnids. (EG. Barth ed.). Springer- Ver- lag, Berlin, Heidelberg, New York, Tokyo. Gould, J.L. 1986. The locale map of honey bees: do insects have cognitive maps? Science 232: 861-863. Henschel, J.R. 1990. The biology of Leucorchestris arenicola (Araneae: Heteropodidae), a burrowing spider of the Namib dunes. Pp 1 15 — 127. In Na- mib Ecology: 25 Years of Namib Research (M.K, Seely ed.). Transvaal Museum. Transvaal Muse- um Monograph No. 7, Pretoria. Henschel, J.R. 1994. Diet and foraging behaviour of huntsman spiders in the namib dunes (Ara- neae: Heteropodidae). Journal of Zoology, Lon- don. 234:239-251. Henschel, J.R. 2002. Long distance wandering and mating by the dancing white lady spider {Leu- corchestris arenicola) (Araneae, Sparassidae) across Namib dunes. Journal of Arachnology 30: 321-330. Hill, D.E. 1979. Orientation by jumping spiders of the genus Phidippus (Araneae: Salticidae). Be- havioural Ecology and Sociobiology 5:301-322. Jager, P. 1999. Sparassidae — the valid scientific 540 THE JOURNAL OF ARACHNOLOGY i name for the huntsman spiders (Arachnida: Ar- eneae). Arachnologische Mitteilungen 17:1-10. Janowski-Bell, M.E. & N.V. Homer. 1999. Move- ment of the male brown tarantula, Aphonopelma hentzi (Araneae, Theraphosidae), using radio te- lemetry. Journal of Arachnology 27:503-512. Mittelstaedt, H. 1985. Analytical cybernetics of spi- der navigation. Pp 298 — 316. In Neurobiology of Arachnids. (EG. Barth Ed,), Springer- Verlag, Berlin, Heidelberg, New York, Tokyo. N0rgaard, T, J.R. Henschel & R, Wehner. 2003, Long-distance navigation in the wandering desert spider Leucorchestris arenicola: can the slope of the dune surface provide a compass cue? Journal of Comparative Physiology A 189:801-809. Ortega-Escobar, J, 2002. Evidence that the wolf- spider Lycosa tarantula (Araneae, Lycosidae) needs visual input for path integration. Journal of Arachnology 30:481-486. Robinson, M.D. & M.K. Seely. 1980. Physical and biotic environments of the southern Namib dune ecosystem. Journal of Arid Environments 3:183- 203. Rossel, S. & R. Wehner. 1986. Polarization vision in bees. Nature (Lond.) 323:128-131 Seyfarth, E.A. & EG. Barth. 1972. Compound slit sense organs on the spider leg: mechanoreceptors involved in kinestetic orientation. Journal of Comparative Physiology 78:176-191. vjeyfarth, E.A., R, Hergenroder, H. Ebbes & EG. Barth. 1982. Ideothetic orientation of a wander- ing spider: compensation of detours and esti- mates of goal distance. Behavioral Ecology and Sociobiology 11:139-148. Suter, R.B. 1991. Balloning in spiders: results of wind tunnel experiments. Ethology, Ecology & Evolution 3:13-25. Tolman, E.C. 1948. Cognitive maps in rats and men. Psycological Review 55:189-208. Tongiorgi, P. 1969. Evidence of a moon orientation I, in the wolf spider Arctosa variane C. L. Koch ' (Araneae, lycosidae). Bulletin du Museum Na- '' tional d'Histoire Naturelle, Paris 41:243-249. i Ugolini, A & A. Pezzani. 1995. Magnetic compass and learning of the Y-axis (sea-land) direction in the marine isopod Idotea baltica basteri. Animal Behavior 50:295-300. von Frisch, K. 1967. The Dance Language and Ori- entation of Bees. The Belknap Press of Harvard University Press. Cambridge, Massachusetts. Wehner, R. 1992. Arthropods. Pp 45 — 144. In An- imal homing (F. Papi Ed.). Chapman and Hall, London. Wehner, R. 1994. The polarization- vision project: championing organismic biology, Fortschritte der Zoologie 39:103-143. ^ Wehner, R. 1997. The ant's celestial compass sys- tem: spectral and polarization channels. Pp 145 — 185. In Orientation and Communication in , Arthropods (M, Lehrer Ed.) Birkhauser Verlag, Basel. I Wehner, R. & P. Duelli. 1971. The spatial orienta- i; tion of desert ants, Cataglyphis bicolor, before i, sunrise and after sunset. Experientia 27:1364- ! 1366. i Wehner, R. & R. MenzeL 1990, Do insects have i: cognitive maps? Annual Review of Neuroscience 13:403-414. Wehner, R., B. Michel, & P. Antonsen. 1996. Visual i navigation in insects: coupling of egocentric and geocentric information. Journal of Experimental ; Biology 199:129-140. j Manuscript received 21 December 2004, revised 21 July 2005. i' 2005. The Journal of Arachnology 33:541-548 USE OF ANOPHELES-SFECIFIC PREY-CAPTURE BEHAVIOR BY THE SMALL JUVENILES OF EVARCHA CULICIVORA, A MOSQUITO-EATING JUMPING SPIDER Ximena J* Nelson^’^ Robert R. Jackson^’^ and Godfrey Sune^: 'Department of Psychology, Animal Behaviour Laboratory, Macquarie University, Sydney, NSW 2109, Australia. E-mail: ximena@galliform.bhs.mq.edu.au; ^School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand; ^International Centre of Insect Physiology and Ecology, RO. Box 30772-00100, Nairobi, Kenya. ABSTRACT. The prey-capture behavior of the juveniles of Evarcha culicivora, an East African mos- quito-eating jumping spider, was investigated in the laboratory using living prey and using dead, motion- less lures made from two mosquito species, Anopheles gambiae sensu stricto and Culex quinquefasciatus. Having tested individuals of E. culicivora that had no prior experience with mosquitoes (rearing diet: only chaoborid and chironomid midges), our findings imply that the small, but not the large, individuals of E. culicivora have an innate predisposition to adopt Anopheles-spQcific prey-capture behavior. Findings from lure tests implicate posture as a primary cue by which the small juveniles of E. culicivora identify Anoph- eles. Each individual of E. culicivora was presented with lures, that were either in the posture typical of Anopheles or in the posture typical of Culex. Small, but not large, juveniles of E. culicivora often re- sponded to Anopheles mounted in the Anopheles posture and Culex mounted in the Anopheles posture by taking an indirect route or a detour to the prey which enabled the salticid to approach the lure from behind. However, detours were not routine for small or for large individuals of E. culicivora when the lure, whether made from Anopheles or Culex, was in the Culex posture. When tested with live mosquitoes, small juveniles of E. culicivora were more effective at capturing Anopheles than Culex. Large juveniles were more effective than small E. culicivora juveniles at capturing Culex, but large and small juveniles had similar success at capturing Anopheles. Keywords: Salticidae, mosquitoes, malaria vectors, predation, detours, predatory versatility Distinctive prey-specific capture behavior has evolved in at least two groups of jumping spiders (Salticidae), the araneophagic species (i.e. species that prey especially on other spi- ders) and the myrmecophagic species (i.e. species that prey especially on ants). Some- times araneophagic and myrmecophagic sal- ticids use specialized tactics to target remark- ably specific prey. For example, Portia fimbriata (Doleschall 1859) from Queensland (Australia) adopts tactics that are specific to a particular prey species, Euryattus sp., a com- mon salticid in the same habitat (Jackson & Wilcox 1990, 1993a). Euryattus females are unusual among salticids because they make a nest by suspending a dead rolled-up leaf by silk lines from the vegetation. Portia fimbriata captures Euryattus females by mimicking the vibratory courtship displays of Euryattus males, luring the females out of their leaf nests. Here we consider another example of re- markable predatory specificity. In this in- stance, the predator is Evarcha culicivora We- solowska & Jackson 2003, a salticid that feeds especially often on female mosquitoes in the field (Wesolowska & Jackson 2003). Here we consider the specificity of the salticid’s pred- atory behavior for a particular mosquito ge- nus, Anopheles. Evarcha culicivora is known only from the vicinity of Lake Victoria in Kenya and Uganda. Its typical habitat is tree trunks and walls of buildings. When quies- cent, it hides in the grass or in other vegeta- tion close to the ground, but feeding individ- uals venture into more exposed locations, such as the inside walls of mosquito-infested hous- es. In preliminary observations, we noticed that the small juveniles, but not the large in- dividuals, of Evarcha culicivora appeared to 541 542 THE JOURNAL OF ARACHNOLOGY be influenced by the mosquito’s posture. In particular. Anopheles is a mosquito genus known for its distinctive resting posture (Cle- ments 1999): hind legs raised; abdomen an- gled up at about 45° from the surface on which the mosquito is standing; abdomen and proboscis form a straight line. This posture contrasts with the posture seen in other mos- quito species. For example, in Culex spp., the abdomen is held parallel to the substrate and the head is tilted ventrally. Larger individuals of Evarcha culicivora typically oriented towards the mosquito, re- gardless of its posture, and then adopted the type of prey-capture sequence that is typical of many salticid species (see Forster 1977, 1982; Richman & Jackson 1992), making a slow, direct approach, with its body lowered, pausing when close, fastening a dragline and then leaping onto the mosquito. However, when the salticid was a small juvenile of E. culicivora and the mosquito was an individual of Anopheles, approach was often by way of a detour that ended with the salticid moving in from behind, walking beneath the mosqui- to’s elevated abdomen, and attacking from un- derneath. If small juveniles of Evarcha culicivora grabbed hold of the dorsal thorax of Culex, and the attacked mosquito often flew away, then when the Culex took flight, the small ju- venile would often lose its grip and fall off. However, when the small juvenile grabbed hold of Anopheles" ventral thorax, it generally would hold on when the mosquito took flight, with the mosquito soon succumbing and drop- ping to the ground, with the salticid on board (Fig. 1). Here we investigate three hypotheses sug- gested by these preliminary observations: (1) the small juveniles, but not the larger individ- uals, of Evarcha culicivora adopt an innate Anopheles-spccific capture tactic; (2) small ju- veniles use the characteristic rest posture of Anopheles as a primary Anop/ze/^^'-identifica- tion cue; (3) their Anopheles-^p^cif^c tactic enables the small E. culicivora juveniles to be especially effective at capturing Anopheles. METHODS General. — All testing was carried out be- tween 0700 and 1900 h (laboratory photope- riod 12L:12D, lights on at 0700) at the Thom- as Odhiambo Campus (Mbita Point) of the Figure 1 . — Small juvenile of Evarcha culicivora feeding on female mosquito {Anopheles gambiae). After attacking by grabbing hold of mosquito’s pos- terior ventral thorax from underneath, the salticid has now shifted to feeding from the side of mos- quito’s thorax. International Centre of Insect Physiology and Ecology (ICIPE) in Kenya. The elevation of the campus at Mbita Point is 1200 m above sea level (0°25'S-0°30'S by 34°10'E- 35°15'E), with 900 mm of rainfall per annum and mean annual temperature of 27 °C. The salticids came from laboratory cultures (for standard salticid-laboratory procedures see Jackson & Hallas 1986). The salticids’ rearing environments were ‘enriched’ (spacious cag- j es, meshworks of twigs within each cage) in j a manner comparable to that described by Carducci & Jakob (2000). Maintenance diet consisted of letting each salticid feed to sati- ation three times per week (Monday, Wednes- day, Eriday) on midges (Chaoboridae & Chi- ronomidae) collected locally at Mbita Point as needed (i.e. the salticids had no prior experi- ence with mosquitoes of any kind). For testing, we used adult females of two mosquito species, Culex quinquefasciatus Say 1 823 and Anopheles gambiae sensu stricto Gi- les 1902. Body length of all mosquitoes used for testing (measured from the head’s anterior end to the abdomen’s posterior end, ignoring proboscis and ovipositor) was 4.5 mm (matched to the nearest 0.5 mm). Procedures for culturing A. gambiae were as described elsewhere (Gougana et al. 2004), and the cul- tures that we used were initiated from speci- mens collected at Mbita Point. Specimens of C. quinquefasciatus were collected as larvae NELSON ET AL.— SALTICID PREDATION ON ANOPHELES 543 I at Mbita Point and maintained in buckets ,i filled with lake water in the laboratory until I the adults emerged. Two size classes (matched to the nearest 0.5 I mm) of Evarcha culicivora juveniles were j used: ‘small’ (body length 1.5 mm) and f ‘large’ (body length 3.5 mm). The small ju- i veniles were individuals that had emerged from their brood sacs 5 days before testing and had not been fed. The large juveniles were kept without prey for 7 days before testing. The 5“day pre-test period was adopted with small juveniles because preliminary trials showed that recent hatchlings became notice- I ably weak after more than 6 days without food. The 7-day pre-test period was adopted for large juveniles because preliminary trials showed that most individuals respond to live prey and to lures after a fast of this length. No individual of E. culicivora and no individ- ual lure was used in more than one test. Data were analyzed using chi-square tests : of independence, with Bonferroni adjustments ' when multiple comparisons were made (Sokal ; & Rohlf 1995). Voucher specimens of Evar- cha culicivora have been deposited at the Mu- seum of Natural History (Wroclaw University, Poland), the National Museums of Kenya (Nairobi) and the Florida State Collection of Arthropods (Gainesville, Florida). Voucher specimens of insects have been deposited at the ICIPE Taxonomy Laboratory and at the Florida State Collection of Arthropods. Testing whether posture of the prey in- fluenced the decision by Evarcha to adopt -specific capture behavior. — Four lure types were made, two from using each of , the two mosquito species, with each species I being in one of two postures (the resting pos- 1 ture typical of Culex or the resting posture i typical of Anopheles). Each lure was made by immobilizing a mosquito with CO2 and then placing it in 80% EtOH for 60 min. The mos- quito was then mounted on the center of one side of a disc-shaped piece of cork (diameter ' 1.25 X the body length of the mosquito; thick- i ness 2 mm). For preservation, the lure and the cork were next sprayed with a transparent j aerosol plastic adhesive and left to air out for I at least 24 h before being used. ' All mosquitoes had been given blood 4-5 j h before being immobilized and used for mak- j ing lures. Previous work (unpubl. data) with I E. culicivora has shown that all instars of these salticids choose blood-fed mosquitoes when the alternative is mosquitoes that have not fed on blood. Each individual of E. culi- civora used for testing was assigned at ran- dom to one of four groups defined by mos- quito species and posture, with the proviso that the number for each group was the same {n = 50). Apparatus and testing procedures were sim- ilar to those detailed elsewhere (Li et al. 1996; Harland & Jackson 2000) except for modifi- cations that facilitated testing small juvenile salticids. The apparatus was a wooden ramp (15 mm thick, 40 mm wide, 140 mm long) that, with the support of a wooden dowel (15 mm thick), angled up at 20°. The ramp and supporting dowel were on a wooden base (50 mm wide, 150 mm long, 15 mm thick). A lure was positioned at the top of the ramp, in front of a wall which served as a background against which salticids could see the lure. The wall was a piece of brown wood (55 mm high, 40 mm wide, 15 mm thick) glued perpendic- ular to the top end of the ramp. The lure was centered on the ramp 15 mm from the base of the wall, leaving 10 mm between the wall and the top edge of the cork disc. The lure was positioned so that it faced 45° away from for- ward (i.e. for E. culicivora walking directly up the ramp, the lure was facing 45° to the left or the right). For each lure, whether it was faced left or right was decided a random. Before testing began, the salticid was kept in a covered pit (diameter 30 mm, depth 10 mm) drilled into the top surface of the ramp (equidistant from left and right side of ramp). The center of the pit was 50 mm from the bottom edge of the ramp (i.e. the lure was positioned 40 mm from the top end of the pit). Tests were allowed to start by removing a transparent glass plate used as a cover. After uncovering the pit, tests were aborted if the salticid failed to come out within 30 min or came out, but then moved off the ramp with- out first moving toward the lure. In successful tests, the salticid came out of the pit within 30 min after the cover was removed, walked up the ramp and, before 30 min elapsed after leaving the pit, contacted the cork disc or the mosquito, or both. The data we recorded were the salticid’s horizontal orientation to the lure and the path it took to reach the lure. Horizontal orientation of the salticid when approaching the lure was defined as follows: 544 THE JOURNAL OF ARACHNOLOGY Lure Figure 2. — Percentage of test spiders (juveniles of E. culicivora) that made detours when approaching lure (dead mosquito female mounted on cork disc). Two size classes of E. culicivrora were used: small (body length 1.5 mm) and large (3.5 mm). Four groups of spiders tested, each group defined by mosquito species and posture used for lures: Anopheles gambiae in Anopheles posture (A), A. gambiae in Culex posture (C), Culex qiiinqiiefcisciatus in Anopheles posture (A) and C. quinquefasciatus in Culex posture (C). For each bar, n = 50 (no individual of E. culicivora and no individual lure used more than once). front (no more than 45° to the left or the right of the anterior end of the sagittal plane of the mosquito’s head); rear (no more than 45° to the left or the right of the posterior end of the sagittal plane of the mosquito’s abdomen); side (between front and rear). “Detours” were defined as instances of salticids approaching the lure from the rear or else approaching the lure from the side in the first instance and then moving around to the rear. “Did not detour” was defined as instances of salticids approach- ing the lure from the front or approaching from the side without shifting to the rear. Testing for prey-capture success. — Large and small juveniles of Evarcha culicivora were tested. In each test, one E. culicivora ju- venile was put inside a clear Plexiglas box (300 mm X 300 mm X 300 mm) with one mosquito (one Anopheles or one Culex that had had a blood meal 4-5 h earlier). Obser- vations were terminated after the salticid cap- tured the mosquito or 30 min after the test elapsed without the salticid capturing the mos- quito. RESULTS Testing whether posture of the prey in- fluenced the decision by Evarcha to adopt Anopheles capture behavior. — When the lures were made from Anopheles, significantly more small juveniles (x^ = 43.46, P < 0.001, df = 1, « = 100), but not large juveniles (x^ — 0.64, P = 0.42, df = I, n = 100), of Evarcha culicivora made detours when the lure was in the Anopheles resting posture rather than in the Culex resting pos- ture (Fig. 2). Likewise, when the lures were made from Culex, significantly more small ju- veniles (x^ = 29.27, P < 0.001, df =1, ^ = 100), but not large juveniles (x^ = 0.09, P = 0.77, df =\, n = 100), of E. culicivora made detours when the lure was in the Anopheles resting posture rather than in the Culex resting posture. Small juveniles significantly more (Fig. 2) often than large juveniles of Evarcha culici- vora made detours when approaching Anoph- eles that were in the Anopheles resting posture (X^ = 55.85, P < 0.001, n = 100) and Culex ' that were in the Anopheles posture (x^ = | 46.54, P < 0.001, n = 100). However, the : numbers of small and large juveniles of E. [ culicivora that made detours when approach- i ing Anopheles in the Culex posture (x^ = 3.73, P — 0.05, n = 100) (Fig. 2) and Culex in the Culex posture (x^ = 2.25, P = 0.13, n = 100) were not significantly different. Prey-capture success. — Large and small juveniles of Evarcha culicivora had greater NELSON ET AL.— SALTICID PREDATION ON ANOPHELES 545 100 80 £ 70 m 8 60 o 3 50 M £ 40 3 30 CO o 20 10 0 Figure 3. — Percentage of test spiders (juveniles of Evarcha culicivora) that captur&d Anopheles gambiae and Culex quinquefasciatus in 30 min test (one spider and one mosquito put together in plexiglas box). N is indicated with each bar (no individual of E. culicivora and no individual mosquito used more than once). Two size classes of E. culicivrora: small (body length 1.5 mm) and large (3.5 mm) (assigned at random to test with one or the other mosquito species). Anopheles gambiae Culex quinquefasciatus Prey success at capturing Anopheles than Culex (small, "" 163.16, P < 0.001, n = 491; large, x" "" 17.78, P < 0.001, n = 594) (Fig. 3). Small juveniles were less successful than large juveniles at capturing Culex (x^ = 63.94, P < 0.001, n = 495), but large and small ju- veniles had similar success at capturing Anopheles (x^ = 4.13, NS with Bonferroni ad- justment, df = 1, n “ 590). DISCUSSION The distinctive resting posture of Anopheles appears to increase the vulnerability of these mosquitoes to predation by the small juveniles of E. culicivora. As shown by their response to our experiments with lures and despite their minute eyes, these small salticids can appar- j ently identify the stationary mosquito’s pos- ture by sight alone. Having identified the mos- quito’s posture, a small E. culicivora juvenile usually makes a detour that enables it to move under Anopheles' raised abdomen from be- hind. The posture of Culex does not afford the small juvenile with comparable easy access to the underside of the mosquito and, upon see- j ing a mosquito in the Culex posture, small E. culicivora juveniles usually do not make de- j tours. Evidently, small E. culicivora juveniles 1 have evolved fine-tuned innate tactics for pre- ; dation on Anopheles. j That Anopheles is generally an easier mos- quito than Culex for Evarcha culicivora to overpower is suggested by how both the large and the small juveniles of E. culicivora had greater success capturing Anopheles than Cu- lex. Furthermore, the limited strength of small juveniles is suggested by the finding that small juveniles were considerably less successful at capturing Culex than large juveniles, yet they were not less successful at capturing Anoph- eles. Evidently, the Anopheles-spcci^c tactic of small juveniles compensates for these spi- ders’ small size, enabling them to be as effec- tive as the larger juveniles when the prey is Anopheles. Large juveniles, being more ca- pable of overpowering the mosquito, usually take direct routes. This way they can quickly attack the mosquito, foregoing the lengthier detours adopted by small juveniles. Although Evarcha culicivora appears to be, along with examples from the myrmecophagic (Jackson & van Olphen 1991, 1992; Jackson & Wilcox 1993b; Jackson et al. 1998; Li & Jackson 1996a; Li et al. 1996; Li et al. 1999; Jackson & Li 2001) and the araneophagic sal- ticids (Li & Jackson 1996b; Li et al. 1997; Jackson & Li 1998; Jackson 2000; Harland & Jackson 2001; Cerveira et al. 2003), a species that adopts distinctive prey-specific prey-cap- ture behavior, E. culicivora seems to target a considerably different kind of prey. It is easy 546 THE JOURNAL OF ARACHNOLOGY il to appreciate how ants (Gillespie & Reimer 1993; Vieira & Hoefer 1994; Halaj et al. 1997; Nelson et al. 2004) and spiders (Foelix 1996; Persons & Rypstra 2000; Bames et aL 2002) can be dangerous prey for a salticid, as they have weapons, such as strong mandibles, strong chelicerae and venom, with which they can seriously, sometimes fatally, injure a saU ticid. However, mosquitoes appear to have no comparable weaponry with which to confront a salticid. Risk may be relevant when a mosquito flies away, with a salticid on board, because the salticid loses control over where it might be tossed. Landing in water or in a spider web, for example, might put a salticid in harm's way. However, in the evolution of Evarcha culicivora's prey-specific behavior, the risk of losing a meal may have outweighed these po- tential risks to life and limb. By attacking from underneath, the small juveniles of E. cuE icivora appear to minimize this risk of being thrown off by the mosquito in flight because they can hold on especially well after an at- tack from underneath. Another way in which Anopheles' posture may be important is by af- fording small juveniles of E. culicivora with the means of getting close without alerting a mosquito (i.e. it would be difficult for E, cuE icivora to move under Culex without first bumping into one of the mosquito's legs). Although it is known that spiders rely to a considerable extent on learned behavior (e.g., Grunbaum 1927; Bays 1962; Edwards & Jackson 1994; Punzo 2004), our methods ruled out prior experience with mosquitoes (i.e. the individuals used in this study had ei- ther not been fed at all, or fed on midges alone before testing). Evidently, an innate Anophe- les-specific tactic (taking a detour and attack- ing the mosquito from behind and underneath) is triggered when E. culicivora sees a mos- quito in the Anopheles posture. This innate tactic appears to be specific to a remarkably precise prey category, female mosquitoes from one particular genus. This study demonstrates another unusual example of prey-specific behavior in a salti- cid. Unlike the better-known examples of pro- nounced prey-specific prey-capture behavior in myrmecophagic and araneophagic salticids, E. culicivora' s Anopheles -spcci^c tactic ap- pears to be expressed by only the smaller ju- veniles. ACKNOWLEDGMENTS We are especially grateful to Hans Herren, ,, Louis-Clement Gouagna, John Githure, Bart I Knols and Charles Mwenda for the numerous '' ways in which they supported the research. Stephen Alluoch, Silas Ouko Orima, Jane j Atieno and Aynsley Macnab provided invalu- ; able technical assistance. For taxonomic as- ! sistance, we thank Arthur Harrison, G.B. Ed- | wards and Louis-Clement Gouagna. Our f research was funded in part by Grant UOC305 i from the Marsden Fund of the Royal Society ' of New Zealand and from Grants 4935-92 and :■ 6705-00 from the National Geographic Sod- i ety. LITERATURE CITED Barnes, M.C., M.H. Persons & A.L. Rypstra. 2002. The effect of predator chemical cue age on an- tipredator behavior in the wolf spider Pardosa ] milvina (Araneae: Lycosidae). Journal of Insect ,, Behavior 15:269-281. ^ Bays, S. M. 1962. Study of the training possibilities !' of Araneus diodematus CL Experientia 18:423- ■ 425. I Carducci, J.P. & E.M. Jakob. 2000. Rearing envL ronment affects behaviour of jumping spiders, i Animal Behaviour 59:39-46. Cerveira, A.M., R.R. Jackson & E.E Guseinov, j 2003. Stalking decisions of web-invading ara- neophagic jumping spiders from Australia, Azer- ^ baijan, Israel, Kenya, Portugal, and Sri Lanka: j the opportunistic smokescreen tactics of Brettus, Cocalus, Cyrba and Portia. New Zealand Journal of Zoology 30:21-30. - Clements, A.N. 1999. The Biology of Mosquitoes. ' CABI Publishing: Wallingford, England. i Edwards, G.B. & R.R. Jackson. 1994. The role of ' experience in the development of predatory be- ^ haviour in Phidippus regius, a jumping spider (Araneae, Salticidae) from Florida. New Zealand i Journal of Zoology 21:269-277. Foelix, R.F 1996. Biology of Spiders. Oxford Uni- ; versity Press. New York, Oxford, Pp. 240. Forster, L.M. 1977. A qualitative analysis of hunt- i ing behaviour in jumping spiders (Araneae: Sal- ■ ticidae). New Zealand Journal of Zoology 4:51- 62. Forster, L.M. 1982, Vision and prey-catching strat- - egies in jumping spiders. American Scientist 70: 165-175. Gillespie, R.G. & N. Reimer. 1993. The effect of alien predatory ants (Hymenoptera: Formicidae) on Hawaiian endemic spiders (Araneae: Tetrag- nathidae). Pacific Science 47:21-33. Gouagna, L.C., H.M. Ferguson, B.A., Okech, G.F. Killeen, E.W Kabiru, J.C. Beier, J.I. Githure & NELSON ET AL.— SALTICID PREDATION ON ANOPHELES 547 G. Yan. 2004, Plasmodium falciparum malaria disease manifestations in humans and traesmis= sion to Anopheles gambiae: a field study in West- ern Kenya. Parasitology 128:235-243. 1 Grunbaum, A. A, 1927. Uber das Verhalten der Spinne Epeira diademata, besoeders gegenuber vibratorischen Reizen. Psychologische For- schung 9:275-299. Halaj, J., D.W. Ross & A.R. Moldenke. 1997. Neg- ative effects of ant foraging on spiders in Doug- las-fir canopies. Oecologia 109:313-322. Harland, D.P. & R.R. Jackson. 2000. Cues by which Portia fimbriata, an araneophagic jumping spi- der, distinguishes jumping-spider prey from other prey. Journal of Experimental Biology 203: 3485-3494. I Harland, D.P. & R.R. Jackson. 2001. Prey classifi- cation by Portia fimbriata, a salticid spider that specializes at preying on other salticids: Species that elicit cryptic stalking. Journal of Zoology (London) 255:445-460. Jackson, R.R. 2000. Prey preferences and visual discrimination ability of Brettus, Cocalus and Cyrba, araneophagic jumping spiders (Araneae: Salticidae) from Australia, Kenya and Sri Lanka. New Zealand Journal of Zoology 27:29-39. Jackson, R.R. & S.E.A. Hallas. 1986. Comparative biology of Portia africana, P. albimana, P. fim- briata, P. labiata, and P. schultzi, araneophagic web-building jumping spiders (Araneae: Saltici- dae): utilisation of silk, predatory versatility, and intraspecific interactions. New Zealand Journal of Zoology 13:423-489. Jackson, R.R. & D. Li. 1998. Prey preferences and visual discrimination ability of Cyrba algerina, an araneophagic jumping spider (Araneae: Sal- ticidae) with primitive retinae. Israel Journal of Zoology 44:227-242. Jackson, R.R. & D. Li. 2001. Prey-capture tech- niques and prey preferences of Zenodorus dur- villei, Z. metallescens and Z orbiculata tropical ant-eating jumping spiders (Araneae: Salticidae) from Australia. New Zealand Journal of Zoology j 28:299-341. Jackson, R.R. & A. van Oiphen. 1991. Prey-capture techniques and prey preferences of Corythalia canosa and Pystira orbiculata, ant-eating jump- ing spiders (Araneae, Salticidae). Journal of Zo- ology (London) 223:577-591. Jackson, R.R. & A. van Oiphen. 1992. Prey-capture techniques and prey preferences of Chrysilla, I Natta and Siler, ant-eating jumping spiders (Ar- j aneae, Salticidae) from Kenya and Sri Lanka. I Journal of Zoology (London) 227:163-170. Jackson, R.R. & R.S. Wilcox. 1990. Aggressive mimicry, prey-specific predatory behaviour and predator recognition in the predator-prey inter- actions of Portia fimbriata and Euryattus sp.. jumping spiders from Queensland. Behavioral Ecology and Sociobiology 26:111-119. Jackson, R.R. & R.S. Wilcox. 1993a. Predator-prey co-evolution of Portia fimbriata and Euryattus sp., jumping spiders from Queensland. Memoirs of the Queensland Museum 33:557-560. Jackson, R.R. & R.S. Wilcox. 1993b. Observations in nature of detouring behavior by Portia fim- briata, a web-invading aggressive mimic jump- ing spider from Queensland, Journal of Zoology (London) 230:135-139. Jackson, R.R., D. Li., A.T. Bairion & G.B. Ed- wards. 1998. Prey-capture techniques and prey preferences of nine species of ant-eating jumping spiders (Araneae: Salticidae) from the Philip- pines. New Zealand Journal of Zoology 25:249- 272. Li, D. & R.R. Jackson. 1996a. Prey-specific capture behaviour and prey preferences of myrmeco- phagic and araneophagic jumping spiders (Ara- neae: Salticidae). Revue Suisse de Zoologie hors Serie:423-436. Li, D.Q. & R.R. Jackson. 1996b. Prey preferences of Portia fimbriata, an araneophagic, web-build- ing jumping spider (Araneae: Salticidae) from Queensland. Journal of Insect Behavior 9:613- 642. Li, D.Q., R.R. Jackson & A. Barrion. 1997. Prey preferences of Portia labiata, P. africana, and P. schultzi, araneophagic jumping spiders (Araneae: Salticidae) from the Philippines, Sri Lanka, Ken- ya, and Uganda. New Zealand Journal of Zool- ogy 24:333-349. Li, D.Q., R.R. Jackson & B. Cutler. 1996. Prey- capture techniques and prey preferences of Ha- brocestum pulex, an ant-eating jumping spider (Araneae, Salticidae) from North America. Jour- nal of Zoology (London) 240:551-562. Li, D.Q., R.R. Jackson & D.P. Harland. 1999. Prey- capture techniques and prey preferences of Aelur- illus aeruginosus, A. cognatus, and A. kocM, ant- eating jumping spiders (Araneae: Salticidae) from Israel. Israel Journal of Zoology 45:341-359. Nelson, X.J., R.R. Jackson, S.D. Pollard, G.B. Ed- wards & A.T. Barrion. 2004. Predation by ants on jumping spiders (Araneae: Salticidae) in the Philippines. New Zealand Journal of Zoology 31: 45-56. Persons, M.H. & A.L. Rypstra. 2000. Preference for chemical cues associated with recent prey in the wolf spider Hogna heiluo (Araneae: Lycosidae). Ethology 106:27-35. Punzo, F. 2004. The capacity for spatial learning in spiders: a review. Bulletin of the British Arach- nological Society 13:65-72. Richman, D. & R.R. Jackson. 1992. A review of the ethology of jumping spiders (Araneae, Sal- ticidae). Bulletin of the British Arachnological Society 9:33-37. 548 THE JOURNAL OF ARACHNOLOGY Sokal, R.R. & EJ. Rohlf. 1995. Biometry: The Prin- ciples of Statistics in Biological Research. Free- man: New York. Vieira, R.S. & H. Hoefer. 1994. Prey spectrum of two army ant species in central Amazonia, with special attention on their effect on spider popu- lations. Andrias 13:189-198. Wesolowska, W. & R.R. Jackson, R. 2003. Evarcha culicivora sp. nov., a mosquito-eating jumping spider from East Africa (Araneae: Salticidae), Annales Zoologici 53:335-338. Manuscript received 4 January 2005, revised 24 August 2005. 2005, The Journal of Arachnology 33:549-557 EGG SAC STRUCTURE OF ZYGIELLA X-NOTATA (ARACHNIDA, ARANEIDAE) T. Gheysees\ L. Beladjal\ K* Gellynck^, E. Van Nimmen^, L. Van Langenhove^ and J. Mertens^: ^Ghent University, Department of Biology, Terrestrial Ecology, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium. E-mail: Tom.Gheysens@UGent.be; ^Ghent University, Department of Textiles, Technologiepark 9, B-9052 Zwijnaarde, Belgium ABSTRACT* A detailed examination of the egg sac of Zygielia x-notata (Clerck 1757) revealed its structure, composition and different fibers. All egg sacs were composed of a basic layer, an insulation layer and an outer layer. The insulation layer consisted of two layers of cylindrical (or tubuliform) fibers with different diameters and probably with different mechanical properties. Knowing the complete struc- ture of the egg sac allows us to locate and extract the needed fibers for further research and to observe how the egg sac composition alters in relation to the habitat. Keywords: Cylindrical (tubuliform) fibers, sticky thread, major ampullate fiber, attachment disc, ad- hesion droplet. Of all natural fibers, silk is the most prom- ising for bioengineering because of its biolog- ical and mechanical properties. Much is al- ready known about the molecular and mechanical properties of dragline silk of Ar- aneus diadematus (Clerck 1757) and Nephila clavipes (Linnaeus 1767) (Shao et aL 1999; Vollrath 1999; Vollrath et al. 1998; Vollrath & Knight 2001); still, regarding the other spi- der silks, these properties have not yet been explored, especially the properties of egg sac silk. Egg sac threads can be very useful for biomedical applications, like sutures, cell sup- port and scaffolds (Gellynck et al. 2003; Van Nimmen et al. 2003). Before one can analyze these properties, the morphology of the egg sac must be investigated to locate these fibers. The primitive role of egg sacs is in giving protection against predators and parasites. Furthermore, the egg sac must create a good microclimate for embryological development, hatching and it must protect the spiderliegs until they leave the egg sac (Hieber 1985). As observed by De Bakker et al. (2002), there appears to be a big difference in egg sac struc- ture between families. Since egg sac threads are possibly the earliest silk used by spiders, it is clear that a detailed analysis of these structures can contribute to spider phylogeny. Further research on egg sacs of other families can perhaps determine, in time, the ancestral construction of the egg sac. In this study, only egg sacs of Zygielia x- notata (Clerck 1757) were investigated. Zyg- iella x-notata is iteroparous and females make most egg sacs in late autumn (Western Eu- rope; November-December). The spiderlings emerge around May of the following year. The egg sacs are elliptical and have a white to yel- lowish brown color. In addition, they have complex airy structures composed of different layers of silk that enclose and protect the eggs (De Bakker et al. 2002). In the present article, a more detailed description of the egg sac of Z. x-notata will be given in which its struc- ture, composition and different fibers will be- come clear. Knowing the complete egg sac structure will allow us to locate and extract the needed fiber types to investigate their use- fulness in several biomedical applications, and to investigate the alteration of egg sac com- position in relation to the habitat choice of Z x-notata. METHODS Egg sacs {n = 20) of Z x-notata were used to analyze their structure and composition. The spiders were collected in Ghent (Bel- gium) at Coupure Right (lat. 51°5'53", long. 3°7T11"), in the beginning of November (± 549 550 THE JOURNAL OF ARACHNOLOGY Table 1 . — Types of fibers measured by one spider (Z. x-notata). Type of fiber Average diameter (pm) St. Dev. MA threads Basic layer {n = 11) 2.63 0.13 Outer layer {n = 13) 2.41 0.35 Dragline {n = 10) 1.54 0.06 TU fibers 1st insulation layer {n = 231) 3.29 0.30 2nd insulation layer {n = 324) 3.84 0.24 80 spiders). They were kept in the laboratory in small, round, plastic (PS) cups, with plastic (PVC) lids (diameter: 50mm, hight: 25mm). The lids were pierced for aeration and to make them an easier surface for walking. All spiders were fed fruit flies {Drosophila sp.). A high level of air humidity was provided by a water- saturated piece of plaster placed on the bottom of the cups. It was moistened every 4 days with a mixture of water and nipagine (Alpha Pharma) to prevent fungal growth. The tem- perature was kept constant (23 ± 1 °C) and light was regulated following a constant day- night period (16h-8h). In this way, a high uni- formity in egg sac structure was obtained which simplified the observations. Draglines, for comparison, were collected from the spi- der while she was hanging. Voucher speci- mens have been deposited in the “Evolution- ary Morphology of Vertebrates & Zoology Museum”, Ghent University in Belgium (UGMD 104091). The morphological study was performed by | means of a stereomicroscope (Wild M5), a i; light microscope (Olympus CH-2) and a Scan- ning Electron Microscope (JEOL JSM-5600 LV, SEM). For light microscopy, slides of strands and connections were prepared with glycerine to prevent air bubbles. Photographs ^ taken by the light- and stereomicroscope were made with a camera (Nikon coolpix 900) mounted on the microscope. For the SEM, samples (connections, fiber types) and a com- pletely dried egg sac were mounted on stubs (standard and large (32mm)) and coated with gold (JEOL JFC— 1200 Fine coater, 8nm). An image processing system (Lucia System for Image Processing and Analysis version 4.51), made it possible to process and analyze real color photographs, like measuring the thick- ness of the draglines and egg sac fibers of one egg sac (Table 1). In order to compare the thickness of the egg sac fibers of the first and second insulation layer, a student t-test {P = 0.05) was performed supposing a normal dis- tribution of the measurements using the Sta- tistica program (StatSoft 2001 Release 6.0). Terminology, except as defined here, is from earlier studies (Peters & Kovoor 1991; ' Zschokke 1999, 2000; Benjamin et al. 2002). RESULTS [ All 20 egg sacs were analyzed and a great I uniformity in their structure was visible. Fig- : ure 1 shows a scheme of this uniform egg sac structure. Figure 1. — Older egg sac structure of Z x-notata: Left: side-view; right: top view, most common placement of the different structures. 1 = substrate (lid), 2 = basic layer (a = attachment discs of the basic layer, b = adhesion droplets), 3 = first insulation layer, 4 = second insulation layer, 5 = eggs forming the egg chamber, 6 = egg sac space, 7 = outer layer (c = attachment discs of the outer layer, d = sticky threads). GHEYSENS ET AL.— EGG SAC STRUCTURE OF ZYGIELLA X-NOTATA 551 Figures 2-4. — Basic layer. 2. Basic layer with an attachment of egg sac fibers of the first insulation layer on the left side (stereomicroscope); 3. Basic layer in more detail with some major ampullate (MA- MA) connections (SEM); 4. Two major ampullate (MA) threads of the basic layer (SEM). Basic layers. — When in the cup, the spi- ders generally walked on the lid (Fig. 1 (1)), and constantly produced draglines (major am- pullate (MA) thread), which they fixed on the substrate by means of attachment discs (Fig. 1 (2a)). In this way a parallel, sheet-like layer was formed and was here named the “walk plate”. The MA thread of this layer was al- ways doubled or sometimes fourfold (Figs. 3, 4). Before egg sac construction, an additional network of fibers was fixed on the center of the walk plate, which was here named the “stitch plate” (observed in 7 egg sacs), be- cause it was on this layer the tubuliform (TU) fibers were attached. In contrast with the walk plate, the stitch plate was a tightly woven net- work that was attached to the substrate by “adhesion droplets” (Fig. 1 (2b)). The threads resembled MA threads such as those of the walk plate but they had a smaller diameter. Attachment discs. — An attachment disc (Fig. l(2a)) consisted of an MA thread (Fig. 5, a) and a big sheet of finer fibers (= the disc; Fig. 5, b). The dragline was continuous and did not stop in the disc. The area of the disc varied and appeared to depend on the impor- tance of the attachment. The number of at- tachment discs per basic plate (Fig. I (2a)) was relatively small according to the number of attachment discs of the outer layer (Fig. l(7c)). Adhesion droplets. — Adhesion droplets (Fig. 1 (2b)) consisted of a glue-like droplet in which a fiber was placed without the pres- ence of a disc (Figs. 7, 8). These attachments were less abundant than attachment discs, they were more centrally placed in the egg sac (Fig. l(2b)) and seemed to attach the stitch plate to the substrate (observed in seven egg sacs). Figure 6 shows in detail an ending fiber in an adhesion droplet. The thread was double stranded and once in the droplet, it spread out in many finer fibers. This seemed to be dif- ferent from the other fibers (Fig. 7, 8) which were only connected to the substrate by means of a glue droplet. M[ajor ampullate (MA-MA) thread con- nection.— The basic layer contained many fixations between MA threads. These connec- tions could be complex and consisted of thread-like glue secretions (Figs. 9, 10). The secretions enveloped the two MA threads in- dividually and there was a stretch zone present in the secretion between the two MA threads (Fig. 10, a). Insulation layer. — This layer formed the actual egg sac and consisted of TU (tubili- form) fibers (Figs. 11, 12). The insulation lay- er could be subdivided into two layers: a “first” and a “second” insulation layer. These layers consisted of crisscrossed, tufted fibers with few or no connections (Fig. 11). Some- times TU fibers were found doubled (Fig. 11, a). They were lying next to each other in close contact, pointing in the same direction and seeming to adhere to one another along a fine line. TU fibers of the first insulation layer (Figs. 1 (3), 13) were attached to the basic layer (Fig. 2), after which the spider pulled the fi- bers out of her spinnerets and attached them again to the basic layer a bit further. In this way she spun around attaching the TU fibers, making a cup in which to put the eggs. After the eggs were laid (Fig. 1 (5)), a second in- sulation layer (Figs. 1 (4), 14) was placed over the eggs and the first insulation layer. This layer was also attached to the basic layer in 552 THE JOURNAL OF ARACHNOLOGY Figures 5-8. — Fiber-substrate connections. 5. Attachment disc, a = dragline, b = disc of finer fibers (SEM); 6. One adhesion droplet fixing an ending thread (stereomicroscope); 7. Fixation of a continuous fiber with one glue droplet (stereomicroscope); 8. Fixation of a continuous fiber with several glue droplets (stereomicroscope). the same way as the first insulation layer, but was more peripheral. Statistical analysis showed a highly signif- icant difference in thickness between the TU fibers of the first and the second insulation layer within the same egg sac. (df = 230, p<0.0001). The first insulation layer (3.29 pm ± 0.30 pm) had finer fibers than the second Figures 9-10. — SEM pictures of a basic layer MA-MA thread connection. 9. Overview of an attach- ment; 10. In more detail; a = stretch zone in the connecting secretion. GHEYSENS ET AL.— EGG SAC STRUCTURE OF ZYGIELLA X-NOTATA 553 Figures 1 1-14. — SEM pictures of the insulation layer. 11. Overview of the fibers of the insulation layer, a: doubled TU fiber; 12. Detail of a egg sac fiber; 13. Detail of a egg sac fiber of the first insulation layer; 14. Detail of a egg sac fiber of the second insulation layer. insulation layer (3.84 jjim ± 0.24 jjLm) (Table 1). Although they differed in thickness, they did not differ in surface structure and appear- ance (Figs. 13, 14). Egg sac fibers of the first and second insulation layer were connected to the basic layer in the same way (Fig. 15, 16). These connections consisted of four to six continuing TU fibers (Fig. 17, a) attached to- gether with a kind of glue to one MA thread (Fig. 17, b) of the basic layer. Egg sac chamber. — In recently placed egg sacs, only fibers surrounding the egg mass (Fig. 1 (5)) were found, and no fibers were found between the eggs. In older egg sacs, a Figures 15-17. — Egg sac fiber-basic layer connection. 15. Attachment of the first insulation layer to the basic layer (stereomicroscope); 16. Multiple egg sac fibers attached at one place on a basic layer thread (MA) (stereomicroscope); 17. Detail of multiple egg sac fibers (a) attached to a basic layer thread (b), somewhat torn apart (SEM). 554 THE JOURNAL OF ARACHNOLOGY Figures 18-19. — SEM picture of the outer layer. 18. Attachment of the outer layer thread to some egg sac fibers of the second insulation layer; 19. Detail of some outer layer threads, a = outer layer threads, b = egg sac fibers. space (Fig. 1 (6)) was found between the eggs and the first insulation layer. Outer layer. — The outer layer (Fig. 1(7)) was placed over the insulation layers and ba- sic layer. It was made up of an airy network of threads attached to the substrate by means of attachment discs (Figs. 1 (7c), 20). These attachment discs were numerous and periph- erally situated, forming the edge of the egg sac. They were similar in structure like those found in the basic layer. The threads were doubled and contained a kind of glue on their surface (Fig. 19, a) in contrast to the MA threads of the basic layer (Fig. 4). There were also connections found between the fibers of the outer layer (Fig. 18, a) and the second insulation layer (Fig. 18, b). These connections consisted, like MA-MA thread connections, of a kind of glue and finer fibers (Fig. 18). Although there was much more var- iation in the number of fibers included in this connection type, there were always two MA fibers (one MA thread), whereas the number of TU fibers was variable. In five egg sacs, a sticky thread was ob- served on the outer layer which resembled the sticky threads of orb webs (Figs. 1 (7d), 20- 22). The sticky threads were winded several times around the egg sac beginning from the attachment discs of the outer layer, going in- wards with an almost constant interval (Fig. 20, arrows). DISCUSSION Egg sacs in spiders vary substantially in structure interspecifically but usually display Figures 20-22. — Stereomicroscope pictures of the outer layer. 20. Outer layer with attachment discs and diagonally on it the sticky threads indicated with arrows; 21. Sticky threads with glue droplets; 22. Detail of the glue droplets. GHEYSENS ET AL.— EGG SAC STRUCTURE OF ZYGIELLA X-NOTATA 555 consistent similarities intraspecifically, that is, most families have egg sacs of only one or two structural types (Austin 1985). The egg sacs of Zygiella were fairly uniform and con- sisted of a basic layer, a double insulation lay- er and an outer layer. This structural compo- sition is consistent among most spiders belonging to the family Araneidae (unpub. data). Basic layer. — The basic layer is composed of a walk plate and a stitch plate and forms the basis for egg sac construction. The MA threads (draglines) of the walk plate are stron- ger and stiffer than TU fibers (Van Nimmen et al. 2003). Their presence in the basic layer and the outer layer is probably to fix the egg sac well to the substrate and to support the egg mass. The fact that the fibers of the stitch plate look like MA threads and have a smaller diameter suggests that they originate from the minor ampullate (MI) glands. The stitch plate is probably present in all egg sacs but due to the difficult observation of these fibers it was only detected in seven egg sacs. In nature however, one can sometimes ob- serve that a new, fresh egg sac is attached to an older one. In this case it can be expected that the ‘‘outer layer” of the older egg sac serves as a basic layer for the new one what costs the spider less energy. Attachment discs. — It is known that the piriform spools (Pi) produce the disc while the MA thread (dragline) is extruded by the MA spigots on the anterior spinnerets (Foelix 1996). The attachment discs used for the basic and outer layer strongly resemble those used for walking and web building described by Schutt (1996), so it is not so surprising that they are used for the egg sac construction as well. Adhesion droplets. — These adhesion drop- lets were no artifacts of oviposition, because no “glue droplets” were observed without fi- bers and the fixation of the fibers to the sub- strate was too specific. We suggest that the MI fibers of the stitch plate are fixed to the sub- strate in this way. The Pi spools are probably not involved in this fixation type because the MI spigots are located to far from them. A glue spool/spigot located closer to the MI spigot is more likely. We hypothesize that the ending thread seen in Fig. 6 does probably not originate from the minor ampullate spigots but rather from the MA spigot and that it is the beginning or end- ing of an MA thread, which would explain their low abundance. Major ampullate (MA-MA) thread con- nection.— The thread-like glue secretions make us suspect that these connections origi- nate from the Pi spools of the anterior spin- nerets, like the attachment discs. Most of the connections and supporting structures in ara- neoid webs are also made up of attachment discs: sticky spiral thread to radius, auxiliary spiral to radius, radius to frame and some con- nections to the hub (Peters & Kovoor 1991; Peters 1993; Benjamin et al. 2002). The stretch zone indicates that the fibers were laid parallel while fixing and reorientated after- wards. Insulation layer,- — Zygiella x-notata si- multaneously produces six TU fibers by six TU spigots, two on the median spinnerets and four on the posterior spinnerets (Foradori et al. 2002), forming the two insulation layers in the egg sac. Measuring the thickness of the fibers of the two insulation layers was only executed on one egg sac but recent experi- ments (unpub. data) indicate that the found difference in thickness is generally applicable. The difference in thickness between the fibers of the two layers is most likely due to the difference in volume of the TU glands before and after placing the eggs. Before oviposition, the glands are limited in space for expansion. Probably a smaller lumen is causing a smaller secretion which results in a finer fiber of the first insulation layer. After oviposition, more space is available in the abdomen causing a bigger lumen and secretion, resulting in a thicker fiber of the second insulation layer. This difference in diameter will probably also be reflected in the mechanical properties of these fibers. MA and TU fibers from Z. x-notata are morphologically very different. MA fibers have a smaller diameter (Tabel 1) and no un- derlying structures (Fig. 4), unlike TU fibers (Fig. 12) (Van Nimmen et al. 2003). The fact that TU fibers are another type of fiber means that they have some advantages compared to MA fibers: 1. In water, TU fibers only increase in diameter without longitudinal shortening (pers. obs.), unlike draglines which supercon- tract (Bell et al. 2002). This observation sug- gests that TU fibers can play a role in moisture regulation in the egg sac. This would also con- 556 THE JOURNAL OF ARACHNOLOGY firm the suggestion of Hieber that TU fibers can serve as a regulator of the relative humid- ity by taking up water if the relative humidity is too high and releasing water if it is too low. It would also explain why fibers of the second insulation layer are thicker than the first, be- cause the second insulation forms the actual barrier with the environment.2. The fact that TU fibers do not supercontract is probably also favorable for the survival of the eggs. If TU fibers should supercontract, the eggs would probably be killed if egg sacs are placed in humid environments. 3. It has been suggested that the egg sacs of N. clavipes pro- tect the egg mass against micro-organisms (Austin 1985). Like sticky threads, it is pos- sible that TU fibers possess a high concentra- tion of potassium dihydrogen phosphate to prevent the eggs from bacterial and fungal degradation (Schildknecht et al. 1972).4. The tufted nature of egg sac fibers protects the eggs against mechanical damage, predation or parasitism (Guarisco 2001) and can also save the eggs and spiderlings from drowning and physical damage (Hieber 1992a, b). Egg sac fibers are always attached to the threads of the basic layer or, in nature, the MA threads of webs or the outer layer of egg sacs and never to the substrate. Apparently they can only be attached to other fibers. The TU fibers (of the two layers) are also attached to other fibers with glue-like fibers, which in- cludes both. The origin of these secretions is however unknown. Egg sac chamber. — In recently placed egg sacs, the eggs are encircled by the first and second insulation layer which forms the egg chamber (Fig. 1 (5)), which is here the same as the egg sac chamber. By older egg sacs however, the mass of the eggs and the upside- down position (horizontal or vertical) causes the egg sac to sag out due to gravity, forming a space (Fig. 1 (6)) between the eggs and the first insulation layer. So here the egg sac chamber is the total of the “egg chamber” (Fig. 1 (5)) and the “egg sac space” (Fig. 1 (6)). It is possible that this egg sac space is vital for the hatching and the survival of the young spiderlings till the first ecdysis. Outer layer. — The threads of the outer lay- er are double stranded, attached to the sub- strate with attachment discs and have a similar morphology like draglines. All these observa- tions suggest that these threads originate from the MA spigots. If the MA threads used for the egg sac are compared with the dragline of the same spider, a remarkable difference (= Ipm) in thickness can be seen (Table 1). Because both fibers are from MA gland origin, this dif- ference can only be explained by the way they were produced. As found by Vollrath et al. (2001), both thread extension and reeling speed affect the diameter of the thread by a constant temperature. Draglines in this experiment were collected from hanging spiders which resulted in a bigger thread extension as well as a higher reeling speed resulting in a fine thread. Threads used for the egg sac structure are never stretched in this way because the spider is at all time attached to the substrate with her legs and the reeling speed was like the walking speed of the spider, resulting in a thicker thread. The difference in thickness between the fibers of the basic layer and the outer layer (Table 1) is probably due to a greater thread extension in the fibers of the outer layer caused by a bigger load on the thread from the mass of the spider. The connections of the outer layer threads to the second insulation layer fibers have probably the same origin as the MA-MA thread connections. In contrast with A. aurantia, the outer layer of Z. x-notata is not as dense, which would suggest that it is rarely or never attacked by generalist predators. This, however, is not so. These egg sacs were protected against preda- tors by use of a defensive layer made up of sticky threads. These sticky threads were only found on the outer layer of the egg sac in a very specific arrangement. It is very likely that these sticky threads are the same sticky thread as those used in the sticky spiral of webs and that they fulfill the same function. The fact that Z. x-notata is iteroparous means that she can replace the sticky thread of the egg sac when it dries out. This sticky thread was only observed in five egg sacs where mites were present in the cup, which may ex- plain the extra security in contrast with the other observed egg sacs. In conclusion, egg sacs are built of four lay- ers; a basic layer, a first insulation layer, a sec- ond insulation layer and an outer layer. This study shows for the first time the details of the different fibers involved in the egg sac, their possible function, their connection types and the role of the different structures they form in the egg sac. The basic- and outer layer are GHEYSENS ET AL.— EGG SAC STRUCTURE OF ZYGIELLA X-NOTATA 557 formed of MA threads which are for support and attachment of the TU fibers. In contrast, the insulation layers are made up by TU fibers and arranged in two layers. The fibers of the first insulation layer are finer than those of the second insulation layer which could indicate that the second insulation layers has a mois- ture regulation function in the egg sac. In some egg sacs there was an additional fiber type present, namely sticky threads. These sticky threads were found on the outer layer and probably protect the egg sac against gen- eralist predators, such as mites. ACKNOWLEDGMENTS Many thanks to Renaat Dasseville for tak- ing SEM pictures and for help interpreting the SEM photographs. The first author especially wants to thank Samuel Zschokke (Basel, Zwitzerland) for reviewing and helping finish this manuscript. His comments and sugges- tions seriously improved the quality of this manuscript. Also many thanks to Rieke Mooens for improving the English of this manuscript. This study was made possible through a special Research Fund of Ghent University (BOF, 01111902). LITERATURE CITED Austin, A.D. 1985. The function of spider egg sacs in relation to parasitoids and predators, with spe- cial reference to the Australian fauna. Journal of Natural History 19:359-376. Bell, EL, IJ. McEwen & C. Viney. 2002. Super- contraction stress in wet spider dragline. Nature 416:73. Benjamin, S.P., M. Duggelin & S. Zschokke. 2002. Fine structure of sheet-webs of Linyphia trian- gularis (Clerck) and Microlinyphia pusilla (Sun- devall), with remarks on the presence of viscid silk. Acta Zoologica 83:49-59. De Bakker, D., K. Gellynck, E. Van Nimmen, J. Mertens & P. Kiekens. 2002. Comparative struc- tural analysis of cocoon and cocoon silk in three spider species through Scanning Electron Mi- croscopy. Samu, F. & C. Szinetar, Poceedings of the 20** European Colloqium of Arachnology, Szombathely 22-26 July 2002, European Arach- nology 2002:81-87. Foelix, R. Biology of Spiders. 1996. New York, Oxford Univ. Press. 330 pp. Foradori, M.J., J. Kovoor, M-J. Moon & E.K. Til- linghast. 2002. Relation between the outer cover of the egg case of Argiope aurantia (Araneae Araneidae) and the emergence of its spiderlings. Journal of Morphology 525:218-226. Gellynck, K., P. Verdonk, F. Almqvist, E. Van Nim- men, D, De Bakker, L. Van Langenhove, J. Mer- tens, G. Verbruggen & P. Kiekens. 2003. A spi- der silk supportive matrix used for cartilage regeneration. Tissue Engineering 9:813-814. Guarisco, H. 2001. Description of the egg sac of Mimetus notius (Araneae, Mimetidae) and a case of egg predation by Phalacrotophora epeirae (Diptera, Phoridae). Journal of Arachnology 29: 267-269. Hieber, C.S. 1985. The “insulation” layer in the cocoons of Argiope aurantia (Araneae, Aranei- dae). Journal of Thermal Biology 10:171-175. Hieber, C.S. 1992a. The role of spider cocoons in controlling desiccation. Oecologia 89:442-448. Hieber, C.S. 1992b. Spider cocoons and their sus- pension systems as barriers to generalist and spe- cialist predators. Oecologia 91:530-535. Peters, H.M. 1993. Functional organization of the spinning apparatus of Cyrtophora citricola with regard to the evolution of the web (Araneae, Ar- aneidae). Zoomorphology 113:153-163, Peters, H.M. & J. Kovoor. 1991. The silk-producing system of Linyphia triangularis (Araneae, Liny- phiidae) and some comparisons with Araneidae. Zoomorphology 111:1-17. Schildknecht, H,, P. Kunzelmann, D. Kraub & C. Kuhn. 1972. Uber die Chemie der Spinnwebe, 1. Naturwissenschaften 59:98-99, Schiitt, K. 1996. Wie Spinnen ihre Netze befestigen. Mikrokosmos 85:273-278. Shao, Z., X.W. Hu, S. Frische & F. Vollrath. 1999. Heterogeneous morphology of Nephila edulis spider silk and its significance for mechanical properties. Polymer 40:4709-4711. StatSoft. 2001. STATISTICA for Windows, (data analysis software system), Release 6,0. Stasoft Inc., Tulsa. Van Nimmen, E., P. Kiekens and J. Mertens. 2003. Some material characteristics of spider silk. In- ternational Journal of Material & Product Tech- nology 18:344-355. Vollrath, E, D.P. Knight & X.W. Hu. 1998. Silk production in a spider involves acid bath treat- ment. The Royal Society 265:817-820, Vollrath, F. 1999. Biology of spider silk. Interna- tional Journal of Biological Macromolecules 24: 81-88. Vollrath, E & D.P. Knight. 2001. Liquid crystalline spinning of spider silk. Nature 410:541-548. Vollrath, E, B. Madsen & Z. Shao. 2001. The ef- fects of spinning conditions on the mechanics of a spider’s dragline silk. Proceedings of the Royal Society of London 268:2339-2346. Zschokke, S. 1999. Nomenclature of the orb- web. Journal of Arachnology 27:542-546. Zschokke, S. 2000. Radius construction and structure in the orb-web of Zilla diodia (Araneidae). Journal of Comparative Physiology A 186:999-1005. Manuscript received 12 January 2005, revised 10 September 2005. 2005. The Journal of Arachnology 33:558-561 NOTES ON THE NATURAL HISTORY OF A TRAPDOOR SPIDER ANCYLOTRYPA SIMON (ARANEAE, CYRTAUCHENHDAE) THAT CONSTRUCTS A SPHERICAL BURROW PLUG Astri Leroy and John Leroy: PO Box 390, RUIMSIG, 1732, South Africa. E-mail: astri @jmLco.za ABSTRACT. Burrows of an unidentified species of Ancylotrypa Simon from the floodplain of the Nyl River in Limpopo Province, South Africa are described. In addition to constructing a thin trapdoor, mem- bers of this species construct a hard, spherical plug or marble from soil particles held together with silk. Burrow structure, the plug and associated behavior are described for the first time. Keywords: Marble spiders, burrows, spherical plug, trapdoor, Ancylotrypa The genus Ancylotrypa Simon 1889 contains 48 species 32 of which occur in southern Africa (Dip- penaar-Schoeman 2002). Members of this genus construct and occupy silk-lined burrows that vary from simple, single-tube structures to Y or U shaped configurations or burrows with multiple arms, not all of which necessarily reach the soil surface. Various forms of soft lids close the burrow entrances (e.g., Dippenaar-Schoeman 2002: p.43 fig. 26). Although burrows of several species of An- cylotrypa have previously been described, this is the first species to be shown to construct a spherical plug or “marble” which is used to close and pos- sibly protect the buiTow. This study was conducted on the floodplain of the Nyl River (24°39"S: 28°42E) in the Limpopo Province of South Africa at Nylsvley Nature Re- serve from the summer of 1992-1993 through 2002-2003. The floodplain is usually inundated during the southern summer (November-March) but in years of poor rainfall the area remains en- tirely dry. When several years of exceptionally high rainfall occur it may remain inundated for more than one season (Barnes et al. in press). In the late 1980s we observed colonies of trap door burrows in the sodic alluvial soils of the Nyl River floodplain. Small spheres made of tight packed sand resembling tiny marbles of various siz- es were noted lying on the ground in the vicinity of the colonies but the connection between these marbles and the spiders that produce them was not made until 1992-1993 when burrows were exam- ined in detail. Burrows were excavated at different times of the year: during dry and wet seasons and in years of average, low and high rainfall. A colony would be located, the ground swept with a hard floor brush and a burrow chosen for excavation. Burrow lids were gently scratched to ascertain which were oc- cupied and it was found that if spiders were present, they would tug at the lids which, because they are soft, made them cave inwards. If the lids were scratched too hard, movement would cease and it was assumed the spider had retreated lower into its burrow. A hole was dug vertically about 60 mm distant from the chosen burrow lid to a depth of about 200 mm at what was hoped to be more or less parallel to the burrow and the hard soil between the initial hole and the burrow was removed. Once the main arm of a burrow was located even more careful digging was carried out to find the direction of the side arm until the whole buiTow was located. The burrow was measured and only then would the burrow wall be breached below the junction of the arms and subsequently sectioned from the bottom towards the top. Burrow shape, the spider, any young or eggsacs, prey remains and the position of the spherical plug were noted. Some plugs were cut open to see how they were constructed. All the burrows (« = 97; Table 1) excavated were roughly Y-shaped with two short arms forming a V above the junction of the main burrow (Figs. 1 & 2). The angle of the burrow to the soil surface var- ied between about 50° and 60° and all burrows had lids (Table 2). The largest burrows were those of adult females, generally about 150 mm deep: one arm between 30 and 40 mm long from the junction to the surface of the soil, ending in a cuff and wa- fer-lid trapdoor, the other ending some 10 mm be- low the soil surface. The trapdoor was soft, folded and asymmetrical. Larger burrows had lids with a raised “cuff” (Leroy & Leroy 2000) of silk around the lid as well as the trapdoor and were found to contain adult female spiders. All the burrows ex- cavated contained hard, spherical plugs or “mar- bles” formed of soil particles bound together by 558 LEROY & LEROY— BURROW PLUGS OF ANCYLOTRYPA SP. 559 Table 1 . — Total number of burrows of Ancylotry- pa sp. excavated over a ten year period from 1992- 2002 including number of burrows containing young or eggsacs. Immature spiders were less than 8 mm in length. No adult males were found in bur- rows. Month Egg Young sacs present present No young or egg sacs present Imma- ture Total Jan. 3 2 — 2 7 Feb. 4 2 — 3 9 Mar. 4 2 — 4 10 April 3 1 — 5 9 May 3 2 1 4 10 June 1 2 — 4 7 July 1 2 2 2 7 Aug. 1 1 1 3 6 Sept. 2 1 1 2 6 Oct. 2 1 1 3 7 Nov. 2 1 2 4 8 Dec. 3 1 — 6 11 Total 29 18 8 42 97 fine, strong silk. The size of the marble correspond- ed closely to the burrow diameter. On cutting open the spherical marbles, all were found to contain only soil particles and no prey remains. Many mar- bles of different diameters were found on the sur- face of the soil and it seems that the spiders peri- odically construct new ones, probably as they grow and enlarge their burrows the spiders discard the old, smaller marbles (Fig. 3). During nocturnal ob- servation, the spiders were found to be sit-and-wait predators. They do not leave their burrows to hunt but lurk below the trapdoor for prey to come close enough to be snatched and taken down into the bur- row. On excavating the burrows, if the spider was undisturbed, the marble would be found at the top of the shorter, blind arm. Likewise, if during ex- cavation, the spider retreated to the bottom of the burrow, the marble would still be at the top of the shorter arm. However, if the burrow wall was care- fully breached for observation and if the trapdoor was then scratched, the spider would pull on the door presumably to test what the disturbance was. I More vigorous scratching, which eventually broke the door, sent the spider scurrying from the open I arm into the blind one, where it retrieved the marble . (Fig. 4) and then positioned the marble below the door, hiding below it. All the burrows excavated housed female or im- mature spiders and those of adult females also con- j tained eggsacs or young at all times of year. No I adult males were collected from burrows. Prey re- mains and exuvia were found to be stored above the marble at the top of the blind arm while eggsacs were generally suspended from the burrow walls near the bottom of the burrow. On checking geomorphological and flooding data it became apparent that the area the spiders inhabit does not become inundated when the river floods but because it is rather flat, will be covered in water from a few to several centimeters deep for varying lengths of time after even a single rain storm. Since the first observations in the southern summer of 1992-1993 we have had the opportunity to observe the effect of showers of varying intensity and du- ration and noted that sheets of water form and, be- cause the soil is virtually impermeable, the humid- ity penetrates it very slowly indeed. In years of steady rainfall these sheets of water persist all sum- mer, being replenished with each successive shower although if it does not rain regularly the shallower parts dry up after a few days. The whole area where the spiders are found is interspersed with vegetated “islands” which are up to half a meter higher than the surrounding bare parts. During the summer months there is good grass cover and considerable termite activity. Ac- cording to Ferrar 1982, 12 species of termites can be found on what he termed “turf vlei” (here called sodic, alluvial soils). Three species are dominant with Macrotermes natalensis being the most visi- ble. It was expected that termites would be the main prey for this spider but a cursory examination of prey remains shows small Coleoptera and ants con- stitute the main prey along with the remains of a few termites and other small unidentifiable hyme- nopterans. The population density of this species of Ancy- lotrypa in the study area is very high, especially on slightly raised and sloping ground. A square meter transect was marked out into 200 mm squares on one of these shallow slopes, the covering of loose soil swept from the top few millimeters and 170 burrow lids counted. While excavating burrows, still more were found which had not been apparent from the surface. At a rough estimate, in optimum areas, there could be around 200 burrows per square meter but these did not extend into areas with dif- ferent soil textures. There are “Y” shaped burrows constructed by other Ancylotrypa species but until this study, there are no records of marbles being constructed by spi- ders in this genus. The only other similar behavior seems to be that of a trapdoor spider in the family Nemesiidae, Stanwellia nebulosa, found in South Australia (Main 1976). This species uses a pebble or stone attached to a sock and stores in a side pocket about halfway down its burrow, counterbal- anced to fall neatly so that when it feels threatened it can be pulled down to close off the bottom half of the bunow. Because the burrows of this species of Ancylo- 560 THE JOURNAL OF ARACHNOLOGY Figure 1 . — Excavated burrow of adult female Acylotrypa showing marble stored at the top of the blind arm. The spider is just above an egg case which is attached to the wall at the bottom of the burrow. Scale bar = 10mm. Figure 2. — Diagram of burrow showing the two positions of the marble: (a) stored at the top of the blind arm and (b) in position to plug and protect the bun'ow. Figure 3. — Size of marbles shown next to a metric scale. Figure 4. — Female spider collecting marble from the blind arm preparatory to plugging the open ai'm of the burrow. Note young still in material buiTOw, Scale bar = 10 mm. LEROY & LEROY— BURROW PLUGS OF ANCYLOTRYPA SP. 561 Table 2. — Numbers and sizes of burrow lids of Ancylotrypa sp. measured in 1 square meter area. Size of burrow lid (diameter in mm) Number <2 42 2-4 83 4-6 32 6-8 13 trypa are found on slightly sloping ground where water can drain away, it appears that the hypothesis that the marbles are used to stop water flooding the burrow is probably not the case. The conclusion is that the marbles are used by the spider to plug the burrow when the trap door is breached and we sug- gest a vernacular name of “marble spiders”. It was not possible to identify the species on which this study is based because the genus Ancy- lotrypa is in need of taxonomic revision. It is ten- tatively identified as Ancylotrypa brevipalpis (Hew- itt 1916) described as Pelmatorycter brevipalpis and originally placed in the family Ctenizidae by Hewitt based on material collected from Pretoria and from one other locality, Crocodile Bridge. Ra- ven (1985) transferred the genus to the family Cyr- taucheniidae and the species to the genus Ancylo- trypa. If it is A. brevipalpis, males have been collected in pit traps and recorded from Gauteng and the North West Provinces of South Africa (Dip- penaar Schoeman 2002) which means that Nylsvley Provincial Nature Reserve (24°39"S:28°42"E) and the nearby Mosdene Private Nature Reserve (24 °3T'S:28 °47"E) in Limpopo Province, South Africa will constitute new locality records. Voucher spec- imens are deposited in The National Collection of Arachnida, ARC-Plant Protection Research Insti- tute, Pretoria, South Africa (PPRI). ACKNOWLEDGMENTS The authors thank The Friends of Nylsvley and the Nyl Floodplain for permission and encourage- ment to list and study the arachnofauna of the Nylsvley Nature Reserve and to Steve Langton for first pointing out the spiders and their burrow plugs. We are especially indebted to Dr. Gail Stratton of the University of Mississippi for final editing of the manuscript and Dr. Eugene Marais, Senior Curator Natural History of The National Museum of Na- mibia for helpful suggestions and initial editing. LITERATURE CITED Barnes, K.N., W.R. Tarboton & J. McLlister. In press. The Important Bird Areas of Southern Af- rica. Avian Demography Unit/Birdlife South Af- rica. Dippenaar-Schoeman, A.S. 2002. Baboon and Trap- door Spiders of Southern Africa: an Identifica- tion Manual. Plant Protection Research Institute Handbook No. 13:1-28. Ferrar, P. 1982. The termites of the Savanna Eco- system Project study area, Nylsvley. South Af- rican National Scientific Programmes Report No. 60. 41 pp. Hewitt, J. 1916. Descriptions of new South African spiders. Annals of the Transvaal Museum 5:180- 213. Leroy, A. & J. Leroy. 2000. Spiderwatch in South- ern Africa. Struik New Holland, Cape Town,l- 96. Main, B.Y. 1976. The Natural History of the Won- gan Hills. Western Australian Naturalists Club, 1977 No. 1. Raven, R.J. 1985. The spider infraorder Mygalo- morphae (Araneae): cladistics and systematics. Bulletin of the American Museum of Natural History 182:1-180. Manuscript received 28 December 2004, revised 15 August 2005. 2005. The Journal of Arachnology 33:562-568 THE SPERMATOZOA OE THE ONE-PALPED SPIDER TIDARREN ARGO (ARANEAE, THERIDHDAE) Peter Michalik\ Barbara Knoflach^, Konrad Thaler^, Gerd Alberti^: ‘Zoologisches Institut und Museum, Ernst-Moritz-Arndt-Universitat, J.-S.-Bach-StraBel 1/12, D- 17489 Greifswald, Germany; ^Institut fiir Zoologie und Limnologie, Leopold-Franzens-Universitat, TechnikerstraBe 25, A-6020 Innsbruck, Austria ABSTRACT. The species of the genus Tidarren are known for their one-palped males and outstanding copulatory behavior. In our ultrastructural observations of T. argo Knoflach & van Harten 2001, we show that this species possesses highly specific spermatozoa which differ from those found in other spiders: The nucleus of the sperm cell is strongly elongated and characterized by a conspicuous implantation fossa. The basis of the axoneme is located close to the acrosomal complex. The axoneme starts in front of the implantation fossa which extends deeply into the postcentriolar elongation. The implantation fossa is filled with dense staining globules and granules as in other theridiid species. Apart from these peculiarities, in T. argo the proximal centriole is located extraordinarily far away from the distal one. The encapsulated cleistospermia are surrounded by a thin secretion sheath. Remarkably, mature spermatozoa are not densely packed, but embedded in a copious secretion. Keywords: Spider sperm ultrastructure, nucleus, implantation fossa, centriole, secretion The theridiid spider Tidarren argo which was first described by Knoflach & van Harten 2001 from Yemen exhibits several peculiari- ties in exomorphology and behavior. The males amputate one palp some hours after the penultimate molt. Such self-amputation is known only from Tidarren and Echinotheri- dion species, but from no other spider (Kno- flach & van Harten 2000, 2001; Knoflach 2002). The reason for palp removal may be to increase locomotor performance as shown for T. sisyphoides (Walckenaer 1842) (see Ramos et al. 2004). In T. argo, the male dies during copulation and even becomes emasculated by the female. Immediately after insertion the fe- male twists off the single male palp, which then continues with sperm transfer discon- nected from the male (for details see Knoflach & van Harten 2001). Based on these outstand- ing features, the present study focuses on the fine structure of the spermatozoa, which are briefly compared with our own unpublished observations on other theridiid spiders. Tidarren argo from Yemen, Khamis Bani Sa’d, 15°11'N 43°25'E, were kept alive in Innsbruck, in plastic boxes at room tempera- ture. From this breeding stock, male speci- mens were dissected and fixed in 3.5% glu- taraldehyde in 0.1 M phosphate buffer, followed by postfixation in buffered 2% os- Figures 1-4. — Late-stage spermatids in Tidarren argo. 1. Overview of part of the testis. Spermatids grouped together in cysts which are sunounded by extensions of the somatic cells (arrow). Scale bar = 10 p.m. 2. Longitudinal section of two spermatids. The ribbon-shaped nucleus coils several times around the axoneme as evident on the right spermatid. Arrow points to nuclear pores. Scale bar = 1 pm. 3. The nucleus is surrounded by a manchette of microtubules. Nuclear canal runs along outer edge of the nucleus. The axoneme possesses a 9 X 2 +3 microtubular pattern. Scale bar = 0.5 pm. 4. Longitudinal section of spermatids. Note the aberrant organization: Axonemal basis (dC) located in front of implantation fossa near acrosomal vacuole; implantation fossa with granular dense material; nucleus strongly elongated, its anterior part triangular. Scale bar = 1 pm. Abbreviations: AV = acrosomal vacuole, Ax = axoneme, dC = distal centriole, FI = flagellum, IF = implantation fossa, M = mitochondria. Mm = manchette of microtubules, N = nucleus, NC = nuclear canal, Sp = spermatozoa, V = vesicles. 562 MICHALIK ET AL.— SPERMATOZOA OF TI DARREN ARGO 563 564 THE JOURNAL OF ARACHNOLOGY i mium tetroxide. After washing, the specimens were rinsed in graded ethanol solutions (60%, 70%, 80%, 96%, absolute) and embedded in Spurrs resin (Spurr 1969). Ultrathin sections were made with a Leica ultramicrotome and stained with uranyl acetate and lead citrate (Reynolds 1963). Examination was performed with a Zeiss EM lOA electron microscope. For depository of voucher specimens see Knofiach & van Harten (2001). Differentiation of the spermatozoa takes place in cysts which are surrounded by exten- sions of the somatic cells. These are located in the periphery of the relatively small testes. The spermatids are loosely distributed within these cysts (Fig. 1). In the following account, the shape of the main cell components of late- stage spermatids will be described, because this stage is most useful for comparative sper- matological studies. A reconstruction of a late-stage spermatid is given in Fig. 5. Nucleus. — In late-stage spermatids the nu- cleus is the most prominent component. It is strongly elongated and turns several times around the axoneme (Fig. 2). In cross-sections the main part of the nucleus forms a flattened ribbon (Figs. 2, 3). Only the anterior part is more or less lens-shaped, containing the axo- nemal basis and the implantation fossa (Fig. 4). Until the coiling process, the nucleus is surrounded by a manchette of microtubules (Fig. 3). In longitudinal sections the anterior part of the nucleus forms a triangle with the main part (postcentriolar elongation, see be- low) (Fig. 4). Along the outer edge of the nu- cleus runs a nuclear canal, which contains the acrosomal filament in its anterior part (Figs. 2, 3, 9). Implantation fossa and axoneme. — Dur- ing spermatogenesis normally an indentation of the nucleus is formed in front of the axo- nemal basis, the so-called implantation fossa. In T. argo the axonemal basis migrates to the anterior part of the nucleus and is finally lo- cated close to the acrosomal vacuole (Fig. 4). The implantation fossa extends behind the ax- onemal basis deeply into the postcentriolar elongation of the nucleus, which in T. argo constitutes the main part of the nucleus (Figs. 4, 5). Within the implantation fossa several globules and granules are accumulated which are very dense in mature spermatozoa (Figs. 6, 7). Embedded in this material, the proximal centriole is located extraordinarily far away Figure 5. — Schematic reconstruction of a late- stage spermatid of Tidarren argo (only main cell components shown). Note the elongated nucleus coiling several times around the axoneme. As a consequence of the extremely positioned axonemal basis close to the acrosomal vacuole the main part of the nucleus (behind the axonemal basis) can be determined as postcentriolar elongation of the nu- cleus (peN). Abbreviations: AV = acrosomal vac- uole, Ax = axoneme, N (peN) = nucleus (postcen- triolar elongation of the nucleus), NC = nuclear canal. from the distal one (Figs. 6, 7). A reconstruc- tion of a longitudinal section of the anterior part of a late-stage spermatid is given in Fig. 12. The axoneme possesses the 9X2 + 3 pattern of microtubules (Fig. 3). Acrosomal complex. — The acrosomal vac- uole has an irregular arrowhead- shape (Fig. 8). The basis of the acrosomal vacuole is very MICHALIK ET AL.— SPERMATOZOA OF TIDARREN ARGO 565 Figures 6-1 L — Late-stage spermatids and mature spermatozoa of Tidarren argo. 6, 7. Sections of spermatids in the region of the implantation fossa. Note the proximal centriole located deeply within the implantation fossa which is filled with globular and granular dense material. Scale bars = 1 fxm. 8. Longitudinal section of the irregularly shaped acrosomal vacuole. Acrosomal filament starts at anterior part of acrosomal vacuole and continues into nuclear canal. Scale bar = 0.25 jxm. 9. Front part of sper- matid. Note irregular shape of acrosomal vacuole. Manchette of microtubules around nucleus continues to acrosomal vacuole. The acrosomal filament seems very short, because of the empty nuclear canal close to the acrosomal vacuole. Scale bar = 0.5 p.m. 10. At the end of spermatogenesis the main cell components (nucleus, axoneme and acrosomal vacuole) coil within the cell. Scale bar = 1 jxm. 1 1 . Section of sper- mophore of palpal organ. Mature spermatozoa possess a thin secretion sheath and are embedded in a dense conspicuous secretion. Scale bar = 1 fxm. Abbreviations: AF = acrosomal filament, AV = acrosomal vacuole. Ax = axoneme, pC = proximal centriole, IF = implantation fossa. Mm = manchette of micro- tubules, N = nucleus, NC = nuclear canal, Sec = secretion, Sp = spermatozoa. 566 THE JOURNAL OF ARACHNOLOGY thin. A layer of dense material is located be- tween the acrosomal vacuole and the nucleus (Fig. 9). The manchette of microtubules sur- rounds more than half of the acrosomal vac- uole (Fig. 9). The subacrosomal space is nar- row and contains the acrosomal filament which continues into the nuclear canal (Fig. 8, 9). The short acrosomal filament ends near the axonemal basis (Fig. 9). Additional cell components. — Other cell components, e.g., mitochondria, Golgi appa- ratus, and vesicles are mainly seen in the cy- toplasm of the posterior part of the spermatid. They seem to be absent in mature spermato- zoa. Mature spermatozoa. — At the end of sper- matogenesis the spermatids coil. The main cell components (acrosomal vacuole, nucleus and axoneme) are involved in this coiling pro- cess within the sperm cell. The nucleus coils up to four times and the axoneme coils at the periphery of the cell (Fig. 10). Finally, in ma- ture spermatozoa, which receive a secretion sheath, cell components are compact and tightly together (Fig. 11). The secretion sheath is rather thin and the mature spermatozoa are embedded in a conspicuous, dense secretion which apparently hinders the fixation process during preparation as seen in Fig. 11. A re- construction of a section of a mature sper- matozoon is given in Fig. 13. Interestingly, the mature spermatozoa are not densely packed in the spermophore of the palpal organ (Fig. 11). The spermatozoa of Tidarren argo possess a highly derivative organization with most ab- errant features in comparison to other spider species (e.g., Osaki 1969, 1972; Reger 1970; Boissin 1973; Alberti & Weinmann 1985; Al- berti et al. 1986; Alberti 1990; Alberti & Coy- le 1991; Michalik et al. 2003). Unfortunately, no other investigations dealing with fine struc- ture of theridiid spermatozoa exist to allow an evaluation of our results. Hence, we compare the results mainly with our personal obser- vations on other theridiid spiders (PM pers. obs.). On this evidence it appears that the spermatozoa of T. argo must be regarded as highly specialized, both within Theridiidae and within spiders in general. The nucleus of the T. argo spermatozoa is strongly elongated and ribbon-shaped over most of its length. As a consequence of the unusual position of the axonemal basis close to the acrosomal vacu- ole, the main part of the nucleus is represented Figure 12. — Schematic reconstruction of a lon- gitudinal section of the front part of a late-stage spermatid in Tidarren argo. Note the proximal cen- triole which is embedded in the globular and gran- ular dense material within the implantation fossa. The acrosomal filament extends only into the most anterior part of the nuclear canal. Abbreviations; AF = acrosomal filament, AV = acrosomal vacu- ole, Ax = axoneme, IF = implantation fossa, dC = distal centriole, pC = proximal centriole, N (peN) = nucleus (postcentriolar elongation of nucleus), NC = nuclear canal. by the so-called postcentriolar elongation of the nucleus. In contrast, in our observations on Argyrodes argyrodes (Walckenaer 1842), Crustulina guttata (Wider 1834), Nesticodes rufipes (Lucas 1846), Steatoda grossa (C.L. Koch 1838) and Theridion nigrovariegatum Simon 1873, the nucleus is completely differ- ent, oval in cross section and more compact. The most striking features are the position of the axonemal basis and the extension and lo- cation of the implantation fossa. In T. argo the axonemal basis is located close behind the ac- rosomal vacuole in front of the implantation fossa. This arrangement completely differs from the above mentioned species. In the lat- ter, the axonemal basis is located behind the implantation fossa which does not extend to the acrosomal vacuole, a situation typical for many other spider species (e.g., Alberti 1990; Alberti & Coyle 1991; Michalik et al. in press). The only exceptions in this respect known until now occur in the genera Tetrag- natha and Cyclosa in which the implantation fossa also reaches the most anterior part of the MICHALIK ET AL.— SPERMATOZOA OF TIDARREN ARGO 567 Figure 13. — Schematic reconstruction of a sec- tion of a mature spermatozoon in Tidarren argo. The nucleus coils as a spiral within the cell. The spermatozoon is surrounded by a thin secretion sheath and a considerable amount of dense secre- tion. Abbreviations: AV = acrosomal vacuole, Ax = axoneme, dC distal centriole, N (peN) = nucleus (postcentriolar elongation of nucleus), Sec = secre- tion, SSh = secretion sheath. 1 i ! I nucleus. However, in contrast to T. argo, in these species the axonemal basis is located in the posterior part of the implantation fossa as usual (Alberti 1990; Michalik et al. in press). Interestingly, in the theridiid spider Neottiura. bimaculata (Linnaeus 1767) the position of the axonemal basis is similar to that seen in T. argo, but the nucleus and acrosomal vac- uole are more compact, their shape therefore resembling that of other theridiid spiders. In T. argo the acrosomal vacuole shows an ir- regular arrowhead- shape and differs from the cylindrical or tube-like acrosomal vacuoles found in other theridiid species, e.g., Argyro- des argyrodes and Theridion nigrovariega- turn. Of special interest is the dense secretion in which mature spermatozoa are embedded. Remarkably, the spermatozoa are loosely ar- ranged in the palpal organ in comparison to other spider species, e.g., Pachygnatha listeri Sundevall 1830 (Michalik et al. in press). In this species no secretion was found and the spermatozoa have a thick protective secretion sheath. We suggest that in T. argo the protec- tive function of the thick secretion sheath might be replaced by the copious secretion. Interestingly, the secretions in which the sper- matozoa are embedded clearly differ between different species. In each of the theridiid spi- ders observed above we found a different structural aspect of the secretion. Since other spider families show different types of secre- tions (unpublished observations by the au- thors), a great diversity in this feature is re- vealed. This may reflect specific importance in the process of reproduction. As nothing is known about the function of male secretions and their possible role in the female genital system, this is an interesting topic for future research. Tidarron argo possesses highly derivative and aberrant spermatozoa in contrast to other theridiid species, but more investigations on further theridiid species are needed to develop evolutionary scenarios and to clarify a possi- ble phylogenetic and functional relevance of spermatological characters. Furthermore, it would be important to know more about the function and chemistry of the secretion in which the mature spermatozoa are embedded. ACKNOWLEDGMENTS We are very grateful for the assistance of Wencke Reiher. This study was financially supported by the Studienstiftung des Deutsch- en Volkes and the German Research Council (DEG; Al 138/11-1) and the Austrian Acade- my of Sciences (APART 10748, Austrian pro- gram for advanced research and technology). LITERATURE CITED Alberti, G. 1990. Comparative spermatology of Ar- aneae. Acta Zoologica Fennica 190:17-34. Alberti, G. & C. Weinmann. 1985. Fine structure of spermatozoa of some labidognath spiders (Filis- tatidae, Segestriidae, Dysderidae, Oonopidae, Scytodidae, Pholcidae; Araneae; Arachnida) with remarks on spermiogenesis. Journal of Morphol- ogy 185:1-35. Alberti, G., B.A. Afzelius & S.M. Lucas. 1986. Ul- trastructure of spermatozoa and spermatogenesis in bird spiders (Theraphosidae, Mygalomorphae, Araneae). Journal of Submicroscopic Cytology 18:739-753. Alberti, G. & EA. Coyle. 1991. Ultrastructure of the primary male genital system, spermatozoa, and spermiogenesis of Hypochilus pococki (Ar- aneae, Hypochilidae). Journal of Arachnology 19:136-149. Boissin, L. 1973. Etude ultrastructurale de la sper- miogenese de Meta bourneti Simon (Arachnides, Araneides, Metinae). Comptes Rendus deuxieme 568 THE JOURNAL OF ARACHNOLOGY de la Reunion Arachnologique d’ Expression Frangaise, Montpellier, 7-22. Knoflach, B. 2002. Copulation and emasculation in Echinotheridion gibberosum (Kulczynski, 1899) (Araneae, Theridiidae). Proceedings of the 19th European Colloquium of Arachnology (Aarhus 2000): 139-144. Knoflach, B. & A. van Harten. 2000. Palpal loss, single palp copulation and obligatory mate con- sumption in Tidarren cuneolatum (Tullgren, 1910) (Araneae, Theridiidae). Journal of Natural History 34:1639-1659. Knoflach, B. & A. van Harten. 2001. Tidarren argo sp. nov. (Araneae: Theridiidae) and its excep- tional copulatory behaviour: emasculation, male palpal organ as mating plug and sexual canni- balism. Journal of Zoology 254:449-459. Michalik, R, M.R. Gray & G. Alberti. 2003. Ultra- structural observations of spermatozoa and sper- miogenesis in Wandella orana Gray, 1994 (Ar- aneae: Filistatidae) with notes on their phylogenetic implications. Tissue & Cell 35: 325-337. Michalik, P., P. Sacher & G. Alberti. In press. Ul- trastructural observations of spermatozoa of some tetragnathid spiders and their phylogenetic implications (Araneae, Tetragnathidae). Journal of Morphology. Osaki, H. 1969. Electron microscope study on the spermatozoon of the liphistiid spider Heptathela kimurai. Acta Arachnologica 22:1—12. Osaki, H. 1972. Electron microscope study on sper- miogenesis in the spider, Oxyopes sertatus. Jap- anese Journal of Zoology 16:184-199. Ramos, M., D.J. Irschick & T.E. Christenson. 2004. Overcoming an evolutionary conflict: Removal of a reproductive organ greatly increases loco- motor performance. Proceedings of the National Academy of Science 101:4883-4887. Reger, J.F. 1970. Spermiogenesis in the spider PT saurina sp.: a fine structure study. Journal of Morphology 130:421-434. Reynolds, E.S. 1963. The use of lead citrate at high pH as an electron-opaque stain in electron mi- croscopy. Journal of Cell Biology 17:208. Spurr, A.R. 1969. A low-viscosity epoxy resin em- bedding medium for electron microscopy. Jour- nal of Ultrastructure and Molecular Structure Re- search 26:31-43. Manuscript received 1 September 2004, revised 13 August 2005, 2005. The Journal of Arachnology 33:569-572 ON THE OCCURRENCE OF THE 9 + 0 AXONEMAL PATTERN IN THE SPERMATOZOA OF SHEETWEB SPIDERS (ARANEAE, LINYPHIIDAE) Peter Michalik and Gerd Alberti: Zoologisches Institut und Museum, Emst-Moritz- Arndt-Universitat, J.-S.-Bach-StraBe 11/12, D“17489 Greifswald, Germany. E-mail: niichalik@uni~greifswald.de ABSTRACT. In general, flagella and cilia of eukaryotes show an axoneme composed of a 9 + 2 microtubular pattern. However, the axoneme of spider spermatozoa is characterized by an exceptional 9 + 3 microtubular pattern, which is knov/n as a synapomorphy of the Megoperculata (Amblypygi, Uropygi and Araneae). In contrast to all other observed spiders, the axoneme of the linyphiid spider Linyphia triangularis, was shown to lack the central microtubules thus representing a 9 + 0 axoneme. In the present study, we investigated the spermatozoa from several linyphiid species of different genera in order to show whether this peculiar pattern also occurs in other linyphiid spiders. Interestingly, in all observed species (Neriene clathrata, N. peltata, Linyphia hortemis, Lepthyphantes sp., Oedothorax gibbosus, Gongylidium rufipes and Drapetisca socialis) we found the 9 + 0 microtubular pattern in the axoneme. Since this study, although considering still a very limited number of species, includes species from Linyphiinae (Linyphiini and Micronetini) and Erigoninae it seems likely that this pattern is an autapomorphy of Linyphiidae. Keywords^ Sperm, phylogeny, axoneme, microtubules The typical and plesiomorphic axoneme found in eukaryote flagella and cilia possess a 9 + 2 arrangement of the microtubules. Nevertheless, there is a wide range of modi- fications within the axoneme of sperm flagel- la, e.g., in insects (Jamieson et al. 1999). Also within the Chelicerata a broader range of pat- terns occurs as shown by the recent species of the early derivative group, Xiphosura, where two different patterns are reported (9 + 2 and 9 + 0; Fahreebach 1973; Yamamichi & Sek- iguchi 1982; Alberti & Janssen 1986). Within arachnids only the spermatozoa of Scorpiones, Uropygi, Amblypygi, Araneae, Pseudoscorpi- ones and Ricinulei possess a flagellum (sum- mary in Alberti 2000), in contrast to the sper- matozoa of Solifugae, Acari, Palpigradi and Opilioees, which are aflagellate (the only ex- ception is the opilionid genus Siro which shows an axoneme during the spermatogene- sis; Juberthie et aL 1976; Alberti in press). The typical 9 + 2 pattern occurs only in Scor- piones, Pseudoscorpiones and Ricinulei. How- ever, in Scorpiones aberrant patterns, e.g., 9 + 0 and 9+1 have also been reported (Hood et al. 1972; Jespersen & Hartwick 1973; Al- berti 1983). The Uropygi, Amblypygi and Ar- aneae (Megoperculata) possess as a syeapo- morphy a 9 + 3 pattern (summary in Alberti 2000; Michalik et al. 2003, 2004, in press and further personal observations). Thus it seems remarkable, that the linyphiid spider Linyphia triangularis (Clerck, 1757) has an unusual 9 + 0 axonemal pattern (Alberti 1990); unfor- tunately until now there have been no other ultrastructural observations on Linyphiidae spermatozoa to know assess if this pattern is typical of this taxon. In the present study, we investigated the spermatozoa of several different linyphiid spi- ders from the subfamilies Linyphiinae (Liey- phiini and Micronetini) and Erigoninae to be- gin a determination of the generality of this peculiar axonemal pattern within the Lieyphi- idae. Male specimens of Neriene clathrata (Sun- devall 1830), N. peltata (Wider 1834), Liny- phia hortemis Suedevall 1830 (Linyphiinae, Linyphiini); Lepthyphantes sp. (Linyphiinae, Micronetini); Oedothorax gibbosus (Black- wall 1841) and Gongyiidium rufipes (Linnaeus 1758) (Erigoninae); and Drapetisca socialis (Sundevall 1833) were dissected and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buff- 569 570 THE JOURNAL OF ARACHNOLOGY Figures 1-7. — Spermatozoa of the observed linyphiid species. 1. Early spermatids of Oedothorax gibbosus with axonemes in cross — and longitudinal sections. Scale bar = 1 pm. Inset: Detail of the axoneme of Linyphia hortensis in cross-section showing the 9 + 0 pattern. Scale bar = 0.1 pm. 2 — 3. Posterior part of an early spermatid in longitudinal section. 2. Neriene clathrata. Scale bar = 0.25 pm. 3. Lepthyphantes sp. Scale bar = 0.5 pm. 4 — 6. Late spermatids. 4. Drapetisca socialis. Scale bar = 1 pm. 5. Gongylidium rufipes. Scale bar = 0.5 pm. 6. Neriene peltata. Scale bar = 0.5 pm. 7. Coiled spermatid of Linyphia hortensis. Scale bar = 0.5 pm. Abbreviations: Ax = axoneme, C = centriole, IF = implantation fossa, N = nucleus, NC = nuclear canal (canal containing the acrosomal filament). er followed by postfixation in buffered 2% os- mium tetroxide. After rinsing, the specimens were dehydrated in graded ethanols and em- bedded in Araldite or SpuiT’s resin (Spun* 1969). Ultrathin sections were made on a Lei- ca ultramicrotome and the sections were stained with uranyl acetate and lead citrate (Reynolds 1963). The examination was per- formed with a Zeiss EM lOA electron micro- scope. Voucher specimens have been depos- ited in the Zoological Museum of the University of Greifswald. f Spermiogenesis starts with spermatids which are mainly characterized by a large, roundish nucleus lying in a homogenous cy- i toplasm (Fig. 1). The axoneme migrates into the posterior pole of the nucleus (Figs. 1-4). The centrioles are orientated in the tandem po- MICHALIK & ALBERTI— SPERMATOZOA OF LINYPHIIDAE 571 sition (Fig. 2). In all sections, the absence of the central tubules in the axoneme is obvious in all investigated species (Figs. 1-7) and in cross sections, the 9 + 0 axonemal pattern is evident (Figs. 1 inset, 4-7). Parallel to the mi- gration of the axoneme, a deep posterior in- dentation into the nucleus is formed, the so- called implantation fossa (Fig. 1) which is filled with dense material at the end of sper- matogenesis (Figs. 4, 7). At the end of sper- matogenesis the nucleus coils once and the ax- oneme turns around the nucleus in the periphery of the cell (Fig. 7). The occurrence of the 9 + 0 axonemal pat- tern is unique among the spermatozoa of Me- goperculata and was first shown by Alberti (1990) for Linyphia triangularis. In spiders a 9 + 3 pattern normally occurs and was studied in detail by Dallai et al. (1995). The present study shows the peculiar 9 + 0 pattern for all observed Linyphiidae. Based on these obser- vations many questions arise concerning the function and the phylogenetic impact of this character. Unfortunately, no studies on the movement of linyphiid spermatozoa exist. However, it was shown from other animal spe- cies that an axoneme which lacks the central tubules can still move. For example, Ishijima et al. (1988) compared the beat pattern from Asian and American horseshoe crabs. The Asian species Tachypleus gigas (Muller 1785) possess a 9 + 0 axoneme which beats in he- lical waves, in contrast to the planar waves of the 9 + 2 axoneme of the American horseshoe crab Limulus polyphemus (Linnaeus 1758). Similar results were also reported in the de- tailed studies of Gibbons et al. (1983, 1985) on the eel Anguilla anguilla (Linnaeus 1758) which possess spermatozoa with a 9 + 0 ax- oneme which lacks the central tubules as well as other structures, e.g., outer dynein arms, ra- dial spokes and spokeheads. The spermatozoa of the eel beats in helicoidal waves. Hence it can be assumed that the spermatozoa of Lin- yphiidae are motile and the movements of the axoneme are different from those of other spi- der spermatozoa that possess a 9 + 3 axo- neme. More investigations are needed to test this hypothesis and clarify the possible influ- ence (selective advantage?) within the female genital system. Furthermore, the occurrence of the 9 + 0 axonemal pattern in all observed species of linyphiids supports the assumption of Alberti (1990) that this peculiar pattern might be an autapomorphy of Linyphiidae. Therefore it would be of much interest to know the situ- ation in the supposed sister taxon Pimoidae as well as in the other linyphiid subfamilies (e.g., Hormiga 1994a, b; 2000). ACKNOWLEDGMENTS We are very grateful for the assistance of Wencke Reiher. Furthermore, we thank Ga- briele Uhl (Bonn) for the provision of the males of O. gibbosus. For the exact identifi- cation of part of the material we are indebted to Konrad Thaler (Innsbruck). This study was financially supported by the German National Merit Foundation (grant to PM) and the Ger- man Research Foundation (DFG, Al 138/11- 1). LITERATURE CITED Alberti, G. 1983. Fine structure of scorpion sper- matozoa {Buthus occitanus; Buthidae, Scorpi- ones). Journal of Morphology 177:205-212. Alberti, G. 1990. Comparative spermatology of Ar- aneae. Acta Zoologica Fennica 190:17-34. Alberti, G. 2000. Chelicerata. Pp. 311-388. In Pro- gress in Male Gamete Ultrastructure and Phylog- eny. Vol. 8. (B. G. M. Jamieson, ed.). Oxford & IBH Publishing Co. PVT. LTD, Queensland. Alberti, G. In press. Double spermatogenesis in Chelicerata. Journal of Morphology. Alberti, G. & H.H. Janssen. 1986. On the fine struc- ture of spermatozoa of Tachypleus gigas (Xip- hosura, Merostomata). International Journal of Invertebrate Reproduction and Development 9: 309-319. Dallai, R., B.A. Afzelius & W. Witalinski. 1995. The axoneme of the spider spermatozoon. Bol- lettino di Zoologia 62:335-338. Fahrenbach, W.H. 1973. Spermiogenesis in the horseshoe crab Limulus polyphemus. Journal of Morphology 140:31-52. Gibbons, B.H., B. Baccetti & I.R. Gibbons. 1985. Live and reactivated motility in the 9+0 flagel- lum of Anguilla sperm. Cell Motility 5:333-350. Gibbons, B. H., I.R. Gibbons & B. Baccetti. 1983. Structure and motility of the 9+0 flagellum of eel spermatozoa. Journal of Submicroscopic Cy- tology 15:15-20. Hood, R.D., O.E Watson, T.R. Deason & C.L.B. Benton, Jr. 1972. Ultrastructure of scorpion sper- matozoa with atypical axonemes. Cytobios 5: 167-177. Hormiga, G. 1994a. A revision and cladistical anal- ysis of the spider family Pimoidae (Araneoidea: Araneae). Smithsonian Contributions to Zoology 549:1-104. Hormiga, G. 1994b. Cladistics and the comparative 572 THE JOURNAL OF ARACHNOLOGY morphology of linyphiid spiders and their rela- tives (Araneae, Araneoidea, Linyphiidae). Zoo- logical Journal of the Linnean Society 1 1 1:1-71. Hormiga, G. 2000. Higher level phylogenetics of Erigonine spiders (Araneae, Linyphiidae, Erigon- inae). Smithsonian Contributions to Zoology 609:1-160. Ishijima, S., K. Sekiguchi & Y, Hiramoto. 1988. Comparative study of the beat patterns of Amer- ican and Asian horseshoe crab sperm: evidence for a role of the central pair complex in forming planar waveforms in flagella. Cell Motility and the Cytoskeleton 9:264-270. Jamieson, B.G.M., R. Dallai & B.A. Afzelius. 1999. Insects: Their Spermatozoa and Phylogeny. Sci- ence Publishers, Inc., Enfield. 555 pp. Juberthie, C., J. E Manier & L. Boissin. 1976. Etude ultrastructurale de la double-spermioge- nese chez lopilion cyphophthalme Siro rubens Latreille. Journal de Microscopie et de Biologie Cellulaire 25:137-148. Michalik, R, M.R. Gray & G. Alberti. 2003. Ultra- structural observations of spermatozoa and sper- miogenesis in Wandellci orana Gray, 1994 (Ar- aneae: Filistatidae) with notes on their phy- logenetic implications. Tissue & Cell 35:325- 337. Michalik, R, J. Haupt & G. Alberti. 2004. On the occurrence of coenospermia in mesothelid spi- ders (Araneae: Heptathelidae). Arthropod Struc- ture & Development 33:173-181. Michalik, R, P. Sacher & G. Alberti. In press. Ul- trastructural observations of spermatozoa of some tetragnathid spiders and their phylogenetic implications (Araneae, Tetragnathidae). Journal of Morphology. Reynolds, E.S. 1963. The use of lead citrate at high pH as an electron-opaque stain in electron mi- croscopy. Journal of Cell Biology 17:208. Spurr, A.R. 1969. A low-viscosity epoxy resin em- bedding medium for electron microscopy. Jour- nal of Ultrastructure and Molecular Structure Re- search 26:31-43. Yamamichi, Y. & K. Sekiguchi. 1982. Axoneme patterns of spermatozoa of Asian horseshoe crabs. Experientia 38:1219-1220. Manuscript received 27 September 2004, revised 13 August 2005. 2005. The Journal of Arachnology 33:573-581 EVIDENCE FOR DIRECTIONAL SELECTION ON MALE ABDOMEN SIZE IN MECOLAESTHUS LONGISSIMUS SIMON (ARANEAE, PHOLCIDAE) Bernhard A. Huber: Zoological Research Institute and Museum Alexander Koenig, Adenauerallee 160, 53113 Bonn, Germany, E-mail: b.huber.zfmk@uni-bonn.de ABSTRACT, Abdomens of male Mecolaesthus longissimus Simon 1 893 are on average more than twice as long as in females, their length is highly variable, and they show extremely steep allometric values when scaled on body size (OLS, b = 2.64). Males cohabit with females, and they likely fight to defend this position as other pholcid spiders do. Male legs, which are usually used in pholcid male-male fights, do not show the usual high allometric values but a very low value (OLS, b = 0.37). Collectively, this lends support to the idea that M. longissimus males do not use their legs in fights and that male abdomens have assumed a role in male-male fights. However, behavioral data are missing and sexual selection by female choice or inter-male display might be involved. A large sample of data from taxonomic revisions is used to document that across pholcids, males consistently have longer tibiae 1 (and probably legs in general) than females. Several possible reasons have been suggested to account for longer male than female legs in various spider groups, but the pattern in pholcids remains to be explained. Keywords: Sexual size dimorphism, sexual selection, allometry, Pholcidae Extreme sexual size dimorphism in spiders has attracted considerable attention for a long time and its evolutionary origin has fueled a lively and ongoing debate (Vollrath & Parker 1992; Coddington et al. 1997; Head 1995; Prenter et al. 1997, 1998, 1999; Hormiga et al. 2000; Schneider et al. 2000; Moya-Larano et al. 2002; Walker & Rypstra 2003). The more common case of slight size dimorphism and the rather exceptional case of males being larger than females have remained compara- tively out of the main focus of size dimor- phism studies in spiders (but see Prenter et al. 1995, 2003; Toft 1989; Schiitz & Taborsky 2003). Different selective forces, both natural and sexual, probably interact in many species, but fecundity selection may be the single ma- jor factor responsible for females usually be- ing larger than males (Beck & Connor 1992; Elgar 1992; Head 1995; Prenter et al. 1999). However, simple size measures derived from the taxonomic literature may result in an over- ly simplistic view of dimorphism. Depending on the structure measured, either males or fe- males may appear to be the ‘larger’ sex, and some or most dimorphism may be in shape rather than in size (Prenter et al. 1995). Few cases of males being larger than fe- males are known in spiders (Prenter et al. 1999; Lang 2001) even though large male size advantage has been documented in numerous species (Vollrath 1980; Elgar & Nash 1988; Nielsen & Toft 1990; Dodson & Beck 1993; Kotiaho et al. 1997, 1999; Elgar 1998; Elgar & Fahey 1996; Taylor et al. 2001; Prenter et al. 2003; Schaefer & Uhl 2003). In most cases in which males are larger than females, male- male fights are intense, and winners of con- tests sire a significant proportion of their mate’s offspring (Rovner 1968; Watson 1990; Elgar 1998). Selection is particularly strong on the fighting structures per se (e.g., chelic- erae in certain linyphiid and salticid spiders: Rovner 1968; Toft 1989; Pollard 1994; Funke & Huber In press) and such intense directional selection usually results in high allometric values (i.e. > 1.0; Petrie 1992; Green 1992; Baker & Wilkinson 2001; Tatsuta et al. 2001; Funke & Huber In press; see also Eberhard et al. 1998; Eberhard 2002a, b). Natural selec- tion may also result in males being the larger sex, as in the exceptional case of the water spider, Argyroneta aquatica (Clerck 1757). In this species, males are on average nearly 30% larger than females, as a result of the unusual habitat (Schtitz & Taborsky 2003). The presence and degree of sexual size di- morphism within and among species can be 573 574 THE JOURNAL OF ARACHNOLOGY Figures 1, 2. — Mecolaesthus longissimus, geni- talic characters measured. 1. Bulb length (b) and procursus length (p), dorsal view; 2. Epigynum width (e), ventral view. used to generate behavioral hypotheses that can then be tested. In this study, I have two main objectives: (1) to document and quantify the apparently unique dimorphism observed in the pholcid Mecolaesthus longissimus, and (2) to use data from the literature to quantify leg length dimorphism across pholcid species. The main object of this study, Mecolaesthus longissimus Simon 1893, is endemic to the Cordillera de la Costa in northern Venezuela (Huber 2000). Nothing is known about its bi- ology except for some very basic habitat data (Simon 1893; Huber 2000). METHODS Males and females of Mecolaesthus longis- simus were collected in a forest above Colonia Tovar (10°25'N, 67°18'W), 2100 m a.s.L, Ar- agua, Venezuela, on 26 November 2002, by the author. The present analysis is based on a sample of 30 males and 14 females preserved in 80% ethanol. They are presently deposited at the Zoological Research Institute and Mu- seum Alexander Koenig, Bonn, but will later be partly transferred to the Museo de La Salle, Caracas. Drawings were made with a camera lucida on a Leitz Dialux 20 compound micro- scope. Photos were made with a Nikon Cool- pix 995 digital camera (1600 X 1200 pixels) mounted on a Nikon SMZ1500 dissecting mi- croscope. Measurements were made with an ocular grid on a Nikon SMZ1500 dissecting micro- scope. Tibia length was measured dorsally; carapace length was measured medially from anterior median eyes to posterior border; ab- domen length was measured ventrally from frontal end to base of frontal spinnerets; an- terior and posterior parts were divided by the epigastric furrow, resulting in two measures; for genitalic measures see Figs, 1 and 2. Gen- italia were included in the analysis to support the assumption that all specimens included are indeed the same species. Statistical analysis was done with SPSS 11.0. Ordinary least squares (OLS) and reduced major axis (RMA) regressions of log-transformed characters were calculated for all traits on carapace length as an indicator of body size (for cri- tique and justification of method see Green 1999 and Eberhard et al. 1999). Carapace length was used rather than carapace width i (the usual indicator of body size in spiders) because lateral carapace borders appeared too soft and indistinct. For comparison of male and female tibia 1 lengths in pholcid spiders, data were taken from recent revisions (Huber 1997a, b, c, 1998a, b, 2000, 2001, 2003a, b, c; Huber & Perez 1998, 2001; Huber et al. In press; B.A. Huber unpubl. data). In order to be included in the analysis, the species (re)description had to give a mean value of at least five measured tibiae 1 in each sex. All together, 2673 tibia 1 measures of 100 species (20 of them un- published) were included, representing 28 genera and all four pholcid subfamily-level taxa. The complete data matrix is available from the author. RESULTS Morphometric analysis of Mecolaesthus longissimus. — Three details are noteworthy in the morphometric analysis (Table 1). First, male abdomens are on average more than twice as long as female abdomens (see also Figs. 4, 8, 9). Second, male abdomens are ex- tremely variable (see also Figs. 5-7 & 9). Third, it is the anterior part of the male ab- domen that accounts for most of the variation in male abdomen length. In females, to the HUBER— DIRECTIONAL SELECTION ON MALE ABDOMEN SIZE 575 Figures 3-8. — Mecolaesthus longissimus. 3. Male (left) and female in the web (photo courtesy B. Striffler); 4. Large male, dorsal view; 5-7. Large, medium, and small male abdomens, ventral views; 8. Medium size female. Figs. 4-8 are to the same scale. contrary, it is the posterior part of the abdo- men that is much more variable than the an- terior part. No appreciable shape variation was seen in the structures usually used in species discrim- ination in pholcids (male procursus, bulbal sclerites, cheliceral armature). The regression coefficients of the three genitalic structures measured were low as is usual for genitalia (Eberhard et al. 1998). Surprisingly low re- gression values were also found for male (but not female) legs. Comparative analysis of pholcid tibiae 1. — There is a consistent trend for males to have longer tibiae 1 than females when 100 species were compared (Fig. 10). The mean ratio of male/ female tibia 1 is 1.28, the me- dian 1.27 (Fig. 11). Strictly speaking, species are linked by phylogeny and not independent data points (Harvey & Pagel 1991). However, my aim here is to document a universal trend within the family and not to claim that there are independent events that might justify some adaptive explanation. Regardless of the details of the phylogeny of pholcids, parsimony clearly suggests that ancestral pholcids had longer male than female legs. DISCUSSION The extremely high allometric value of male abdomen length in M. longissimus in- dicates that directional selection is operating on this body part. Structures used as weapons in male-male fights or as visual display char- acters in the context of sexual selection tend to show high allometric values (Petrie 1992; Green 1992; Baker & Wilkinson 2001; Tatsuta et al. 2001; Funke & Huber In press; see also Eberhard et al. 1998; Eberhard 2002a, b). The exact nature of this selection cannot be de- rived from allometric values alone but only by behavioral observations and experiments. However, circumstantial evidence suggests that males might use their abdomens in a most unusual and unexpected way: as display or even fighting devices. First, male-female postinsemination non- contact guarding (sensu Alcock 1994) is rare in spiders (Elgar 1998) but common in phol- cids (Eberhard & Briceno 1985; Raster & Ja- kob 1997; pers. obs.). For example, during a monthly survey of a population of Modisimus guatuso Huber 1998 in Costa Rica from No- vember 1995-September 1997, I counted 398 pairs involving adult males and adult females, not a single pair involving a juvenile female, and 65% of 596 males seen were cohabiting (unpub. data). During several collecting ex- peditions I have become used to the expecta- tion that seeing one adult pholcid often means that another one of the opposite sex is nearby. Most webs at the collection site of the present species contained a male and a mature female. 576 THE JOURNAL OF ARACHNOLOGY Table 1 . — Mecolaesthus longissimus, male and female characters measured (in mm), with sample sizes (n), ranges, means, standard deviations (SD), coefficients of variation, corrected for sample size (CV*), significance values of Kolmogorov-Smirnov tests for normal distribution (KS), estimates on measurement error (± 1/2 unit on the measuring grid), and slopes (b) of regressions on carapace length as an indicator of body size, using ordinary least squares (OLS) and reduced major axis (RMA) regression. Slopes sig- nificantly different from 0 are indicated by *(P < 0.05), **(P < 0.01), and ***(P < 0.001). RMA regressions were not calculated when OLS values were non-significant. Characters n Range Mean SD CV* KS Measure- ment error (± mm) b (OLS) b (RMA) Males tibia 1 length 30 10.53-12.80 11.59 0.54 4.7 0.57 0.07 0.37*** 0.57 tibia 3 length 30 5.15-6.40 5.78 0.31 5.3 0.69 0.05 Q 49*** 0.66 abdomen total length 30 2.90-6.50 4.85 1.21 25.1 0.33 0.07 2.64*** 3.14 abdomen frontal part 30 1.15-3.80 2.42 0.90 37.5 0.32 0.03 3 72*** 4.67 abdomen post, part 30 1.75-2.95 2.44 0.36 15.1 0.75 0.07 1.60*** 1.93 carapace length 30 0.90-1.22 1.09 0.088 8.2 0.84 0.01 — — bulb length 30 0.35-0.38 0.36 0.009 2.4 0.10 0.005 Q 2j*** 0.29 procursus length 30 0.39-0.43 0.41 0.012 2.9 0.16 0.005 0 23*** 0.35 Females tibia 1 length 10 6.55-8.10 7.38 0.46 6.4 0.71 0.07 1,91** 2.37 tibia 3 length 14 3.05-3.78 3.51 0.21 6.1 0.71 0.03 1.39** 2.11 abdomen total length 14 2.00-2.70 2.34 0.18 7.9 1.00 0.02 1.07 n.s. — abdomen frontal part 14 0.82-0.92 0.88 0.031 3.6 0.52 0.02 0.45 n.s. — abdomen post, part 14 1.17-1.80 1.46 0.17 11.6 1.00 0.02 1.43 n.s. — carapace length 14 0.80-0.90 0.85 0.026 3.1 0.78 0.01 — — epigynum width 14 0.34-0.39 0.36 0.013 3.6 0.90 0.005 0.37 n.s. — (The reason that many more males were col- lected is simply that I always collected the males first in order to maximize the male sam- ple, and females often dropped out of the web before I could capture them.) Pholcus phal- angioides (Fuesslin 1775), the pholcid species E E £ 05 c c 7 6 5 4 3 2 1 • male ■ female t • 3 4 5 6 Tibia 3 length (mm) Figure 9. — Mecolaesthus longissimus, scatter of male and female abdomen lengths on tibia 3 lengths. studied in most detail, is apparently unusual in this regard as there is no evidence for mate guarding (Uhl 1998). Fights have been observed in pholcids (Eberhard 1992; Eberhard & Briceno 1985), and it is probable that males gain something by cohabiting with or guarding females and that they will fight to defend whatever re- source there is. The exact benefit males derive from staying with females is unknown. They might protect their sperm investment from competition with rival male ejaculates, be- cause in pholcids the second males may fer- tilize a large proportion of eggs (Eberhard et al. 1993; Kaster & Jakob 1997; Yoward 1998; Schafer & Uhl 2002). They might improve fe- male foraging efficiency, but chivalrous be- havior in pholcids might rather be a means to induce the female not to leave and thus make her defensible (Eberhard & Briceno 1983). They might aid the female to repel other mo- tivated males (Parker 1970). Finally, they might provide postinsemination signals to in- crease their chances of fathering their mate’s HUBER— DIRECTIONAL SELECTION ON MALE ABDOMEN SIZE 577 Female tibia 1 (log) 11 20' 2.0 Figures 10, 11. — Tibia 1 length dimorphism in Pholcidae. 10. Scatter of log-transformed male tibia 1 lengths on female tibia 1 lengths for 100 pholcid species. The line indicates monomorphism; 11. Histogram showing the ratio of male/female tibia 1 lengths in 100 pholcid species. offspring (Eberhard 1985; Alcock 1994). Whatever the details, male M. longissimus probably fight intruders, or try to expel resi- dents. Second, exaggerated morphologies and high variability of sexually dimorphic char- acters often seem to result from sexual selec- tion (Pomiankowski & Mpller 1995; Baker & Wilkinson 2001). For example, extreme male size variation in the salticid Zygoballus rufipes Peckham & Peckham 1885 was attributed to alternative male mating strategies (Faber 1994). Comparative evidence strongly sug- gests that female M. longissimus have retained the plesiomorphic abdomen size, and that males vary from ‘normal’ to extreme. All oth- er known species of Mecolaesthus have ‘nor- mal’ abdomens, not appreciably different from the abdomens of females and of other closely related genera (Huber 2000). Thus, male M. longissimus abdomens are exagger- ated sexual modifications. Third, there is no evidence pointing to eco- logical determinants of male abdomen size. The webs in which the specimens were col- lected appear identical to those of many New World pholcids, i.e. a distinct, loosely meshed and more or less domed sheet (Eberhard & Briceno 1985). Further observations on ecol- ogy are not available. Thus, sexual selection on male abdomen size appears as the most plausible explanation for the dimorphism in this species. Female choice might be involved, and a large abdo- men may be a costly and thus honest indicator of male quality (cf. Uetz et al. 2002). Alter- natively, cryptic female choice might select for exaggerated male testes or accessory gen- ital glands (cf. Eberhard 1996). However, nu- merous studies indicate that male-male fights are the most important force selecting for large male size (Christenson & Goist 1979; Watson 1990, 1991; review in Andersson 1994). Therefore, I hypothesize that male M. longissimus use their abdomens either to fight or to assess each other before fights. A large brown spot ventrally on the abdomen (Figs. 5-7) might be significant in this respect: the spot marks the posterior border of the anterior part of the abdomen, i.e. that part that is most extremely size dimorphic, has the highest re- gression coefficient, and is therefore the most reliable predictor of male size (cf. Taylor et al. 2000). Male M. longissimus carry their ab- domen more or less vertically (Fig. 3; see also fig. 439 in Simon 1893), making the spot po- tentially visible to conspecifics in the same web. Whether pholcids have the appropriate visual capabilities is unknown. A surprising but revealing result is the low regression value of male (but not female) tibia 1 (b[OLS] = 0.37) in M. longissimus. It is consistently higher in other pholcids studied: 0.88 in Metagonia mariguitarensis (Gonzalez- Sponga 1998) (Huber 2004), 1.00 in Buitinga safura Huber 2003 (Huber & Hopf 2004), 1.22 in Physocyclus globosus (Taczanowski 1874) (Eberhard et al. 1998). This would seem to indicate stabilizing selection in M. longis- simus, in contrast to other pholcids. I hypoth- 578 THE JOURNAL OF ARACHNOLOGY esize that the unusual regression value of male leg length and the unusual exaggerated ab- domen are directly correlated and that M. lon- gissimus males have changed from leg fights (the usual strategy in pholcids: Eberhard 1992; Eberhard & Briceno 1985) to abdomen fights, thus relaxing selection on leg length. However, this still requires an explanation for longer legs in M. longissimus males than in females. One potential explanation is phylo- genetic inertia, as nearly all pholcids have lon- ger male than female legs (see below). The tibia 1 measures across the entire fam- ily clearly show that male pholcids have con- sistently longer tibiae than females. Unfortu- nately, there are no comparable data on other size measures, as for example total body size. However, the reason for this missing data is that male and female pholcids usually are monomorphic regarding total size (Elgar 1992; pers. obs.). Collectively, this lends fur- ther support to the idea that single size mea- sures may not reliably reflect sexual size di- morphism in spiders (Prenter et al. 1995). The reasons for leg length dimorphism in pholcids are unknown. Longer legs may help cursorial males in their search for females (Montgom- ery 1910), they may provide males with a wide sensory radius and keep them relatively safe from female aggression (Elgar et al. 1990), or they may play a role in male-male fights (Eberhard 1992; Dodson & Beck 1993; Eberhard & Briceno 1985; Prenter et al. 1995; Bridge et al. 2000). Whatever the details, the consistent and fairly uniform pattern argues for a widely responsible cause or set of causes rather than for varying explanations in differ- ent taxa. ACKNOWLEDGMENTS I am indebted to Osvaldo Villarreal and Abel Perez Gonzalez for invaluable help get- ting a collection permit, and I thank the Di- reccion General de Eauna y Oficina Nacional de Diversidad Biologica in Caracas for issuing permit N° 01-1 1-0966. Paul Watson and three anonymous referees provided valuable com- ments on previous versions of the manuscript. Boris Striffler was a great companion in the field, and the Deutsche Eorschungsgemein- schaft provided the funds that made the col- lecting trip possible (DEG grant HU 980/1- 1). LITERATURE CITED Alcock, J. 1994. Postinsemination associations be- tween males and females in insects: the mate- guarding hypothesis. Annual Review of Ento- mology 39:1-21. Andersson, M. 1994. Sexual Selection. Monographs in Ecology and Evolution. Princeton University Press, Princeton. Baker, R.H. & G.S. Wilkinson. 2001. Phylogenetic analysis of sexual dimorphism and eye-span al- lometry in stalk-eyed flies. Evolution 55:1373- 1385. Beck, M.W. & E.E Connor. 1992. Factors affecting the reproductive success of the crab spider Mis- umenoides formosipes: the covariance between juvenile and adult traits. Oecologia 92:287-295. Bridge, A.R, R.W. Elwood & J.T.A. Dick. 2000. Imperfect assessment and limited information preclude optimal strategies in male-male fights in the orb-weaving spider Metellina mengei. Pro- ceedings of the Royal Society London B 267: 273-279. Christenson, T.E. & K.C. Goist. 1979. Costs and benefits of male-male competition in the orb- weaving spider Nephila clavipes. Behavioral Ecology and Sociobiology 5:87-92. Coddington, J.A., G. Hormiga & N. Scharff. 1997. Giant female or dwarf male spiders? Nature 385: 687-688. Dodson, G.N. & M.W. Beck. 1993. Pre-copulatory guarding of penultimate females by male crab spiders, Misumenoides formosipes. Animal Be- haviour 46:951-959. Eberhard, W.G. 1985. Sexual Selection and Animal Genitalia. Harvard University Press, Cambridge, Massachusetts. Eberhard, W.G. 1992. Notes on the ecology and be- haviour of Physocyclus globosus (Araneae, Phol- cidae). Bulletin of the British Arachnological So- ciety 9:38-42. Eberhard, W.G. 1996. Female Control: Sexual Se- lection by Cryptic Female Choice. Princeton University Press, New Jersey. Eberhard, W.G. 2002a. Natural history and behavior of Chymomyza mycopelates and C. exophthalma (Diptera: Drosophilidae), and allometry of struc- tures used as signals, weapons, and spore collec- tors. Canadian Entomologist 134:667-687. Eberhard, W.G. 2002b. The relation between ag- gressive and sexual behavior and allometry in Palaeosepsis dentatiformis (Diptera: Sepsidae). Journal of the Kansas Entomological Society 75: 317-332. Eberhard, W.G. & R.D. Briceno. 1983. Chivalry in pholcid spiders. Behavioral Ecology and Socio- biology 13:189-195. Eberhard, W.G. & R.D. Briceno. 1985. Behavior and ecology of four species of Modisimus and HUBER— DIRECTIONAL SELECTION ON MALE ABDOMEN SIZE 579 Blechroscelis (Araneae, Pholcidae). Revue Ar- achnologique 6:29-36. Eberhard, W.G., S. Guzman-Gomez & K.M. Catley. 1993. Correlation between spermathecal mor- phology and mating systems in spiders. Biolog- ical Journal of the Linnean Society 50:197-209. Eberhard, W.G., B.A. Huber, S.R.L. Rodriguez, R.D. Briceno, I. Salas & V. Rodriguez. 1998. One size fits all? Relationships between the size and degree of variation in genitalia and other body parts in twenty species of insects and spi- ders. Evolution 52:415-431. Eberhard, W.G., B.A. Huber & R.L. Rodriguez. 1999. Don’t forget the biology: a reply to Green. Evolution 53:1624-1627. Elgar, M.A. 1992. Sexual cannibalism in spiders and other invertebrates. Pp. 128-155. In Canni- balism: Ecology and Evolution among Diverse Taxa. (M.A. Elgar & B.J. Crespi, eds.). Oxford University Press, Oxford. Elgar, M.A. 1998. Sperm competition and sexual selection in spiders and other arachnids. Pp 307- 339. In Sperm Competition and Sexual Selec- tion. (T.R. Birkhead & A.R Mpller, eds.). Aca- demic Press, London. Elgar, M.A., N. Ghaffer & D. Read. 1990. Sexual dimorphism in leg length in orb- weaving spiders: a possible role for sexual cannibalism. Journal of Zoology London 222:455-470. Elgar, M. & B. Fahey. 1996. Sexual cannibalism, competition, and size dimorphism in the orb- weaving spider Nephila plumipes Latreille (Ar- aneae: Araneoidea). Behavioral Ecology 7:195- 198. Elgar, M.A. & D.R. Nash. 1988. Sexual cannibal- ism in the garden spider, Araneus diadematus (Araneae: Araneidae). Animal Behaviour 36: 1511-1517. Faber, D.B. 1994. Przibram’s rule and male body- size dimorphism in Zygoballus rufipes (Araneae: Salticidae), Journal of Zoology London 232: 191-198. Funke, S. & B.A. Huber. In press. Allometry of genitalia and fighting structures in Linyphia tri- angularis (Araneae, Linyphiidae). Journal of Ar- achnology Green, A.J. 1992. Positive allometry is likely with mate choice, competitive display and other func- tions. Animal Behaviour 43:170-172. Green, A.J. 1999. Allometry of genitalia in insects and spiders: one size does not fit all. Evolution 53:1621-1624. Harvey, P.H. & M.D. Pagel. 1991. The Comparative Method in Evolutionary Biology. Oxford Univer- sity Press, Oxford. Head, G. 1995. Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (class Araneae). Evolution 49: 776-781. Hormiga, G., N. Scharff & J.A. Coddington. 2000. The phylogenetic basis of sexual size dimor- phism in orb-wcaving spiders (Araneae, Orbi- culariae). Systematic Biology 49:435-462. Huber, B.A. 1997a. On the distinction between Modisimus and Hedypsilus (Araneae, Pholcidae), with notes on behavior and natural history. Zoo- logica Scripta 25:233-240. Huber, B.A. 1997b. On American "Micromerys' and Metagonia (Araneae, Pholcidae), with notes on natural history and genital mechanics. Zoologica Scripta 25:341-363. Huber, B.A. 1997c. The pholcid spiders of Costa Rica (Araneae: Pholcidae). Revista de Biologia Tropical 45:1583-1634. Huber, B.A. 1998a. Notes on the neotropical spider genus Modisimus (Pholcidae, Araneae), with de- scriptions of thirteen new species from Costa Rica and neighboring countries. Journal of Ar- achnology 26:19-60. Huber, B.A. 1998b. Report of some pholcid spiders collected in Guatemala and Honduras (Araneae, Pholcidae). Revue Suisse de Zoologie 105:49- 80. Huber, B.A. 2000. New World pholcid spiders (Ar- aneae: Pholcidae): a revision at generic level. Bulletin of the American Museum of Natural History 254:1-348. Huber, B.A. 2001. The pholcids of Australia (Ara- neae: Pholcidae): taxonomy, biogeography, and relationships. Bulletin of the American Museum of Natural History 260:1-144. Huber, B.A. 2003a. Cladistic analysis of Malagasy pholcid spiders reveals generic level endemism: Revision of Zatavua n. gen. and Paramicromerys Millot (Pholcidae, Araneae). Zoological Journal of the Linnean Society 137:261-318. Huber, B.A. 2003b. High species diversity in one of the dominant groups of spiders in East African montane forests (Araneae: Pholcidae: Buitinga n. gen., Spermophora Hentz). Zoological Journal of the Linnean Society 137:555-619. Huber, B.A. 2003c. Southern African pholcid spi- ders: revision and cladistic analysis of Quamtana n. gen. and Spermophora Hentz (Araneae: Phol- cidae), with notes on male-female covariation. Zoological Journal of the Linnean Society 139: 477-527. Huber, B.A. 2004. Evidence for functional segre- gation in the directionally asymmetric male gen- italia of the spider Metagonia mariguitarensis (Gonzalez-Sponga) (Pholcidae: Araneae). Jour- nal of Zoology London 262:317-326. Huber, B.A. & S. Hopf. 2004. Color dimorphism in adults and juveniles of Buitinga safura (Araneae, Pholcidae). Journal of Arachnology 32:336-340. Huber, B.A. & A. Perez G. 1998. Leptopholcus de- licatulus (Araneae, Pholcidae) is a valid name. Journal of Arachnology 26:251-256. 580 THE JOURNAL OF ARACHNOLOGY Huber, B.A. & A. Perez G. 2001. A new genus of pholcid spiders (Araneae: Pholcidae) endemic to Western Cuba, with a case of female genitalic dimorphism. American Museum Novitates 3329: 1-23. Huber, B.A., Rheims, C.A. & A.D. Brescovit. 2005. Two new species of litter-dwelling Metagonia spiders (Araneae, Pholcidae) document both rap- id and slow genital evolution. Acta Zoologica. (Stockholm) 86:33-40. Raster, J.L. & E.M. Jakob. 1997. Last-male sperm priority in a haplogyne spider (Araneae: Pholci- dae): correlations between female morphology and patterns of sperm usage. Annals of the En- tomological Society of America 90:254-259. Kotiaho, J., R.V. Alatalo, J. Mappes & S. Parri. 1997. Fighting success in relation to body mass and drumming activity in the male wolf spider Hygrolycosa nibrofasciata. Canadian Journal of Zoology 75:1532-1535. Kotiaho, J., R.V. Alatalo, J. Mappes & S. Parri. 1999. Honesty of agonistic signalling and effects of size and motivation asymmetry in contests. Acta Ethologica 2:13-21. Lang, G.H.P. 2001. Sexual size dimorphism and ju- venile growth rate in Linyphia triangularis (Lin- yphiidae, Araneae). Journal of Arachnology 29: 64-71. Montgomery, TH. 1910. The significance of the courtship and secondary sexual characters of Ar- aneads. American Naturalist 44:151-177. Moya-Larano, J., J. Halaj & D.H. Wise. 2002. Climbing to reach females: Romeo should be small. Evolution 56:420-425. Nielsen, N. & S. Toft. 1990. Alternative male mat- ing strategies in Linyphia triangularis (Araneae, Linyphiidae). Acta Zoologica Fennica 190:293- 297. Parker, G.A. 1970. Sperm competition and its evo- lutionary consequences in the insects. Biological Reviews 45:525-567. Petrie, M. 1992. Are all secondary sexual display structures positively allometric and, if so, why? Animal Behaviour 43:173-175. Pollard, S D. 1994. Consequences of sexual selec- tion on feeding in male jumping spiders (Ara- neae: Salticidae). Journal of Zoology London 234:203-208. Pomiankowski, A. & A.R Mpller. 1995. A resolu- tion of the lek paradox. Proceedings of the Royal Society London B 260:21-29. Prenter, J., 1. Montgomery & R.W. El wood. 1995. Multivariate morphometries and sexual dimor- phism in the orb-web spider Metellina segmen- tata (Clerck, 1757) (Araneae, Metidae). Biolog- ical Journal of the Linnean Society 55:345-354. Prenter, J., 1. Montgomery & R.W. Elwood. 1997. Sexual dimorphism in northern temperate spi- ders: implications for the differential mortality model. Journal of Zoology, London 243:341- 349. Prenter, J., R.W. Elwood & W.I. Montgomery. 1998. No association between sexual size dimor- phism and life histories in spiders. Proceedings of the Royal Society London B 265:57-62. Prenter, J., R.W Elwood & W.I. Montgomery. 1999. Sexual size dimorphism and reproductive investment by female spiders: a comparative analysis. Evolution 53:1987-1994. Prenter, J., R.W. Elwood & W.I. Montgomery. 2003. Mate guarding, competition and variation in size in male orb-web spiders, Metellina seg- mentata: a field experiment. Animal Behaviour 66:1053-1058. Rovner, J.S. 1968. Territoriality in the sheet-web spider Linyphia triangularis (Clerck) (Araneae, Linyphiidae). Zeitschrift fiir Tierpsychologie 25: 232-242. Schafer, M.A. & G. Uhl. 2002. Determinants of pa- ternity success in the spider Pholcus phalangioi- des (Pholcidae: Araneae): the role of male and female mating behaviour. Behavioral Ecology and Sociobiology 51:368-377. Schaefer, D. & G. Uhl. 2003. Male competition over access to females in a spider with last-male sperm priority. Ethology 109:385-400. Schneider, J.M., M.E. Herberstein, EC. de Crespig- ny, S. Ramamurthy & M.A. Elgar. 2000. Sperm competition and small size advantage for males of the golden orb-web spider Nephila edulis. Journal of Evolutionary Biology 13:939-946. Schiitz, D. & M. Taborsky. 2003. Adaptations to an aquatic life may be responsible for the reversed sexual size dimorphism in the water spider, Ar- gyroneta aquatica. Evolutionary Ecology Re- search 5:105-1 17. Simon, E. 1893. Histoire Naturelle des Araignees. 2nd edition, l(2):257-488. Roret, Paris. Tatsuta, H., K. Mizota & S.-I. Akimoto. 2001. Al- lometric patterns of heads and genitalia in the stag beetle Lucanus maculifemoratus (Coleop- tera: Lucanidae). Annals of the Entomological Society of America 94:462-466. Taylor, P.W., O. Hasson & D.L. Clark. 2000. Body postures and patterns as amplifiers of physical condition. Proceedings of the Royal Society Lon- don B 267:917-922. Taylor, P.W., O. Hasson & D.L. Clark. 2001. Initi- ation and resolution of jumping spider contests: roles for size, proximity, and early detection of rivals. Behavioral Ecology and Sociobiology 50: 403-413. Toft, S. 1989. Mate guarding in two Linyphia spe- cies (Araneae: Linyphiidae). Bulletin of the Brit- ish Arachnological Society 8:33-37. Uetz, G.W, R. Papke & B. Kilinic. 2002. Influence of feeding regime on body size, body condition and a male secondary sexual character in Schi- HUBER— DIRECTIONAL SELECTION ON MALE ABDOMEN SIZE 581 zocosa ocreata wolf spiders (Araneae, Lycosi- dae): conditioe^dependeece in a viseal signaling trait, Joumai of Arachnology 30:461-469. Uhi, G. 1998. Mating behaviour in the cellar spider, Pholcm phalangioides, indicates sperm mixing. Animal Behaviour 56:1155-1159. Vollrath, E 1980. Male body size and fitness in the web-building spider Nephila clavipes. Zeitschrift fur Tierpsychologie 53:61-78. Vollrath, E & G.A. Parker. 1992. Sexual dimor- phism and distorted sex ratios in spiders. Nature 360:156-159. Walker, S.E. & A.L. Rypstra. 2003. Sexual dimor- phism and the differential mortality model: is be- haviour related to survival? Biological Journal of the Linnean Society 78:97-103. Watson, PJ. 1990. Female-enhanced male compe- tition determines the first mate and principal sire in the spider Linyphia litigiosa (Linyphiidae). Behavioral Ecology and Sociobiology 26:77-90. Watson, P.J. 1991. Multiple paternity and first mate sperm precedence in the sierra dome spider, Lin- yphia litigiosa Keyserling (Linyphiidae). Animal Behaviour 41:135-148. Toward, P.J. 1998. Sperm competition in Pholcus phalangioides (Fuesslin, 1775) (Araneae, Phol- cidae)— shorter copulations gain higher paternity reward than first copulations. Pp. 176-170. In Proceedings of the 17th European Colloquium of Arachnology, Edinburgh 1997. (P.A. Selden, ed.). British Arachnological Society, Burnham Beeches, Bucks. Manuscript received 9 September 2004, revised 26 February 2005. 2005. The Journal of Arachnology 33:582-590 EFFECTS OF PREY QUALITY ON THE LIFE HISTORY OF A HARVESTMAN Aino Hvam and S0ren Toft: Department of Ecology & Genetics, University of Aarhus, Denmark, Bldg. 1540 DK-8000 Arhus C, Denmark. ABSTRACT. Information on the value of various food types for harvestmen is sparse. The aim of this study was, therefore, to clarify the quality of six different food types to a harvestman. Survival, growth and development were used as measures of fitness in a laboratory experiment. Recently hatched Oligo- lophus tridens were fed the following experimental diets until maturity: Drosophila melanogaster (Dip- tera), entomobryid Collembola (Tomocerus bidentatus! Sinella curviseta), Folsomia Candida (Collembola), Sitobion avenae (Aphidoidea), Rhopalosiphum padi (Aphidoidea), and a mixed diet containing the five prey types. Survival and growth rate were high on the D. melanogaster and entomobryid diets, and low on the F. Candida, S. avenae and R. padi diets. The mixed diet caused a high early mortality, later a good survival and a high growth rate. The majority of harvestmen on the D. melanogaster and entomobryid diets matured. None of the harvestmen fed pure aphid diets developed beyond the fourth instar, and only few from the F. Candida diet matured. Overall, the diets separate in three levels: D. melanogaster and the entomobryid diet were high-quality, the mixed diet was intermediate, and the two aphid diets and F. Candida diet were low-quality. In general, the quality ranking agrees with that of other generalist predators, though there are differences in details. Keywords: Opiliones, Oligolophus tridens, fitness, diet Harvestmen are omnivorous generalists, with a variety of feeding habits, ranging from plant eating to predation. They eat a range of small invertebrates, which are probably caught live and killed. Examples of the inver- tebrate diet are: springtails, aphids, snails, earthworms, other harvestmen and spiders (Sankey & Savory 1974). Harvestmen are also scavengers and will scavenge both on in- vertebrates (Sankey & Savory 1974) and ver- tebrates (Sankey 1949). Studies have shown that harvestmen generally prefer small prey, such as Hemiptera and Collembola (Adams 1984). In the laboratory, harvestmen have successfully been fed odd diets, e.g., bananas, cooked vegetables, ham, cream cheese (Gnas- pini 1996), dried eggs, whole meal flour, yeast (Todd 1949), together with live animal food. Gnaspini (1996) tested which food types the harvestman Goniosoma spelaeum (Mello-Lei- tao 1932) will accept and came to the conclu- sion that these harvestmen should be consid- ered “omnivores tending to carnivory”. There are only few laboratory studies of the feeding ecology of harvestmen, so information on the quality of specific diets is very sparse. Laboratory studies on the quality of differ- ent food types have been conducted on several generalist predators such as spiders (Toft 1995; Toft & Wise 1999a) and carabid beetles (Bilde & Toft 1999). The quality of different prey can be evaluated by comparing fitness parameters of the predator kept on different dietary treatments. Different fitness parame- ters can be used as quality measures: survival, body mass, time used in development, size of body parts, fecundity etc. In this study, sur- vivorship, growth and development were used. Young harvestmen {Oligolophus tridens (C.L. Koch 1836)) were reared on six diets and the quality of each diet was assessed by comparison of the fitness measures. In the ex- periment we tested monotypic diets of a fly, two springtails, two aphids, and a mixed diet of all five. These prey types were chosen be- cause they represent ordinary harvestman prey from different invertebrate orders and most of the prey can be found in the same habitats as O. tridens. Furthermore, the aphids {Sitobion avenae (Fabricius) and Rhopalosi- phum padi (Linneaus)) are pests in agricultur- al fields, and are among the most abundant aphids in cereal fields (Wiktelius 1982). The prey types used in the present study 582 HVAM & TOFT— EFFECTS OF PREY QUALITY ON A HARVESTMAN 583 have been evaluated in other studies of gen- eralist predators as well. In the light of these results we expected that fruit flies, Drosophila melanogaster (Meigen), and the entomobryid springtails Tomocerus bidentatus (Fo\som)/Si- nella curviseta (Brook) would be of good quality to the harvestmen (Toft & Wise 1999a; Vanacker et aL 2004). The aphids S, avenae and R. padi and the collembolan Folsomia Candida (Willem) are of poor quality to spi- ders and carabid beetles. Only a few individ- uals molted when wolf spiders (Pardosa pra- tivaga (L. Koch 1870)) were fed the aphids (Toft 2000). Wolf spiders (Schizocosa sp.) fed F. Candida survived for a shorter period than the starved controls (Toft & Wise 1999a) and F. Candida was therefore considered a toxic prey. We expected that these findings would also apply to harvestmen, because both wolf spiders and harvestmen are generalist preda- tors and can be found in the same habitats. It is more difficult to predict the consequences of the mixed diet. The effect depends on the effect of each prey type, how much the har- vestmen eat of each prey and if there is an interaction of the effects between some of the prey when harvestmen are fed a mixed diet. METHODS The harvestman. — The harvestman Oli- golophus tridens occurs all over northern and central Europe (Martens 1978) and has also been reported from North America (Bell 1974), The species is abundant in Denmark and can be found in a variety of habitats, es- pecially in woodlands, roadsides and in gar- dens. The harvestman has a body length of 4- 5 mm (males) or 5-6.5 mm (females) (Sankey 6 Savory 1974). The life cycle is annual and the harvestmen overwinter in the egg stage. Hatchlings emerge in spring (in Denmark April-May, Meinertz 1964; per. obs.). The first molt takes place a few hours after the harvestmen emerge from the egg (Martens 1978) and they pass through 6 juvenile instars before they mature (Pfeifer 1956; Phillipson 1962). Prey.— All the prey animals came from laboratory cultures. The prey was freeze killed and provided in surplus amounts. Wild type fruit flies {Drosophila melanogaster) were reared on instant Drosophila medium (For- mula 4-24, Carolina Biological Supply; Bur- lington, NC, USA) mixed with crushed dog food (Techni-Cal® ADULT, Martin Pet Foods, Ontario, Canada) in a proportion of 100 g of Drosophila medium to 54.5 g dog food. The enrichment ensured a high nutritional quality of the flies, especially regarding proteins. En- riched fruit flies increased growth and survival in a wolf spider (Mayntz & Toft 2001) and supported a high egg production in a carabid beetle (Bilde et al. 2000). Folsomia Candida was raised on baker’s yeast. The entomobryids Tomocerus bidentatus! Sinella curviseta were both raised on baker’s yeast and Drosophila medium. At the beginning of the experiment the harvestmen in the entomobryid group were fed T. bidentatus, but as the culture was slow and there was a risk of food shortage, the harvestmen were fed S. curviseta from week 6. Both T, bidentatus and S. curviseta are considered to be prey of high quality {T. bidentatus: Toft & Wise 1999a; S. curviseta: Vanacker et al. 2004). Rhopalosiphum padi and Sitobion avenae were both raised on wheat seedlings of mixed cultivars. Mixed stages of springtails and aphids were used to feed the harvestmen. The experiment. — Young O. tridens in the second instar were collected in a small forest near Arhus, Denmark, 56°07'N, 10°00'E, in late April 2003 by sifting leaf litter over a white tray. The harvestmen were kept individ- ually in plastic tubes (diameter 2 cm, height 6 cm) with a moistened bottom layer of plas- ter mixed with charcoal and a foam rubber plug. Throughout the experiment the harvest- men were kept at a constant temperature of 17 °C, and a photoperiod of 16L:8D. The har- vestmen were weighed the day after collection and assigned to one of six diet treatments with roughly the same distribution of body masses. The treatments were: D. melanogaster, T. bb dentatusiS. curviseta (both Entomobryidae), F. Candida (Isotomidae), S. avenae, R. padi, and a mixed diet with about equal amounts of the five prey types. Some of the replicates were discarded because of escapes and acci- dents. The number of replicates in each treat- ment therefore varied from 18-22. The har- vestmen were transferred to larger plastic tubes (diameter 3.5 cm, height 8 cm) after the third molt. Prey and water were renewed, and mortality and molts were checked three times per week. The duration of instar 2 was recorded as the number of days from collection to the next 584 THE JOURNAL OF ARACHNOLOGY 1 Weeks Figures 1-2. — 1. Survivorship curves for harvestmen Oligolophus tridens. The harvestmen were raised in the laboratory from the second instar to maturity on six different diets. Curves with different letters are significantly different. 2. Growth curves for harvestmen Oligolophus tridens. The harvestmen were raised in the laboratory from the second instar to maturity on six different diets. “Males” are males from the diets: Drosophila melanogaster, entomobryid springtails and mixed diet. Harvestmen fed aphids died before the sex could be determined. Folsomia Candida data for males and females were pooled because of the low number. Error bars are only shown every second week for the sake of clarity. molt. A few molts were missed. As the molts progressed synchronously within each diet treatment a molt date was estimated for the missing molts, using the average molt date for the harvestmen in the same treatment. When a molt was observed, the midpoint between two days in which the tubes were checked, was used to compute the parameter “days in j; instar”. The harvestmen were weighed week- ly (Sartorius electronic balance MC5: 0.001 mg accuracy) to measure growth rate. The most recent weighing before the molt was used for the parameter “weight at molt”. The experiment was terminated after 10 weeks. HVAM & TOFT— EFFECTS OF PREY QUALITY ON A HARVESTMAN 585 Statistical analysis. — The survivorship data were tested with the Log Rank test (Pyke & Thompson 1986). The pairwise Log Rank comparisons were not corrected with sequen- tial Bonferroni adjustment (Moran 2003), be- cause the prey types were chosen based on prior assumptions and the relatively high number of prey types would make it unrea- sonably difficult to obtain any significance af- ter adjustment. The growth curves were com- pared using multivariate analysis of variance (MANOVA) with repeated measures, with time (weeks) as the repeated factor. The time * diet interaction term was used to detect dif- ferences in growth over time between the treatments. However, animals that died before the end of the experiment were excluded from the analysis. We analyzed the growth data for all treatments for only three weeks or approx- imately 50% of their maturation time, at which time there were still harvestmen in all treatments. Body mass changes from start of the experiment to week three were tested with one-way ANOVA. The data were log trans- formed to achieve variance homogeneity (Levene’s test a > 0.05). A post hoc test was used to locate the differences indicated in the overall ANOVA; because the treatments were chosen to test potential harvestman food, and all the comparisons therefore were planned, a Student’s t-test was applied. For the treat- ments: D. melanogaster, entomobryids and the mixed diet, repeated measures analysis of body mass was carried out for the full exper- imental period. The duration of the instars and the “weight at molt” were analyzed with one- way ANOVA. The data were transformed when the assumption of homogeneity of var- iance was not met (for details, see Results). Post hoc mean comparisons between treat- ments were done with Student’s t-test. Fur- thermore a two-way ANOVA was used to test for any interaction between sex and treatment on development. All statistical analyses were performed with IMP 5.0 for windows (SAS institute). RESULTS Survivorship. — There was an overall sig- nificant treatment effect on survival (Log Rank test, = 78.7382, P < 0.0001, Fig. 1). The pairwise comparisons separated treat- ments into three groups: D. melanogaster and the entomobryids were of the same high qual- Figure 3.— -Body mass change in the harvestman OUgolophus tridens, from the beginning of the ex- periment to week 3 (mg, mean + SE). Bars with different letters are significantly different (ANOVA, Student’s t-test). ity. The mixed diet was intermediate and the two aphids and F. Candida were of low qual- ity. Four individuals from the F. Candida diet survived to the end of the experiment. None of the aphid-fed harvestmen survived. Growth. — For the first three weeks there was a significant overall time * diet interac- tion on the body masses (MANOVA, n = 86, Wilk’s X - 0.1355, F = 15.2856, NumDF = 15, DenDF - 215.73, P < 0.0001, Fig. 2). The ranking of the diets was: D. melanogaster > entomobryid = the mixed diet >> the two aphid diets > F. Candida. This is supported by an ANOVA test on the body mass change over the first three weeks of the experiment (overall ANOVA test on In-transformed data, n = 86, ^5^80 - 94.9672, P < 0.0001, Fig. 3). The repeated measures test was also done for the first three weeks on the animals that ma- tured (from the treatments D. melanogaster, entomobryids and mixed diet), with both treatment and sex as factors. The test showed that there was no significant time * diet * sex interaction on the body mass (MANOVA, n - 53, Wilk’s X - 0.9305, F = 0.5502, NumDF — 6, DenDF = 90, P < 0.77). There was a significant time * diet interaction on the body mass among the females from the high- quality treatments over all 10 weeks of the experiment (MANOVA, n = 26, Wilk’s X = 0.0528, F = 4.6934, NumDF = 20, DenDF - 28, P < 0.0001). Contrast tests showed that 586 THE JOURNAL OF ARACHNOLOGY the three diets all differed in body mass over time (P < 0.006), though D. melanogaster and mixed diet ended up at the same level. If males were included in the test there was a significant time * diet * sex interaction on the body mass over 10 weeks (MANOVA, n = 53, Wilk’s X = 0.2133, F = 4.4273, NumDF = 20, DenDF = 16, P < 0.0001). Development. — The harvestmen were in the second instar at collection. The maturation success was high on the D. melanogaster (95%), and the entomobryid (100%) diets, and the majority of harvestmen from the mixed diet matured (67%). Development was re- stricted on the aphid and F. Candida diets and many of the harvestmen on these diets never molted. None of the harvestmen fed aphids molted to the fifth instar and only 19% from the F. Candida diet matured. Generally the harvestmen from the D. melanogaster and mixed diet were the fastest to complete an in- star and harvestmen from the F. Candida treat- ment were the slowest (Fig. 4, right column). The total number of days from collection to the last molt showed a significant effect of diet (ANOVA test on In-transformed data, n - 58, ^3,54 = 58.2450, P < 0.0001). The har- vestmen from the D. melanogaster diet were the first to complete their development (39.0 ± 0.68 days, mean ± SE), mixed diet and the entomobryid took a few days more (41.1 ± 0.62; 43.6 ± 0.86) and those from F. Candida were the last (68.5 ± 3.06 days). As to the “weight at molt”, the D. melanogaster and entomobryid diets resulted in the heaviest an- imals and the F. Candida diet resulted in a low body mass, which is particularly evident at the last molts (Fig. 4, left column). Both male and female data are included in Fig. 4. After five weeks the sex of the surviving harvestmen be- came apparent. The males reached a body mass of approximately 17 mg which was maintained with minor fluctuations (Fig. 2). The females increased their body mass con- siderably after maturation. A two-way analy- sis of variance was used to test for any sex- specific growth patterns (only the animals that matured from the treatments: D. melanogas- ter, entomobryids and mixed diet). There was no interactions (treatment * sex, P > 0.11), but there were significant effects of sex on the duration of instar 2, 5 and 6; and on the weight at the 6th molt (Table 1). Weight at molt Duration of instar 20 15 10 5 0 15 10 5 0 20 15 10 5 0 20 15 10 5 0 15 10 5 Q” Figure 4. — Weight at molt (mg, mean + SE) and duration of instars (days, mean -I- SE) of harvest- men, Oligolophus tridens, reared on six diets. The dataset were tested with ANOVA, data were trans- formed if necessary. Overall P-values indicate sig- nificance. Pair wise comparisons were made with Students t-test; bars with different letters are sig- nificantly different. Male and female data are pooled in the figures. DISCUSSION When the three fitness parameters, survival, growth and development are combined, the overall conclusion is that the diets separate in three different quality levels. The two aphids, S. avenae and R. padi, and the springtail F. Candida were low-quality diets; both diets af- fected survival and growth. The development of the harvestmen was slow on the F. Candida HVAM & TOFT— EFFECTS OF PREY QUALITY ON A HARVESTMAN 587 diet, whereas harvestmen on the two aphid di- ets were only slightly slower than from the high-quality diets. Overall the mixed diet was of intermediate quality. It was of high quality regarding growth and development. Among the small juvenile harvestmen, the mixed diet caused a high mortality, an affect not seen in older animals. Drosophila melanogaster and entomobryids were high-quality prey, with re- spect to all three parameters. These results agree in general with the findings of other studies of generalist predators. Drosophila melanogaster and entomobryids have been re- ported to be of high quality to spiders (Toft & Wise 1999a; Mayntz & Toft 2001). Aphids are usually found to be of low quality to spiders (Toft 1995, 2000) and beetles (Bilde & Toft 1999), and F. Candida was classified as a toxic prey to wolf spiders (Toft & Wise 1999a). A pronounced sexual size dimorphism was de- tected in the present experiment. The growth of the males stopped when the males were subadult, whereas the females gained body mass throughout the experiment and became much larger than the males. A large body mass is more important for females than for males, because females invest more in repro- duction. The result of this fecundity selection is that females often are larger than males in invertebrates (Head 1995). The females in this experiment were generally faster to complete an instar and they reached maturity about 3 days before the males. In a study on linyphiid spiders, Toft (1995) found that when the females were fed normal fruit flies, the hatching success of the spider eggs was high for the first two or three egg sacs, but then the hatching success declined. The quality of fruit flies can be improved by enrichment of the media with extra proteins, for example by adding dog food (Mayntz & Toft 2001). However, even a fruit fly diet, with or without enrichment, has its restrictions. Al- though it was the best prey of the study, pro- tein enriched fruit flies was not fully sufficient for a wolf spider, as mortality and molting failures were higher than expected (Mayntz & Toft 2001). In the present study, mortality was low on the fruit fly diet and there were ap- parently no molting failures. The effects of mixed diets are varied. Some studies have shown that dietary mixing is ben- eficial and essential to survival and develop- ment (Lowrie 1987; Uetz et al. 1992), others that it depends on what the mixed diet consists of, i.e. it has to be the right mix (Marcussen et al. 1999; Toft & Wise 1999a). In this study the mixed diet caused a high mortality at the beginning, which might be due to the low- quality parts of the diet, i.e., F. Candida and the aphids. If low-quality and potentially toxic prey comprise a large part of the diet, a mixed diet may not be beneficial. Those that sur- vived the first few weeks may either have had a physiological tolerance to the low-quality prey or been able to reject them. If the sur- viving harvestmen in the mixed diet group de- veloped an increased preference for the high- quality prey, their diet basically consisted of a mix of two high-quality preys. If high-qual- ity prey is provided, there might be no or even negative effects of adding other prey types. In this study it seems that a monotonous high- quality diet, as for example D. melanogaster or the entomobryids, is better than a mixed diet of high-quality and low-quality or poten- tially toxic elements. In this experiment the F. Candida diet was of low quality to the harvestmen, both regard- ing survival, growth and development. Some of the harvestmen from the F. Candida treat- ment survived, gained weight and molted to maturity. This shows that they did eat F. Can- dida and that some of the harvestmen must have been more tolerant to the potentially tox- ic components in this diet than others. Fol- somia Candida is toxic to spiders, as spiders fed F. Candida died faster than starved con- trols (Toft & Wise 1999a) and they cannot complete their development on a diet of pure F. Candida (Fisker & Toft 2004); furthermore F. Candida induced a specific feeding aversion in a spider (Toft & Wise 1999b). It was there- fore a surprise that some of the harvestmen in the present experiment survived and devel- oped. This could indicate genetic variation in the ability to cope with the toxic collembolan (cf. Beck & Toft 2000). It is possible that at least some of the harvestmen are better able to overcome the chemical defenses in F Can- dida than the spiders are. The harvestman Mi- topus morio (Fabricius 1799) can tolerate the defensive alkaloids of their leaf beetle prey, by avoiding bioactivation and by rapid elim- ination of the detoxification products via the feces (Hartmann et al. 2003). Perhaps a sim- ilar process is operating in O. tridens, but with a high variation in individual ability to tolerate 588 THE JOURNAL OF ARACHNOLOGY Table L — Weight at molt (mg, mean ± SE) and duration of instar (days, mean ± SE) in male and female Oligolophus tridens from the treatments: Drosophila melanogaster, entomobryids and mixed diet. Only animals that matured are included in the tests. Welch ANOVA was used if the assumption of homogeneity of variance was not met. Asterisks indicate level of significance (*P < 0.05, **p < 0.01, < 0.001). Weight at molt (mg) Duration of instar (days) d $ d 2nd molt/instar 2 1.09 ± 0.06 1.16 ± 0.07 — 7.15 ± 0.33 5.74 ± 0.34 ** 3rd molt/instar 3 2.24 ±0.15 2.26 ± 0.15 — 6.40 ± 0.30 6.23 ± 0.31 — 4th molt/instar 4 3.72 ± 0.22 3.67 ± 0.23 — 7.59 ± 0.25 7.71 ± 0.26 — 5 th molt/instar 5 9.13 ± 0.28 8.80 ± 0.29 — 8.68 ± 0.22 7.99 ± 0.22 6th molt/instar 6 16.65 ± 0.50 19.42 ± 0.53 12.92 ± 0.31 11.86 ± 0.32 * F. Candida. It is possible that freeze-killing of the prey, as used in this study, can alter the chemical composition, compared to live ani- mal prey. However, a study of the carabid bee- tle Bembidion lampros (Herbst) showed that freeze-killing did not change the palatability of the springtails used as food (Bilde et al. 2000). The few harvestmen fed F. Candida that survived and matured obtained a lower body mass compared to the harvestmen from the other three diets. It is possible that F. Candida contains toxic substances that impede devel- opment. In a study of a linyphiid spider it was suggested that Folsomia fimetaria (Linneaus) “contains an element that inhibits digestion” (Marcussen et al. 1999). A similar result was seen in a study of a wolf spider (Pardosa pra- tivaga) in which F. Candida apparently inhib- ited the utilization of a better quality prey (D. melanogaster) (Fisker & Toft 2004). Pardosa prativaga compensated for the toxic effect of F. Candida by increasing the intake of D. me- lanogaster, but the spiders still showed a higher mortality and grew more slowly than spiders fed only D. melanogaster (Fisker & Toft 2004). If the harvestmen on the mixed diet ate F. Candida, they might have been ex- posed to toxins that decrease the digestion and/or utilization of the high-quality parts of the diet, and thereby caused a high mortality in the first few weeks. The high early mortal- ity in the mixed diet and in the F. Candida diet can also be explained by the size of the harvestmen. Studies have shown that small ju- venile wolf spiders are more dramatically af- fected by F. Candida than are larger juveniles (Toft & Wise 1999b; Fisker & Toft 2004). This also seems to be the case for harvestmen. Low ranking of aphids as food is wide- spread among generalist predators (Bilde & Toft 1994, 1999; Toft 1995). This experiment shows that O. tridens cannot survive on a pure aphid diet. High mortality was also the result in an experiment with larvae of the staphyli- nid beetle Tachyporus hypnorum (Fabricius) (Kyneb & Toft 2004). Aphids can also affect development. Wolf spiders {Pardosa amentata (Clerck 1757)) were unable to go through the first molt, and all the spiders died within two weeks, when fed a pure aphid diet (Toft 1995). These studies also indicate a limitation on the quantity of aphids the spiders and beetles can tolerate (Bilde & Toft 1994; Toft 1995). The food consumption was not measured in the present experiment, but it is very likely that the harvestmen consumed considerably fewer aphids than fruit flies. The prey in this study was freeze killed, before being offered to the harvestmen. This process neutralized the si- phuncular defense system of the aphids, mak- ing predation easier (Toft 1995). The low quality of the aphid diet must therefore rely on a deterrent or toxic substance in the aphids which prevents the harvestmen from utilizing the nutrients. Dixon & McKinlay (1989) stud- ied aphid predation by harvestmen in a potato field. They state that opilionids have been ne- glected and probably undervalued as predators of crop pests. A microcosm study with R. padi and different generalist predators showed that O. tridens was the most efficient predator; re- ducing aphid numbers up to 97% as compared to predator-free controls (Madsen et al. 2004). HVAM & TOFT— EFFECTS OF PREY QUALITY ON A HARVESTMAN 589 These studies of aphids and harvestmen are in contrast to our results ^ from which it seems unlikely that harvestmen, at least not O. tri- dens, can act as a powerful biocoetrol agent. Harvestmen may, however, contribute to the combined effect of the generalist predator complex on aphid population growth (Sy- moedson et aL 2002). ACKNOWLEDGMENTS We thank J. Mertens, Ghent, for the Sineiia culture as well as J. Viegborg and T Bilde for comments on the manuscript. LITERATURE CITED Adams, J. 1984, The habitat and feeding ecology of woodland harvestmen (Opiliones) in England. Oikos 42;361--370. Bell, R.T 1974. A European harvestman in North America (Phalangida, Phalangiidae). Entomology ical News 85:154. Bilde, T. & S. Toft. 1994. Prey preference and egg production of the carabid beetle Agonum dorsale. Entomologia Experimentalis et Applicata 73: 15N156. Bilde, T. & S. Toft. 1999. Prey consumption and fecundity of the carabid beetle Calathus melan- ocephalus on diets of three cereal aphids: high consumption rates of low-quality prey. Pedo- biologia 43:422-429. Bilde, T, J.A. Axelsen & S. Toft. 2000. The value of Collembola from agricultural soils as food for a generalist predator. Journal of Applied Ecology 37:672-683. Dixon, P.L. & R.G. McKinlay. 1989. Aphid pre- dation by harvestmen in potato fields in Scot- land. The Journal of Arachnology 17:253-255. Fisker, E.N. & S. Toft. 2004. Effects of chronic ex- posure to a toxic prey in a generalist predator. Physiological Entomology 29:129-138. Gnaspini, P. 1996, Population ecology of Gonioso- ma spelaeum, a cavernicolous harvestman from south-eastern Brazil (Arachnida: Opiliones: Gon- yleptidae). Journal of Zoology, London 239:417- 435. Hartmann, T., H. Haggstrom, C. Theuring, R. Lin- digkeit & M. Rahier. 2003. Detoxification of pyr- rolizidine alkaloids by the harvestman Mitopus morio (Phaiangidae) a predator of alkaloid de- fended leaf beetles. Chemoecoiogy 13:123-127. Head, G. 1995. Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (Class Araneae). Evolution 49: 776-781. Kyneb, A. & S. Toft. 2004. Quality of two aphid species (Rhopalosiphum pad! and Sitobion av- enae) as food for the staphylinid generalist pred- ator Tachyporus hypnorum. Journal of Applied Entomology 128:658-663. Lowrie, D.C. 1987. Effects of diet on the devel- opment of Loxosceles iaeta (Nicolet) (Araneae, Loxoscelidae). The Journal of Arachnology 15; 303-308. Madsen, M., S. Terkildsen & S. Toft. 2004. Micro- cosm studies on control of aphids by generalist arthropod predators: Effects of alternative prey. BioControl 49:483-504. Marcussen, B.M., J.A. Axelsen & S. Toft. 1999. The value of two Collembola species as food for a linyphiid spider. Entomologia Experimentalis et Applicata 92:29-36. Martens, J. 1978, Weberknechte, Opiliones. Die Tierwelt Deutschlands 64. Fischer, Jena. Mayntz, D. & S. Toft. 2001. Nutrient composition of the prey’s diet affects growth and survivorship of a generalist predator. Oecologia 127:207-213. Meinertz, N.T. 1964. Der Jahreszyklus der Danisch- en Opilioniden. Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening i K0benhavn 126: 451-464. Moran, M.D. 2003. Arguments for rejecting the se- quential Bonferroni in ecological studies. Oikos 100:403-405. Pfeifer, H. 1956. Zur okologie und larvalsystematik der weberknechte. Mitteiluegen aus dem Zoolo- gischen Museum in Berlin 32:59-104, Phillipson, J. 1962. Respirometry and the study of energy turnover in natural systems with particu- lar reference to harvestspiders (Phalangiida). Oi- kos 13:311-322. Pyke, D.A. & J.N. Thompson. 1986. Statistical analysis of survival and removal rate experi- ments. Ecology 67:240-245. Sankey, J.H.P. 1949. Observations on food, enemies and parasites of British harvest-spiders (Arach- nida, Opiliones). The Entomologist’s Monthly Magazine 85:246-247. Sankey, J.H.P. & T.H. Savory. 1974. British har- vestmen. Synopses of the British Fauna: 4. Lin- nean Society of London. Symondson, W.O.C., K.D. Sunderland & M.H. Greenstone. 2002. Can generalist predators be ef- fective biocontrol agents? Annual Review of En- tomology 47:561-594. Todd, V. 1949. The habits and ecology of the Brit- ish harvestmen (Arachnida, Opiliones), with spe- cial reference to those of the Oxford District. Journal of Animal Ecology 63:209-229. Toft, S. 1995. Value of the aphid Rhopalosiphum padi as food for cereal spiders. Journal of Ap- plied Ecology 32:552-560. Toft, S. 2000. Species and age effects in the value of cereal aphids as food for a spider (Araneae). Ekologia (Bratislava) 19:273-278. Toft, S. & D.H. Wise. 1999a. Growth, development, and survival of a generalist predator fed single- 590 THE JOURNAL OF ARACHNOLOGY and mixed-species diets of different quality. Oec- ologia 119:191-197. Toft, S. & D.H. Wise. 1999b. Behavioral and eco- physiological responses of a generalist predator to single- and mixed-species diets of different quality. Oecologia 119:198-207. Uetz, G.W., J. Bischoff & J. Raver. 1992. Survi- vorship of wolf spiders (Lycosidae) reared on different diets. The Journal of Arachnology 20: 207-211. Vanacker, D., K. Deroose, L. van Nieuwenhuyse, V. Vandomme, J. Vanden Borre & J.-R Maelfait. In press. The springtail Sinella curviseta: the most suitable prey for rearing dwarf spiders. Arthrop- oda Selecta. Wiktelius, S. 1982. Flight phenology of cereal aphids and possibilities of using suction trap catches as an aid in forecasting outbreaks. Swed- ish Journal of Agricultural Research 12:9-16. Manuscript received 16 November 2004, revised 15 June 2005. 2005. The Journal of Arachnoiogy 33:591-596 CHROMOSOMAL DATA OF TWO PHOLCIDS (ARANEAE, HAPLOGYNAE): A NEW DIPLOID NUMBER AND THE FIRST CYTOGENETICAL RECORD FOR THE NEW WORLD CLADE Douglas de Araujo^^ Antonio Domingos Brescovit^^ Cristina Anne Rheims^’^ and Doralice Maria Cella^: Universidade Estadual Paulista — UNESP, lestituto de Biociencias, Departamento de Biologia, Av. 24- A, 1515, CEP: 13506-900, Bela Vista, Rio Claro, SP, Brazil. E-mail: daraujo@rc.unesp.br; E-mail: dmcella@rc.unesp.br; ^Institute Butantan, Laboratorio de Artropodes Pe^onheetos, Av. Vital Brasil, 1500, CER: 05530-900, Sao Paulo, SP, Brazil; ^Departamento de Zoologia, Instituto de Biociencias, Universidade de Sao Paulo, Sao Paulo, SP, Brazil ABSTRACT. Mesabolivar luteus (Keyserling 1891) and Micropholcus fauroti (Simon 1887) specimens were collected in Ubatuba and Rio Claro, both in the state of Sao Paulo, Brazil. Mesabolivar luteus showed 2n (d) = 15 = 14 + X and 2n (?) = 16 = 14 + XX in mitotic metaphases and 7II + X in diplotenic cells. During late prophase I, all bivalents presented a ring shape, evidencing two chiasmata per bivalent. In this species, some diplotenic cells appear in pairs, maybe due to specific characteristics of the intercellular bridges. The metaphases II showed n = 7orn = 8 = 7 + X chromosomes. Micro- pholcus fauroti evidenced 2n (d) = 17 = 16 + X in spermatogomal metaphases and 8II+X in diplotenic cells, with only one chiasma per bivalent, contrasting with M luteus. In both species, all chromosomes were metacentrics. The sexual chromosome X was the largest element and appeared as a univalent during meiosis 1. These are the first cytogenetical data for the genera Mesabolivar and Micropholcus. Additionally, M. luteus is the first chromosomally analyzed species of the New World clade and the observed diploid number for M. fauroti had not yet been recorded in Pholcidae. Keywords: Arachnida, Meiosis, chromosomal morphology, spider, diplotene pair Pholcids are small to medium sized spiders, with total length ranging from 1-1 5mm, usu- ally with six or eight eyes, and legs several times longer than the body length. Specimens are found in low and high elevations, forests and deserts, leaf litter and tree canopies. There are several synanthropic species with cosmo- politan distribution. These characteristics tak- en together make the family Pholcidae Koch 1851 the most diverse among the haplogyne group, comprising 75 extant genera and 866 extant species (Huber 2000, 2005). According to the cladogram proposed as a working hypothesis by Huber (2000) for the New World pholcids, the family is strongly supported as a monophyietic group and is di- vided into four clades: “ninetines”, “pholci- nes” {Metagonia Simon 1893 + Pholcus group sensu Huber, 1995), “holocnemines” (Holocnemus group sensu Timm, 1976 + Ar- tema Walckenaer 1837 + Physocyclus Simon 1893 + Priscula Simon 1893) and “New World clade”. The latter includes most of the genera and is the only one that is exclusive for the New World. However, Huber (2000) himself pointed to “ninetiees” and “holoc- nemines” as questionable monophyietic groups. Despite the high number of Pholcidae spe- cies, only nine species (1%) of five genera have been chromosomally analyzed, i.e., “pholciees”: Pholcus crypticolens Boseeberg & Strand 1906, 2n (6) - 24 = 22 + X1X2 (Suzuki 1954); Pholcus manueli Gertsch 1937 (under Pholcus qffinis Scheekel 1953), 2e (A) - 25 - 24 + X (Wang et al. 1997); Pholcus phalangioides (Fuesslin 1775), 2e (A) = 24 = 22 + X1X2 (Rodriguez-Gil et al, 2000) and Spermophora senoculata (Duges 1836) (under Spermophora meridionalis Heetz 1841, mis- spelled as Spermaphora meridionalis), 2e (A) = ? -? + X1X2 (Painter 1914), and “holoc- eemines”: Artema atlanta Walckenaer 1837 (misspelled as Artema atlenta), 2ri (A) = 32 591 592 THE JOURNAL OF ARACHNOLOGY = 30 + X1X2 (Parida & Sharma 1987; Sharma & Parida 1987); Crossopriza lyoni (Blackwall 1867), 2ii (d) - 27 - 26 + X (Bole^Gowda 1958), 2n (d) = 25 = 24 + X (Srivastava & Shukla 1986), 2n (d) = 24 = 22 + X1X2 (Sharma et aL 1959) and 2n (d) = 23 = 22 + X (Parida & Sharma 1987; Sharma & Par- ida 1987); Physocyclus californicus Chamber- lin & Gertsch 1929, 2n (d) “ 15 ^ 14 +X (Cokeedolpher 1989); Physocyclus enaulus Crosby 1926, 2n (d) - 15 - 14 + X (Cok- endolpher 1989) and Physocyclus sp., 2n(d) ==15 = 14 + X (Cokeedolpher & Brown 1985; Cokeedolpher 1989). In the species whose chromosomal morphology has been de- termined, all chromosomes are metacentric, with the exception of the X, and X2 chromo- somes of C lyoni described by Sharma et ah (1959) and P. crypticolens, which are acro- centric. There are no cytogenetical data on “ninetines” and “New World clade’k The genus MesaboUvar Gonzalez-Sponga 1998, included in the New World clade by Huber (2000), includes 34 species from which 24 occur in Brazil (Huber 2005). This genus arises as a sister group of Coryssocnemis Si- mon 1893; however, this position is not yet clearly established. The genus MesaboUvar has been divided into four “operationaT’ groups, based on morphological characters: a “northern group with spines on male metatar- si” (5 species), a “northern group without spines on male metatarsi” (6 species), a “southern/eastern group” (15 species) proba- bly not monophyletic, and a “miscellaneous group” (7 species), certainly polyphyletic, that will probably be partly transferred to oth- er geeera/group (Huber 2000). MesaboUvar luteus (Keyserling 1891) is a species belonging to the “miscellaneous group,” probably related to MesaboUvar levii Huber 2000, and is distributed in the states of Rio de Janeiro, Sao Paulo, Parana and Rio Grande do Sul, in Brazil. The genus Micro- pholcus Deeleman-Reinhold & Prinsen 1987 (pholcine) includes only two species, of which only the Pantropical species Micropholcus fauroti (Simon 1887) occurs in Brazil by in- troduction and lives as a synanthropic species (Huber 2000). The use of chromosomal data in phyloge- netic analysis is relatively new, and the cri- teria to codify these data are controversial (Modi 1987; Borowik 1995). Additionally, cy- togenetic analysis may have some difficulties when compared with other kinds of analysis: the specimens must be kept alive until the slide preparations, some of them do not have f cell division at the moment of analysis, and some techniques are expensive. Nevertheless, : chromosomal data have a potential usefulness ‘ for phylogenetic inference, because they are | heritable, homologue states can be identified, and the characters are independent from each other (Borowik 1995). Basically, a chromo- somal phylogeny can be constructed based on the minimum number of rearrangements re- quired or the maximum number of shared seg- ments (Rokas & Holland 2000). Although ' chromosomal data has not been used for cla- distic analysis in spiders, there have been some attempts in other groups, such as mam- mals, to obtain characters by conventional (Nagamachi et al. 1999; Garcia et al, 2000) or molecular cytogenetic techniques (Oliveira et ah 2002). The aim of this study is to characterize the ' chromosomes of the species M, luteus and M. fauroti, analyzing standard stained mitotic and meiotic cells, in order to begin an effort to establish karyotypic relationships among spe- cies in the Pholcidae. METHODS Three males and one female of M. luteus \ were collected at Maranduba beach, Ubatuba j (23°43'S 45W'W), and five males of M. fau- \ roti were collected in buildings in Rio Claro i and Ubatuba (22°4rs 47°56'W and 23°43'S * 45°07'W), both in the state of Sao Paulo, ! southeastern Brazil. The specimens are depos- ited in the collection of the Laboratorio de Ar- tropodes Pegonhentos, Instituto Butantan, Sao j Paulo (IBSP, A.D. Brescovit) under the num- ; bers IBSP 42785 {MesaboUvar luteus), 42782, 42783, 42784, 47504 and 47505 {MicrophoU cus fauroti). \ Gonads were dissected in Ringers solution for insects, transferred to colchicine solution (0.16% in Ringer for insects) and left for 2 hrs.; a volume of hypotonic solution (tap wa- ter) equal to that of the colchicine solution was added and after 15 mins, the material was placed in Carnoy I fixative solution for 60 min., after which it was macerated in 60% ' acetic acid on the surface of the slide. The : slide was dried on a metal heating plate (35- 40 °C) and stained with a 3% Giemsa solution ARAUJO ET AL.— CHROMOSOMAL DATA OF TWO PHOLCIDS 593 Figures 1-6. — MesaboUvar luteus cells. 1. Spermatogonial metaphase, with 2n = 15 = 14 + X. 2. Oogonial metaphase, with 2n = 16 = 14 + XX. The asterisks indicate overlapped chromosomes. 3, Diplotene, with 7II + X, Arrows indicate the chiasma location, 4. Dipiotene nuclei, constituting a pair of cells. 5. Metaphase II, with n = 8 = 7 + X. The X could not be identified in this spread. 6. Metaphase II, with n = 7. Scale =10 fxm. for 13-15 min. The cells were photographed under a Zeiss microscope and the chromo- some morphology classification was deter- mined according to Levan et al. (1964). The number of analyzed chromosomal spreads was 65 for M. luteus and 40 for M. fauroti. In each of these spreads, the chromosome number was determined and no intraspecific variation was detected. RESULTS MesaboUvar luteus,— The mitotic meta- phases showed 2n= 15" 14 + Xin males (Fig. 1) and 2e~ 16™ 14 + XX in females 594 THE JOURNAL OF ARACHNOLOGY Figures 7-8. — Micropholcus fauroti cells. 7. Spermatogonial metaphase, with 2n = 17 = 16 -h X. 8. Diplotene, with 8II + X. Arrow indicates a terminal chiasma and arrowhead points to an interstitial chiasma. Scale =10 pm. (Fig. 2). In the spermatogonial metaphases, the X chromosome is always easily identified as the largest element (Fig. 1). The chromo- somal morphology is not clear in the mitotic metaphases due to the low degree of chro- mosome condensation. Diplotene cells showed 7II + X (Fig. 3). All bivalents present a ring shape, evidencing the occurrence of two terminal chiasmata per bivalent, and the X chromosome constitutes an univalent during all meiosis I (Fig. 3). Some diplotene cells appeared in pairs (Fig. 4). Metaphases II showed n = 8 = 7 + X (Fig. 5) or n = 7 (Fig. 6) chromosomes. The X chromosome cannot be recognized in the n = 8 cells due to the irregular chromosome appearance. De- spite the low staining contrast, the chromo- somal morphology of this species was deter- mined as metacentric. Micropholcus faurotL—Th^ spermatogo- nial metaphases showed 2n= 17 = 16 + X (Fig. 7). The largest chromosome of comple- ment is X, which is easily identified in all an- alyzed metaphases (Fig. 7). Despite the low staining contrast and the low morphology res- olution, the chromosomes seem to be biarmed (Fig. 7). Diplotene cells possessed 8II + X (Fig. 8) and each bivalent shows only one chi- asma, terminal or interstitial (Fig. 8). The X chromosome appears as a univalent during meiosis I (Fig. 8). DISCUSSION Despite high diversity of pholcid species among haplogynes, this family is poorly known from the cytogenetic point of view. This could be due to the lack of Pholcidae cytogenetic researchers, the relatively small size of pholcid species and their chromo- somes, and the difficulty in obtaining good quality chromosomal preparations. As the generic name suggests, Microphol- \ cus fauroti is a very small spider, l-2mm in length. Thus, dissection of the specimens, as ; well as the removal of the testis, is very dif- ficult, Additionally, only one slide, with few cells, can be obtained per specimen due to ex- ! tremely minute size of the testis. In relation to the chromosome length, Paint- ^ er (1914), Suzuki (1954) and Bole-Gowda (1958) emphasized the very small size of the ; elements. The largest chromosome of P. cryp- \ ticolens, obtained by Suzuki (1954), measured only around 2.4 pm. The largest chromosome ‘ of C lyoni is the X chromosome, which mea- sures 5.8 pm, but the largest autosome mea- sures only around 2.3 pm (Bole-Gowda 1958). The measurements of the largest chro- mosomes of M. luteus and M. fauroti were respectively 9 and 7 pm (for the X chromo- some), and 6 and 5 pm (for the autosomes). Thus, the chromosomes of the studied species ■ are not as small as those obtained by Suzuki (1954) and Bole-Gowda (1958). On the other . hand, they are not as large as those of other haplogyne genera, such as Loxosceles Heine- „ ken & Lowe 1832 (Sicariidae) in which the largest chromosomes measure around 15 pm (Silva et ah 2002). j Concerning the preparation quality. Painter ' (1914) and Suzuki (1954), using different types of fixative solutions, called attention to |; the unfavorable fixation of pholcid chromo- i ARAUJO ET AL.— CHROMOSOMAL DATA OF TWO PHOLCIDS 595 somes. A similar problem occurred with M. luteus and M. fauroti chromosomes, when Carnoy I fixative solution was used, resulting in low staining contrasts. Alternative fixation methods should be tested in pholcid species. MesaboUvar luteus is the first cytogeneti- cally studied species from the “New World clade” and showed a diploid number equal to that found in three Physocyclus species (hoi- oceemiiies) analyzed by Cokeedolpher (1989), despite the fact that these genera be- long to different clades. Thus, the 2n = 15 could have arisen iedepeedeetly at least two times within the pholcids. Micropholcus fau- roti is the first cytogenetically analyzed spe- cies from this genus and until now, its diploid number had not yet been recorded in Pholci- dae. The presence of biamied chromosomes in both species of this study is a feature shared among most of the haplogyne group species, as stated by Rodriguez-Gil et aL (2002). In both species, the largest chromosome of the complement is the X chromosome. This is in agreement with the data obtained by Bole- Gowda (1958) for C lyoni and by CokeiidoL pher (1989) for three Physocyclus species. During ieterphase, the observed X chromatin positive heteropycnosis of M. fauroti is simi- lar to that recorded for C. lyoni by Bole-Gow- da (1958). The studied species showed significant dif- ferences from each other in relation to the chi- asrna number, during meiosis. However, in- formatioe on chiasma iiumbei and position was not provided by previous papers on phol- cid cytogenetics. Thus, these characteristics cannot be used as parameters to compare re- lated species or to establish a pattern within the pholcid groups. The diploteee pairs found in. M. luteus are probably a consequence of the germ cell ar- rangement and interaction, which constitute “cysts” with synchronously dividing cell con- nections via intercellular bridges due to the lack of cytokinesis during speiinatogenesis. Alberti & Weiemaee (1985) described the presence of similar cysts in the testis of P. phalangiodes. In relation to these grouped cells, two ques- tions are crucial: why do they appear in pairs and not in largei groups of cells; and why do these pairs only appear in the diplotene phase? Concerning the first question, Pepling & Spra- dling (1998) have verified a tendency towards the increase in number by the power of two in mouse embryo oogoeial mitotic cells, being more frequently found in clusters of two cells. Clusters with more cells are probably more susceptible to breaks during slide preparation. However, the possibility of finding such clus- ters in future analysis cannot be discarded. With respect to the second question, this fea- ture is probably a consequence of the skewed cellular phase ratio in the sample, because from the 105 spreads obtained, only 9 were mitotic metaphases and the others were almost all diplotenes. Possibly, paired mitotic meta- phases should be also found in MesaboUvar luteus. An ultrastmctural analysis of sper- matogenesis would be of interest to answer these questions. Additionally, further analysis of other pholcid species is needed to verify whether this pairing of ceils also occurs. The possibility of the occurrence of poly- ploidy in M. luteus was discarded, at least in the first instance, dee to two main reasons: the lack of polyploid metaphase II cells (despite the low frequency of cells in this meiotic stage) and the lack of tetravaleets or chro- mosomal chains at meiosis L The formation of chr’omosomal chains is not a strict rale in polyploids, but they are frequently observed (John 1990). The cytogenetic analysis of the pholciees Leptopholcus Simon 1893 and Metagonia Si- moe 1893, and of the holocnemines Smerin- gopus Simon 1890, Holocnemus Simon 1873 and Priscula Simon 1893 seems to be ex- tremely important to establish the karyotypic evolution in these two clades. The cytogenet- ical study of the nieetiees and of the New World clade requires more exhaustive re- search, considering that only MesaboUvar was analyzed and that there are numerous genera belonging to these two clades. Finally, when a full cytogenetical data set becomes available for Pholcidae, it could be used to improve the proposed phylogenetic hypothesis for the family. ACKN’OWLEDGMENTS We wish to thank Dr. Bernhard A. Huber, from the Zoological Institute and Museum Al- exander Koenig, Germany, for the bibliogra- phy assistance. Financial support was provid- ed by CAPES and FAPESP (99/05446-8). 596 THE JOURNAL OF ARACHNOLOGY LITERATURE CITED Alberti, G, & C. Weinmann. 1985. Fine structure of spermatozoa of some labidognath spiders (Filis- tatidae, Segestriidae, Dysderidae, Oonopidae, Scytodidae, Pholcidae; Araneae; Arachnida) with remarks of spermiogenesis. Journal of Morphol- ogy 185:1-35. Bole-Gowda, B.N. 1958. A study of the chromo- somes during meiosis in twenty-two species of Indian spiders. Proceedings of the Zoological So- ciety of Bengal 11:69-108. Borowik, O.A. 1995. Coding chromosomal data for phylogenetic analysis: phylogenetic resolution of the Pan-Homo-Gorilia trichotomy. Systematic Biology 44:563-570. Cokendolpher, J.C. 1989. Karyotypes of three spi- der species (Araneae: Pholcidae: Physocyclus). Journal of the New York Entomological Society 97:475-478. Cokendolpher, J.C, & J.D. Brown. 1985. Air-dry method for studying chromosomes of insects and arachnids. Entomological News 96:114-118. Garcia, L., M. Ponsa, J. Egozcue & M. Garcia. 2000. Comparative chromosomal analysis and phylogeny in four Ctenomys species (Rodentia, Octodontidae). Biological Journal of the Linnean Society 69:103-120. Huber, B.A. 1995. Copulatory mechanism in HoU ocnemus pluchei and Pholcus opilionoides, with notes on male cheliceral apophyses and stridu- latory organs in Pholcidae (Araneae). Acta Zool- ogica (Stockholm) 76:291-300. Huber, B.A. 2000. New World pholcid spiders (Ar- aneae: Pholcidae): a revision at generic level. Bulletin of the American Museum of Natural History 254:1-348. Huber, B.A. 2005. Catalogue of Pholcidae. Zoolog- ical Research Institute and Museum Alexander Koenig, on-line at http://b.a,huber.bei, t-online.de/ homepage. John, B. 1990. Meiosis. Cambridge University Press.Pp. 79-85. Levan, A., K. Fredga & A. A. Sandberg. 1964, No- menclature for centromeric position on chromo- somes. Hereditas 52:201-220, Modi, W.S. 1987. Phylogenetic analyses of chro- mosomal banding patterns among the neartic Ar- vicolidae (Mammalia: Rodentia). Systematic Zo- ology 36:109-136. Nagamachi, C.Y., J.C. Pieczarka, J.A.P.C. Muniz, R.M.S. Barros & M.S. Mattevi, 1999. Proposed chromosomal phylogeny for the South American primates of the Callitrichidae Family (Platyrrhi- ni). American Journal of Primatology 49:133- 152. Oliveira, E.H.C,, M. Neusser, W.B, Figueiredo, C. Nagamachi, J.C. Pieczarka, I.J. Sbalqueiro, J. Wienberg & S. Muller, 2002. The phylogeny of howler monkeys (Alouatta, Platyrrhini): recon- struction by multicolor cross-species chromosome painting. Chromosome Research 10:669-683. Painter, T.S. 1914, Spermatogenesis in spiders. Zoologische Jahrbuecher Abteilung fuer Anato- !■ mie und Ontogenie der Tiere 38:509-576. j Parida, B.B. & N.N. Sharma. 1987. Chromosome l! number, sex mechanism and genome size in 27 species of Indian spiders. Chromosome Infor- mation Service 43:11-13. Pepling, M.E. & A.C. Spradling. 1998. Female I mouse germ cells form synchronously dividing l- cysts. Development 125:3323-3328. , Rodrfguez-Gil, S.G., L.M. Mola, A.G. Papeschi & I; C.L. Scioscia. 2000. Cytogenetic heterogeneity | in common argentine spiders. XXI International ' Congress of Entomology 1:584. i Rodrfguez-Gil, S.G., L.M. Mola, A.G. Papeschi & i C.L. Scioscia. 2002. Cytogenetic heterogeneity j in common haplogyne spiders from Argentina i’ (Arachnida, Araneae). Journal of Arachnology I 30:47-56. Rokas, A. & P.W.H. Holland. 2000. Rare genomic changes as a tool for phylogenetics. Tree 15: 454-459. Sharma, G.P., B.L. Gupta & R. Parshad. 1959. Cy- tological studies on the Indian spiders. III. An i' analysis of the chromosomes in the male germ cells of the spider, Crossopriza lyoni (Blackwall), fam. Pholcidae. Research Bulletin (N.S.) of the Panjab University 10:49-53. Sharma, N. & B.B. Parida. 1987. Study of chro- mosomes in spiders from Orissa, Pranikee 8:71- I 76. : Silva, R.W., D.R. Klisiowicz, D.M. Celia, O.C. i Mangili & I.J. Sbalqueiro. 2002. Differential dis- tribution of constitutive heterochromatin in two . species of brown spider: Loxosceles intermedia ' and L. laeta (Araneae, Sicariidae), from the met- ropolitan region of Curitiba, PR (Brazil). Acta Biologica Paranaense 31:123-136. [ Srivastava, M.D.L. & S. Shukla. 1986. Chromo- ! some number and sex-determining mechanism in forty-seven species of Indian spiders. Chromo- ' some Information Service 41:23—26, Suzuki, S. 1954. Cytological studies in spiders. III. ! Studies on the chromosomes of fifty-seven spe- cies of spiders belonging to seventeen families, with general considerations on chromosomal ' evolution. Journal of Science of the Hiroshima ' University. Series B. Division 1 15:23-136. Timm, H. 1976. Die Bedeutung von Genitalstmk- turen fiir die Klarung systematischer Fragen bei Zitterspinnen (Arachnida: Araneae: Pholcidae). j Entomologica Germanica 3:69-76. Tres, L.L., E. Rivkin & A.L. Kierszenbaum. 1996. Sak 57, an intermediate filament keratin present in intercellular bridges of rat primary spermato- cytes. Molecular Reproduction and Development 45:93-105. [ Wang, X., S. Cui, Z. Yang, J. Wang & Y. Wang. ' 1997. On karyotype of the Pholcus affinis (Ar- aneide: Pholcidae). Acta Arachnologica Sinica 6: 19-22. Manuscript received 10 December 2004, revised 5 July 2005. 2005. The Journal of Arachnology 33:597-603 SIX STRIDULATING ORGANS ON ONE SPIDER (ARANEAE, ZODARIIDAE): IS THIS THE LIMIT? Rudy Jocque: Royal Museum for Central Africa, Leuvensesteenweg 13, B-3080 Tervuren, BELGIUM. E-mail: rudy.jocque@africamuseum.be ABSTRACT. A new type of stridulatory organ is described and figured occurring in three species of Mallineila Strand from Thailand and Singapore. In one species there are four stridulatory organs, with the ridges on femora I and II and the pegs in the shape of granulations on femora II and III. In both the other species an additional pair occurs, with ridges on femora III and pegs on femora IV. To date no more than four stridulatory organs have been recorded on a single spider. Examples of various known forms of stridulatory organs on spiders are illustrated and their significance briefly discussed. Keywords: Stridulation, Thailand, Singapore, courtship, mate check Stridulating organs are manifold in spiders and were reported for at least 22 families of spiders in Uetz & Stratton (1982) and since then, several cases in other families (Corin- nidae, Tetragnathidae, Zodariidae, see below) have been mentioned. Some of the organs are single, but paired stridulatory organs appear to occur more commonly. These organs in- evitably comprise two elements: the “pars stridens”, a sclerotized area provided with a series of ridges referred to as “the file” or simply “the ridge” in those cases where there is only one, and the “plectron” which may be one or a series of stiff setae or pegs, some- times called “the plectrum” or “the scraper” in the case of a single peg. In single stridulatory organs, sounds are produced by rubbing the front of the abdomen against the rear of the cephalothorax, the sur- faces of which are provided either with ridges, pegs or stiff setae (Figs. 1-3) (e.g. Maddison & Stratton 1988a, b). The same applies to the sound produced by files on the inner surface of the chelicerae that occurs in Mygalomor- phae or on the inner surface of the anterior lateral spinnerets as found in some Theridi- idae (Forster et al. 1990; Agnarsson 2004). Paired stridulatory organs, usually in the form of ridges and pegs, occur in many other taxa. Most commonly, these ridges occur on the chelicerae (externally), and are rubbed by the palps; on the booklung covers, rubbed by the fourth legs; or on the coxae of the first legs rubbed by a peg on the second trochanters (e.g. Hinton & Wilson 1970). Rovner (1975) recorded a stridulatory device in males of Ly- cosa and Schizocosa: a plectrum on the male palpal tibia rubs against a file on the cym- bium. Edwards (1982) later found a similar device in the salticid Phidippus mystaceus (Hentz). Legendre (1963) and Uetz & Stratton (1982) provided a fairly complete overview of the different types of stridulatory organs, sum- marized here in Figs. 1-10. Starck (1985) gave a complete list of the known stridulatory organs, analyzed the structure of the elements that compose a stridulating organ and dis- cussed the function and the evolutionary as- pects of the devices. He stressed the homo- plasy of these structures in different taxa and concluded that there has been parallel devel- opment of similar organs, even within the same family. Since these early papers, several more types of stridulatory organs have been described. Maddison (1987) found Marchena minuta (Peckham & Peckham 1888) and other jump- ing spiders (Salticidae) to be provided with ridges or a row of stout setae on the dorsal base of the femora I combined with respec- tively a row of setae or a stridulatory file on the carapace just under the eyes. Simon (1937) was apparently aware of this structure and mentions the femoral tubercles in Icius Simon. Wunderlich (1995) reported on an ex- ternal longitudinal ridge on the chelicerae in Zygiometella Wunderlich (Tetragnathidae), supposedly combined with setae on the inner side of the male palp to form a stridulatory organ. Ramirez et al. (2001) found a similar 597 598 THE JOURNAL OF ARACHNOLOGY Figures 1-3. — Examples of spiders with one stridulatory organ (appendages omitted). 1. Steato- dci Siindevall (Theridiidae), with file on carapace, pegs on abdomen; 2. Cambhdgea L. Koch (Stiphi- diidae), with file on abdomen, pegs on pedicel (dor- sally); 3. Cambhdgea L. Koch (Stiphidiidae), with file on abdomen, pegs on pedicel (ventrally) (after Legendre 1963 and Uetz & Stratton 1982). Dotted arrows = files and ridges, solid arrows = pegs. stridulating system in Olbiis Simon 1880 (Corinnidae): a retrolateral ridge on femur IV corresponding with a field of modified seta ba- ses on the abdomen. This organ combines with a field of prolateral setae with modified bases on the prolateral side of the same femur opposed to a field of evenly spaced setae with transversely arranged bases. If both these combinations represent stridulating organs, this was the first case reported of four such organs in a spider. Only one putative stridulating organ was so far reported in the Zodariidae: the species Ak- yttara homunculus (Jocque 1991) has warts on the anterior surface of the abdomen corre- sponding with ridges on the posterior part of the carapace (Jocque 1 99 1 ). The present paper reports on a remarkable case of multiple stridulatory organs and pro- vides a concise overview of the present knowledge on these structures. METHODS Stridulating organs on the femora of Mal- lineila Strand 1906 (Zodariidae) were noted for the first time while sorting through collec- ! tions of representatives of the family from j; Thailand collected by R Schwendinger. A dark I area around the stridulating file of the species ' with four stridulating organs made that region " conspicuous. Without the color contrast, the structures would probably have passed unno- ticed. Stridulatory structures were found on the males of two species; each species was represented by only one male. A more inter- esting example was found in the collection of J. Murphy. This collection included both sexes of what appeared to be M. cinctipes (Simon 1 892) according to the drawing of the epigyne in Workman ( 1 896) and a photo of the spider by Koh (1989). All the examined specimens belong to the palaeotropical genus MaUmeUci that has a vast distribution from West Africa to northern Aus- tralia (Jocque 1993). They are typical forest ; soil-dwellers and compulsory termite feeders that hide in silk-lined spherical buried retreats during daytime. ! Specimens examined. — Mallinella sp. 1: 1 6 (with four stridulatory organs), Thailand, Penang Hill, 150-330 m, 02.xii.l991 (P. , Schwendinger). Mallinella sp. 2: 1 d (with six ' stridulatory organs), Thailand, Doi Chiang Dao, 510 m, 25.x-23.xi. 1990 (P. Schwendin- ger). Mallinella cinctipes: 1 S : Singapore, Up- per Pierce Reservoir, iii.1986 (Murphy collec- tion 13418); Id: Singapore, Upper Pierce Reservoir, ii.l988 (Muiphy collection 15443); 3d, 1$: Singapore, Bukit Timah, ii.l988 (Murphy collection 15471). The voucher specimens of the unknown species from Thai- land shall be deposited in the Musee d’Histoire Naturelle de Geneve, Switzerland. Males were preserved in ethanol 75% in the field and examined in the lab. The male specimen from Penang Hill was scanned using a XL30 ESEM scanning elec- tron microscope in wet mode with cooling cell that leaves the specimen undamaged. Images (Fig. 13) were taken of the entire specimen at different depths and composed into a single photomontage by using an analogue camera and composition software. (Automontage of JOCQUE— MULTIPLE STRIDULATION ORGANS IN SPIDERS 599 Figures 4-7. — Examples of spiders with two stridulatory organs (appendages omitted). 4. Linyphiidae, Hahniidae (and many other families), with file on chelicerae, pegs on palps; 5-. Linyphiidae, with file on abdomen (ventrally), pegs on leg IV; 6. Linyphiidae, with file on palps, pegs on coxae I; 7. Linyphiidae, with file on coxae I, pegs on trochanters II (after Legendre 1963). Synoptics). Other SEM images were taken with a JEOL 6480LV. RESULTS Mallinella sp. 1 (Eigs. 11, 13) appears to have four and Mallinella sp. 2 and Mallinella cinctipes (Eig. 12) appear to have six stridu- latory organs. In the latter, the ridges are sit- uated on a conspicuous swelling (Eigs. 15, 16) of the dorsal base of the anterior femora and are apparently rubbed by prolateral ventral granulations (Eig. 18) at the base of tiny setae on the following femur. In Mallinella sp. 1, the file area is rounded and has a diameter of 0.45 mm, with 48 ridges which means that the ridges are slightly less than 0.01 mm apart. The granulations are 0.045 mm apart. In Mal- linella sp. 2 the ridges are somewhat thinner (52 in an area with diameter 0.41 mm) and the granulations more densely set at a distance of just under 0.01 mm. In Mallinella cinctipes the stridulation area on Ee II is on average 0.34 mm across and has 72 ridges which means that they are again thinner and about half as far apart as in the first species (5 pim). The granulations are between 0.04 and 0.06 mm apart. The female of M. cinctipes has not the slightest indication of a femoral stridula- tion organ. DISCUSSION To date, no spider species with more than four stridulatory organs has been reported. The only probable case is that of Olbus jaguar Ramirez et al. 2001 mentioned in the intro- duction, As far as I am aware, no cases exist in which a single central stridulatory organ is combined with a symmetrical double organ. The number of stridulatory organs on a single spider is now known to be 1, 2, 4 or 6. The last two mentioned cases are especially sur- prising, as even in the case of other animals, such a high number has apparently never been recorded. The organ depicted here bears some 600 THE JOURNAL OF ARACHNOLOGY Figures 8-10.— 8. Lycosidae, file on cymbium, peg on palpal tibia (after Rovner, 1975); 9. Salti- cidae, with file on side of carapace under the eyes, pegs on inner side of first femur (after Maddison, 1987). Dotted arrows = files and ridges, solid ar- rows = pegs; 10. Example of a spider with four stridulating organs and the only one known with a mixed set-up: Olbus jaguar (Corinnidae). One sys- tem (A) consists of a file on femur IV and pegs on femur III, the other one (B) of a file on the abdomen and pegs on femur IV (after Ramirez et al. 2001). Dotted arrows = files and ridges, solid arrows = pegs. resemblance to those mentioned in Olbus and in Marchena Peckham & Peckham described by Maddison (1987) but in the zodariids the carapace and the abdomen are smooth and de- void of setae or ridges. In the Mallinella, the hie is dorsolateral (Figs. 11, 12, 14-16) and directed towards the following femur with the granulations that apparently function as pegs. This is corroborated by the fact that the seta bases on the hrst femur (Fig. 17) have no granulate extensions whereas those of femora II, III and in some cases IV do (Fig. 18). Stridulation may have two clearly different functions, i.e. defense and courtship (Starck 1985; Uhl & Schmitt 1996). Although there are no studies available for spiders with more than two stridulatory organs, there is no rea- son to expect that the function should be dif- ferent for multiple organs. Defense stridulation in larger animals, such as mygalomorphs (Legendre 1963), is often audible to the human ear. However, in araneo- morph spiders, stridulation probably originat- ed as part of courtship (Starck 1985). Several hypotheses have been formulated regarding the function of this stridulation during court- ship. These include mate recognition, antag- onistic behaviour between males (Gwinner- Hanke 1970; Maddison & Stratton 1988a), stimulation of the female by the male (Eber- hard 1996) and information transfer (Jocque 1998). It is difficult to accept that more than one stridulating organ would be needed if the aim is to recognize the partner: the possibili- ties for variation with one “instrument” are endless and it is therefore unlikely that mul- tiple stridulation organs are developed for that purpose. Stimulation of the partner is another possibility that has been invoked to explain the development of secondary sexual organs. The question always remains why species with similar morphology and life style would evolve such different degrees of partner stim- ulation. “Mate check,” (Jocque 1998, 2002) on the other hand, assumes that the quantity of in- formation transferred during courtship is di- rectly related to the ecological specialization of the species. Via an array of signals, com- bined in a so-called “mating module” (Jocque 2001, 2002), the presence of crucial adapta- tions in the male mate is verified during court- ship and mating. In the speciose genus Mal- linella, species of which have a highly specialized biology, the development of a complex stridulatory apparatus as part of the mating module is a plausible explanation cer- tainly because the females appear to be devoid of such organs. Another fascinating question to be an- swered is how these stridulatory organs are operated. It is very unlikely and physically ap- parently impossible that they are all activated at the same time in an orchestra-like manner. It is therefore to be expected that the organs are activated consecutively and in pairs, one on either side of the animal. This prompts the question whether spiders with more than six stridulating organs can be expected. If the or- gans are arranged as in Olbus jaguar (see JOCQUE— MULTIPLE STRIDULATION ORGANS IN SPIDERS 601 Figures 11-12.- — Example of a spicier with four and six stridulatory organs. 11. Mallinella sp. 1 (Zo- dariidae), with files on femora I and II, pegs on femora II and III; 12, Mallinella sp. 2 (Zodariidae), with files on femora I, II and III, pegs on femora II, III and IV. Dotted arrows = files and ridges, solid arrows = pegs. above) a total of eight belongs to the possi- bilities. The arrangement in that species gives the impression that different types of stridu- latory organs are present in one spider. Yet, as in the Mallinella, it can be expected that the movement involved is similar for both pairs: moving the femur with the pegs relative to the adjacent file. The main difference with the sit- uation in Olbus is that in Mallinella the pegs are behind the file and it is difficult to imagine that a series of pegs would evolve on the ab- domen. In Olbus it is the other way round and the last femur scratches a file on the abdomen. In this way it is possible to have four stridu- lating files on the same side operating them in sequence and with the same movement, like it must be the case for the organs in M. cine- tipes. A spider with an eight instrument or- chestra thus theoretically belongs to the pos- sibilities. ACKNOWLEDGMENTS I am very greatly indebted to John Murphy for the Singapore specimens and to Peter Schwendinger (Geneve, Switzerland), who made his zodariid collections from Thailand available for study. I thank Gabriele Uhl and Gail Stratton who commented on an earlier version of the manuscript. I thank J. Cillis (Royal Institute for Natural Sciences Brussels) for help with the SEM-images and A. Reygel for the drawings. Figures 13-14. — Mallinella species 1 from Thailand S. 13, Habitus with arrows indicating files on femoral base I and II; 14. Stereoscan micrograph of femoral file. Scale bars = 1 mm (13); 0.1 mm (14). 602 THE JOURNAL OF ARACHNOLOGY Figures 15-18. — Mallinellci cinctipes Singapore S. 15. Left femur I showing dorsal swelling with stridulating file (dotted arrow); 16. Right femur II showing dorsal swelling with stridulating file (dotted arrow); 17. detail of Fig. 15 showing knobless hair bases (white arrow); 18. detail of Fig. 16 showing pegs in the shape of a knob at hair base (solid black arrow). LITERATURE CITED Agnarsson, 1. 2004. Morphological phylogeny of cobweb spiders and their relatives (Araneae, Ar- aneoidea, Theridiidae). Zoological journal of the Linnean Society 141:447-626. Eberhard, W. G. 1996. Female Control: Sexual Se- lection by Cryptic Female Choice. Princeton University Press, Chichester, West Sussex, 501 PP- Edwards, G.B. 1982. Sound production by courting males of Phidippiis mystaceus (Araneae: Salti- cidae). Psyche 88:199-214. Forster, R. R., N. I. Platnick & J. Coddington. 1990. A proposal and review of the spider family Syn- otaxidae (Araneae, Araneoidea), with notes on theridiid interrelationships. Bulletin of the Amer- ican Museum of Natural History 193:1-116. Gwinner-Hanke, H. 1970. Zum Verhalten zweier stridulierender Spinnen, Steatoda bipunctata Linne und Teutana grossa Koch (Theridiidae, Araneae), unter besonderer Berticksichtigung des Fortzpflanzungsverhaltens, Zeitschrift fur Tierp- sychologie 27:649-678. Hinton, H.E. & R.S. Wilson. 1970. Stridulatory or- gans in spiny orb-weaver spiders. Journal of Zo- ology of London 162:482-484. Jocque, R. 1991. A generic revision of the spider family Zodariidae (Araneae). Bulletin of the American Museum of Natural History 201:1- 160. Jocque, R. 1993. “We’ll meet again”, an expression remarkably applicable to the historical biogeog- raphy of Australian Zodariidae. Memoirs of the Queensland Museum 33:561-564. Jocque, R. 1998. Female choice, secondary effect of “mate check”? A hypothesis. Belgian Journal of Zoology 128:99-117. Jocque, R. & T Sziits. 2001. Bacelarella (Araneae, Salticidae) in eastern Cote d’Ivoire: salticid ra- diation in a poorly lit environment. Annales du Musee royal de V Afrique centrale 285:93-99. Jocque, R. 2002. Genitalic polymorphism — a chal- lenge for taxonomy. Journal of Arachnology 30: 298-306. Koh, J. 1989. A Guide to the Common Spiders of Singapore. Singapore Science Centre 160 p. Legendre, R. 1963. L’ audition et remission de sons chez les Araneides. Annales de Biologic 2:371- 390. JOCQUE— MULTIPLE STRIDULATION ORGANS IN SPIDERS 603 Maddison, W. 1987. Marchena and other jumping spiders with an apparent leg-carapace stridula- tory mechanism (Araneae: Salticidae: Helio- phaninae and Thiodininae). Bulletin of the Brit- ish Arachnological Society 7:101-106. Maddison, W.P. & G.E. Stratton. 1988. Sound pro- duction and associated morphology in male jumping spiders of the Habronattiis agilis spe- cies group (Araneae, Salticidae). Journal of Ar- achnology 16:199-211. Maddison, W.P. & G.E. Stratton. 1988. A common method of sound production by courting jumping spider (Araneae, Salticidae). Journal of Arach- nology 16:267-269. Ramirez, M.J., L. Lopardo & A.B. Bonaldo. 2001. A review of the Chilean spider genus Olbus, with notes on the relationships of the Corinnidae (Arachnida, Araneae). Insect systematics and evolution 31:441-462. Rovner, J.S. 1975. Sound production by nearctic wolf spiders. A substratum-coupled stridulatory mechanism. Science 190:1309-1310. Simon, E. 1937. Les arachnides de France. Tome VI. Synopsis generale et catalogue des especes frangaises de Pordre des Araneae; 5e et derniere partie. Paris, 6:979-1298. Starck, J.M. 1985. Stridulationsapparate einiger Spinnen — Morphologic und evolutionsbiologis- che Aspekte. Zeitschrift fiir zoologische Syte- matik und Evolutionsforschung 23:1 15-135. Uetz G.W. & G.E. Stratton. 1982. Acoustic com- munication and reproductive isolation in spiders. Pp. 123 — 159. In P. Witt & J. Rovner, eds. Spider Communication. Mechanisms and Ecological Significance. Princeton University Press. Uhl, G & M. Schmitt. 1996. Stridulation in Palpi- manus gibbulus Dufour (Araneae, Palpimani- dae). Revue Suisse de Zoolologie. hs. 649-660. Workman, T. 1896. Malaysian spiders. Vol. I. Bel- fast, pp. 25-104. Wunderlich, J. 1995. Beschreibung der neuen Spin- nen-Gattung Zygiometella der Tetragnathidae: Metinae und eines bisher unbekannten Typs von Stridulations-Organen (Arachnida: Araneae). Beitrage zur Araneologie 4:639-642. Manuscript received 20 December 2004, revised 17 June 2005. 2005. The Journal of Arachnology 33:604-612 FIRST ULTRASTRUCTURAL OBSERVATIONS ON THE TARSAL PORE ORGAN OF PSEUDOCELLUS PEARSEI AND R BONETI (ARACHNIDA, RICINULEI) Giovanni Talarico\ Jose G. Palacios- Vargas^, Mariano Fuentes Silva^ and Gerd Alberti^: ^Zoological Institute & Museum, Ernst-Moritz-Arndt-University Greifswald, J.-S.-Bach-Str. 11/12, D-17489 Greifswald, Germany. E-mail: g.talarico@gmx.net; ^Laboratorio de Ecologia y Sistematica de Microartropodos, Departamento de Ecologia y Recursos Naturales, Facultad de Ciencias, UNAM, Mexico. ABSTRACT. Due to their relative rareness and restricted distribution, little is known about the ultra- structure of ricinuleids. In particular, sense organs have not been the subject of electron microscopic research until now. Ricinuleids use their forelegs to explore their surroundings with tentative movements. The distal tarsomeres of legs I and II of two cavernicolous Mexican species, Pseudocellus pearsei from the Yucatan Peninsula and Pseudocellus honeti from GueiTero, were examined in this study with light microscopy, scanning (SEM) and transmission electron microscopy (TEM). A conspicuous feature of the distal tarsomeres of legs I and II is a single circular opening that extends as a deep tube-like pit into the tarsus. This pore organ is lacking in the 6-legged larvae. Comparable organs are present in Araneae, Scoipiones, Amblypygi and Anactinotrichida. The tarsal organs of the mentioned groups possess several types of sensilla (olfactory, thermo- and hygrosensitive and mechanosensitve). The pore organ is located in the distal third of the dorsal half of the tarsus. In longitudinal sections it shows a long oval shape. In cross sections it is nearly circular. The pore organ contains a large number of long, slightly curved setae. These setae are localized on the bottom and the lower two thirds of the wall of the pit and project into the lumen. The upper third of the wall is free of setae and shows folds which extend parallel to the opening. All setae inside the pit seem to be of the same type. In sections they show a complex inner structure and likely represent chemoreceptive wall pore single-walled (wp-sw) sensilla. This indicates a possible olfactory function. The pore organ is underlain by numerous gland cells which represent char- acteristics of unicellular “class I” gland cells. Keywords: Tarsus, ultrastructure, sensory organs The order Ricinulei Thorell 1892 is one of the smallest arachnid groups. Only 56 recent species, all belonging to the family Ricinoidi- dae Ewing 1929, have been described. The re- cent species are divided into three genera. Ri- cinoides Ewing 1929 is from Western Central Africa, and Cryptocellus Westwood 1874 and Pseudocellus Platnick 1980 are both from Central America. Ricinuleids inhabit humid layers of soil and litter in tropical rainforests or caves (Cooke 1967; Mitchell 1970; Adis et al. 1989). They pass through 5 postembryonic life stages: a 6-legged larva, 3 nymphal stages (proto-, deuto- and tritonymph) and the adult stage (Mitchell 1970). Most available studies about ricinuleids are taxonomic (Mitchell 1970). The knowledge about their internal morphology is based on relatively few old fundamental studies (e.g., Hansen & Sprensen 1904; Millot 1945, 1949). ■ Until recently, little has been known about the , ultrastructure of ricinuleids as there have been few scanning and transmission electron micro- I scopic studies on this animal group (for SEM see Legg 1976, 1977; Dumitresco & Juvara- |: Bals 1977; Platnick & Shadab 1976, 1977; j Harvey 1984; Adis et al, 1999; for TEM see ; Alberti & Palacios- Vargas 1984; Ludwig & ' Alberti 1990; Ludwig et al. 1994). In partic- f ular, sensory organs have not been subjected j! to electron microscopic research until now. Ricinuleids use their forelegs, especially the ,! elongated second leg, to explore their sur- i roundings with tentative movements (Pollock | 1967). Hence the presence of different sensilla ‘ on the distal tarsomeres of the forelegs can be |j I 604 TALARICO ET AL.— TARSAL PORE ORGAN OF TWO RICINULEIDS 605 expected. Some authors identified different types of setae and other surface structures on the tarsi and expected them to be sensilla (e.g., Hansen & Sprensen 1904; Pittard & Mitchell 1972; Dumitresco & Juvara-Bals 1973, 1976; Legg 1976), but information about their ultra- structure and possible function are still not available. In the present work, we intend to present the first ultrastructural study of the pore organ of the foreleg tarsi of Ricinulei. METHODS The distal tarsomeres of leg I and II of two cavernicolous Mexican species were exam- ined in this study. Specimens of Pseudocellus pearsei (Chamberlin & Ivie 1938) from Yu- catan peninsula were collected in three differ- ent caves, Gruta Actun Chen (Quintana Roo; 20° 20H3" N & 87° 20' 45" W), Gruta X-Caret (Quintana Roo; 20° 33' 54"N & 86° 58' 49" W) and Gruta Sabac-Ha (Yucatan; 20°10'18"N & 89°16'03"W). Pseudocellus boneti (Bolivar and Pieltain 1941) from Guerrero was col- lected in the caves Grutas de Acuitlapan (Mexico; 18°38'00" N & 99° 31' 55'" W). Both species have been found in bat guano or under flat stones. For SEM, 7 specimens of P. pearsei (1 larva, 1 protonymph, 1 deuto- nymph, 3 adult males and 1 adult female) and 4 specimens of P. boneti (1 larva, 1 trito- nymph and 2 adult males) stored in ethanol (70%) were dehydrated in graded ethanols, critical-point dried and coated with gold-pal- ladium. Examination was performed on a LEO DSM 940. For TEM the distal tarsomer- es I and II of 5 specimens of P. pearsei (3 deutonymphs and 2 adult males) were dis- sected in ice-cold Sorensen phosphate buffer (pH 7.4; 0.1 M) and then fixed in 3.5% glu- taraldehyde buffered in Sorensen phosphate buffer overnight. Further processes included postfixation with OSO4 (2%) for two hours, rinsing in buffer, dehydration in graded etha- nols and embedding mainly in Spurrs medium (Spurr 1969) and alternatively in Epon-Aral- dite. Ultrathin sectioning with a Diatome di- amond knife took place on a Leica Ultracut. Sections were stained with saturated uranyla- cetate (in 70% methanol) for 5 minutes and lead citrate according to Reynolds (1963) for 15 minutes. The sections were examined with a Zeiss EM 10 A. For general orientation sem- ithin sections (400-700 nm) were used which were stained according to Richardson et al. (1960). All sections and voucher specimens are housed in the Zoological Institute & Mu- seum of the University of Greifswald. RESULTS The pore organ is located in the distal third of the dorsal half of the distal tarsomeres of legs I and II (Figs. 1, 3). The width of the opening is about 32 pim in P. pearsei and 36 jam in P. boneti (Figs. 2, 4). The edge of the opening differs slightly in both species. In P. pearsei the edge is smooth without any pro- jections (Fig. 2), while in P. boneti there are some short and thin microtrichae which pro- ject radially into the center of the opening (Fig. 4). Except for the larvae (Figs. 5, 6), this structure is present on the forelegs of each in- vestigated life stage and both sexes of P. pear- sei and P. boneti. The pore organ extends as a deep tube-like pit into the tarsus. In longi- tudinal sections it shows a long oval shape (Figs. 7-10). In cross sections it is nearly cir- cular (Figs. 11, 12). Sexual dimorphism could not be observed in the present material. In both species the pore organ contains a large number of long slightly curved setae. These setae are localized on the bottom and the lower two thirds of the wall of the pore organ and project into the lumen but do not reach the opening (Figs. 2, 4, 13). The upper third of the wall is free of setae and shows folds which extend parallel to the opening (Fig. 13, 14). Some small openings in the wall are visible (Fig. 13, 14). In SEM micrographs, the setae show a great number of wall pores (Fig. 15) but in some parts of the shaft the openings of these pores are covered by drop- lets of different size (Figs. 16, 17). In P. pear- sei all setae inside the pore organ seem to be of the same type (Fig. 12). Sections reveal the complex wall of these setae. It consists of two layers: a thick inner wall with up to 25 pores per section and a thin outer wall with a similar number of pores which are plugged by elec- tron dense bodies (Figs. 18-20). Some setae are partly surrounded by secretions (Figs. 19). This is very evident in the basal part of the pore organ where most of them arise (Figs. 23, 26). The sockets of the setae are inflexible (Fig. 26). The setae are innervated by 4-7 out- er dendritic segments (Figs. 21, 22). These are surrounded by an enveloping cell and many densely arranged microvilli (Fig. 21). The lat- ter are formed by the tormogen cell which 606 THE JOURNAL OF ARACHNOLOGY Figures 1-6. — ^Distal tarsomeres. 1. Tarsus II of Pseudocellus pearsei (adult male). Scale bar = 100 (xm. 2. Pore organ opening of that tarsus. Scale bar = 10 p.m. 3. Tarsus I of Pseudocellus boned (adult male). Scale bar = 50 fxm. 4. Opening of the pore organ of tarsus I of P. boneti (tritonymph). Note the small microtrichae (arrow). Scale bar = 10 pm. 5. Tarsus I of P. pearsei (larva). Scale bar = 100 pm. 6. Tarsus II of P. pearsei (larva). Scale bar = 100 pm. Note the dorsofrontal region of the tarsi without pore organ (arrows). TALARICO ET AL.— TARSAL PORE ORGAN OF TWO RICINULEIDS 607 Figures 7-12. — Light and TEM micrographs of the pore organ of Pseudocellus pearsei. 1. Sagittal section of tarsus 11. Scale bar = 50 (xm. 8. Detail of pore organ with longitudinal and oblique sections of sensilla. Scale bar = 10 ixm. 9. Horizontal section of tarsus 1. Scale bar = 50 pm. 10. Detail of the pore organ base and some oblique sections of sensilla. Scale bar = 10 pm. 1 1. Transversal section of tarsus 1. Scale bar = 50 pm. 12. Detail of the lumen with cross sections of sensilla. Scale bar = 10 pm. Abbre- viations; Cl = claw, Cu = cuticle, PS = pore organ-sensilla. Sec = secretion. 608 THE JOURNAL OF ARACHNOLOGY Figures 13-17. — Surface of the pore organ integument and the pore organ-sensilla. 13. Longitudinal section of the pore organ of tarsus I of Pseudocellus pearsei (adult female) with three lateral inserted sensilla and a gland opening (arrow). Scale bar = 30 pm. 14. Detail of the integument with folds and a gland opening (arrow). Scale bar = 3 pm. 15. The sensilla shaft of P. pearsei with many wall pores. Note the damaged area (arrow). Scale bar = 1 pm. 16. Some sensilla with numerous droplets covering the wall pores. Scale bar = 1 pm. 17. Pore organ-sensillum of Pseudocellus boneti with a totally covered surface. Scale bar = 1 pm. TALARICO ET AL.— TARSAL PORE ORGAN OF TWO RICINULEIDS 609 produces the slightly electron dense receptor lymph (Figs, 21, 26), A dendritic sheath is lacking. The dendritic segments terminate in the basal part of the shaft. Pore tubules be- neath the wall pores are lacking. The apical part of the shaft is completely filled with re- ceptor lymph of different electron densities (Fig, 18), Large gland cells which are formed by modified epidermal cells occur between the sensilla forming cells (Figs. 23, 26). The glands appear sack-like and each one forms a large secretion reservoir which is filled with an almost electron lucent material (Figs. 23, 26). Large nuclei, numerous mitochondria, se- cretion vesicles and microvilli, which project into the reservoirs, are present in these cells (Figs. 23, 26). The secretion seems to be de- livered through at least 1 pore, partly filled with granular material, into the lumen of the tarsal pore organ (Figs. 24, 25) but it can not be excluded that a gland cell exhibits more than 1 pore. DISCUSSION The first short description of the tarsal pore organ was given by Pittard & Mitchell (1972). They named it “deep pit” and found that this structure is not present in the larva but on the distal tarsomeres of leg I and II of all further life stages. Our observations confirm these re- sults for P. pearsei and partly for P. boneti. Dumitresco & Juvara-Bals (1973) suggested the “organe tarsal” may be comparable to the tarsal organs of other Arachnida. These are I present in Araneae, the tarsal organs on palps f and walking legs (e.g., Blumenthal 1935; Foe- lix & Chu-Wang 1973), Scorpiones (Foelix & Schabronath 1983), Amblypygi (Foelix et al. 1975) and in Anactinotrichida, the well known I Haller’s Organ on tarsus I of Ixodida and Hol- I othyrida and the telotarsal organ on tarsus I j of Opilioacarida (summarized by Alberti & ' Coons 1999; Coons & Alberti 1999). The tar- sal organs of the mentioned groups possess I several types of sensilla. Olfactory, thermo-/ hygrosensitive and mechanosensitve receptors could be identified in numerous studies. ' The tarsal pore organ of ricinuleids shows j similarities to the proximal part of Haller’s or- i gan, the capsule, in Ixodida. These capsules bear 2-7 sunken sensilla (see Foelix & Axtell 1972; Coons & Alberti 1999) which have ol- factory function. According to the concepts of pore structures and the function of arthropod sensilla (e.g., Altner 1977; Altner & Prillinger 1980; Tichy & Barth 1992; Steinbrecht 1997; Hallberg & Hansson 1999), three main types of olfactory sensilla are known: 1) single- walled sensilla with simple wallpores, 2) sin- gle-walled sensilla with plugged wallpores and 3) double-walled sensilla with spoke ca- nals. Single-walled sensilla with plugged wallpores are present in the capsule of Haller’s organ (Foelix & Axtell 1972). The wall of the pore organ-setae of P. pearsei differs in struc- tural details from the main types described above and also from the capsule-sensilla of Ixodida. Although wall pores with some kind of pore plugs are clearly present, the complex thin outer layer (Figs. 18-20), which may con- sist of another type of secretion instead of cu- ticle, makes it difficult to assign the pit-setae to one type of sensilla. Foelix & Axtell (1972) described a thin layer of “extracellular mate- rial” which often covers the capsule-sensilla, but this layer has no complex structure. It is not clear whether this layer consists of recep- tor lymph or other secretions but the authors note that it was only prominent after simul- taneous glutaraldehyde-0s04 fixation which was not performed in this study. Indeed the phenomenon of droplets appearing on the sur- face of sensilla (Figs. 16, 17) is explained as dried receptor lymph (Foelix & Schabronath 1983). Altner (1977) pointed out that pore structures exist which do not fit to the classi- fication system of sensilla types. However, the presence of wall pores and innervating den- drites (Figs. 15-22) in the pore organ-setae of P. pearsei indicate an olfactory function. The limited material does not allow the reconstruc- tion of the exact innervation pattern (e.g. number and organization of neurons) of this organ. Therefore further investigations are needed. According to Foelix & Axtell (1972) and with regard to the more or less endogeous living of ricinuleids we believe that the tarsal pore organ serves, similar to the capsule of Haller’s organ of Ixodida, mainly as a protec- tive device for numerous olfactory sensilla, which could easily be damaged mechanically if been exposed to the tarsal surface. However, only electrophysical proofs can verify the sen- sory function of an organ (see e.g., Dumpert 1978; De Bruyne & Guerin 1994). In Ixodida a large multicellular gland be- neath the capsule is known (Foelix & Axtell 610 THE JOURNAL OF ARACHNOLOGY Figures 18-26. — Ultrastructure of the tarsal pore organ of PseiidoceUiis pearsei. 18. Cross section of a pore organ-sensillum (apical shaft). Scale bar = 0.5 pm. 19. Cross section of the basal shaft with droplets of secretion. Scale bar = 0.5 pm. 20. Detail of the wall. Scale bar = 0.2 pm. 21. Transverse section of a sensillum socket with 4 outer dendritic segments (inset). Scale bars = 0.5 pm, 0.2 pm. 22. Horizontal section of dendrites beneath a sensillum. Scale bar = 1 pm. 23. Horizontal section of the pore organ base with gland cells between the sensilla forming cells (asterisks). Scale bar = 5 pm. 24. Transverse section of pores (arrows) in the integument between sensilla sockets. Scale bar = 0.5 pm. 25. Horizontal section of a pore filled with granular material (arrow). Scale bar = 0.5 pm. 26. Detail of Fig. 23. Scale bar = 2 pm. Abbreviations: Cu = cuticle, eC = enveloping cell, gR = glandular reservoir, iL = inner layer. Mi = mitochondria, Mv = microvilli, N = nucleus, oD = outer dendritic segment, oL = outer layer, PP = pore plug, RLy = receptor lymph. Sec = secretion, tC = tormogen cell. TALARICO ET AL.— TARSAL PORE ORGAN OF TWO RICINULEIDS 611 1972), Their glandular openings were found in the capsule walL It was suggested that this gland might be the origin of the material sur^ rounding the capsule-sensilia. The large glands beneath the pore organ of P, pearsei are supposed to produce the secretion present between the sensilla and on their surface (Figs. 19, 23~26). They are believed to rep- resent enlarged unicellular “class F’ epider- mai glands according to the classification of Noirot & Queenedy (1974, 1991). Such glands pour their secretions through a simple pore without any special canal formation (Figs. 13, 14, 24, 25). The secretion may sup= port the binding of odorants or probably rinses the sensilla surfaces to keep them clean. However, some main differences between Haller’s organ of Ixodida and the tarsal organ of ricieuleids are evident. The tarsal pore or- gan of ricieuleids occurs on leg I and leg II not only on leg I like in ticks and it contains many more sensilla than the capsule of ticks. Furtherm.ore Haller’s organ is present in ixo- did larvae but the tarsal pore organ is not pre- sent in the larva of ricinuleids. In ticks Hall- er’s organ is the main receptor for host detection in all life stages (Foelix 1985). Ri- cinuleids are not parasitic. If olfactory func- tion can be confirmed in the future, detection of other odorants can be expected. Like in Ar- aneae (Dumpert 1978) pheromone detection is imaginable, because this might not be impor- tant for the larvae. Unfortunately, the knowl- edge of the biology of these animals, in par- ticular the dynamics between individuals in their habitats is still too poor to enable any suggestions in this case. For these reasons, further investigations including also species of the other two genera and on the biology of Ricieulei are required. ACKNOWLEDGMENTS G.T wishes to express his special thanks to Aeja Klann and Peter Michalik for comments and support. He also thanks Romano Daliai, Fabiola Giusti, Valerio Vignoli and all their colleagues at the Department of Evolutionary Biology, Siena for their generous help and kind hospitality. Some specimens of Pseudo- cellus pearsei from Yucatan Peninsula were collected and donated to P.-V. by Mrs. Josiane Lips. All authors appreciate the technical as- sistance of Christine Putzar and Hartmut Fi- scher. G.T, P.-V. and G.A. remember Mariano Fuentes Silva who died in April 2004. LITERATURE CITED Adis, J.U., B. Messner & N.L Platnick. 1999. Mor- phological structures and vertical distribution in the soil indicate facultative plastron respiration in Cryptocellus adisi (Arachnida, Ricinulei) from Central Amazonia. Studies on Neotropical Fauna & Environment 34:1—9. Adis, J.U., N.L Platnick, J.W. de 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. Alberti, G. & L.B. Coons. 1999. Acari: Mites. Pp. 515 — 1215. In Microscopic Anatomy of Inver- tebrates. VoL 8C: Chelicerate Arthropoda, (F.W. Harrison & R.F. Foelix, eds.). Wiley-Liss, New York, Chichester. Alberti, G. & J.G. Palacios- Vargas. 1984. Fine structure of spermatogenesis and mature sper- matozoa in Cryptocellus boned Bolivar y Piel- tain, 1941 (Arachnida, Ricinulei). Journal of Ul- trastructural Research 87:1-12. Altner, H. 1977. Insektensensillen: Bau und Funk- tionsprinzipien. Verhandluegen der Deutschen Zoologische Gesellschaft 70:139-153. Altner, H. & L. Prillinger. 1980. Ultrastructure of invertebrate chemo-, thermo-, and hygrorecep- tors and its functional significance. International Review of Cytology 67:69-139. Blumenthal, H. 1935. Untersuchungen fiber das “Tarsalorgan” der Spinnen. Zeitschrift fiir Mor- phologic und Okologie der Tiere 29:667-719. Cooke, J.A.L. 1967. Observations on the biology of Ricinulei (Arachnida) with description of two new species of Cryptocellus. Journal of Zoology, London 151:31-42. Coons, L.B. & G. Alberti. 1999. Acari: Ticks. Pp. 267 — 514. In Microscopic Anatomy of Inverte- brates. Vol. 8B: Chelicerate Arthropoda. (F.W. Harrison & R.F. Foelix, eds.). Wiley-Liss, New York, Chichester. De Bruyne, M. & P.M. Guerin. 1994. Isolation of 2,6-Dichlorophenol from the cattle tick Boophi- lus microplus: Receptorcell responses but no ev- idence for a behavioural response. Journal of In- sect Physiology 40:143-154. Dumitresco, M. & L Juvara-Bals. 1973. Cryptocel- lus cubanicus n. sp. (Arachiiida-Ricinulei). Re- suitats des Expeditions Biospeologiques Cubano- Roumaines a Cuba 1:259-275. Dumitresco, M. & L Juvara-Bals. 1976, Position systematique de Heteroricinoides bordoni n.g. n.sp. dans la famille Ricinuleididae (Arachnida). Boletm de la Sociedad Venezolana de Espeleo- logia 7:147-180. Dumitresco, M. & 1. Juvara-Bals. 1977. Quelques 612 THE JOURNAL OF ARACHNOLOGY details de morphologic tegumentaire chez Cryp- tocelhis ciibaniciis Dumitresco, Juvara. Resultats des Expeditions Biospeologiques cubano-rou- maines a Cuba 2:145-146. Dumpert, K. 1978. Spider odor receptor: Electro- physical proof. Experientia 34:754-756. Foelix, R.F. 1985. Mechano- and chemoreceptive sensilla. Pp. 118 — 137. In Neurobiology of Arachnids. (EG. Barth, ed.). Springer- Verlag, Berlin, Heidelberg, New York. Foelix, R.F. & R.C. Axtell. 1972. Ultrastructure of Haller’s organ in the tick Amblyomma american- um (L.). Zeitschrift fiir Zellforschung 124:275- 292. Foelix, R.F. & I. Chu-Wang. 1973. The morphology of spider sensilla: II. Chemoreceptors. Tissue & Cell 5:461-478. Foelix, R.F, I. Chu-Wang & L. Beck. 1975. Fine structure of tarsal sensory organs in the whip spi- der Admetiis piimilio (Amblypygi, Arachnida). Tissue & Cell 7:331-346 Foelix, R.F. & J. Schabronath. 1983. The line struc- ture of scorpion sensory organs: I. tarsal sensilla. Bulletin of the British Arachnological Society 6: 53-67. Hallberg, E. & B.S. Hansson. 1999. Arthropod sen- silla: Morphology and phylogenetic consider- ations. Microscopy Research and Technique 47: 428-439. Hansen, H.J. & W. Sprensen. 1904. The order Ri- cinulei. Pp. 114 — 157. In On Two Orders of Arachnida. (H.J. Hansen & W. Sprensen, eds.). Cambridge University Press, Cambridge. Harvey, M.S. 1984. The female of Ricinoides wes- termanni (Guerin-Meneville), with notes on those of R. afzelii (Thorell) and R. kcirschii (Han- sen & Sprensen) (Ricinulei). Bulletin of the Brit- ish Arachnological Society 6:205-210. Legg, G. 1976. The external morphology of a new species of ricinuleid (Arachnida) from Sierra Le- one. Zoological Journal of the Linnean Society 59:1-58. Legg, G. 1977. Sperm transfer and mating in Ri- cinoides hanseni (Ricinulei: Arachnida). Journal of Zoology, London 182:51-61. Ludwig, M. & G. Alberti. 1990. Pecularities of arachnid midgut glands. Acta Zoologica Fennica 190:255-259. Ludwig, M., J.G. Palacios-Vargas & G. Alberti. 1994. Cellular details of the midgut of Crypto- cellus boneti (Arachnida: Ricinulei). Journal of Morphology 220:263-270. Millot, J. 1945. L’anatomie interne des Ricinulei (Arachnides). Annales des Sciences Naturelles, Zoologie 7:1-29. Millot, J. 1949. Ordre des Ricinuleides. Pp. 744 — 760. In Traite de Zoologie. Vol. 6. (PR Grasse, ed.). Masson, Paris. Mitchell, R.W. 1970. Population size and dispersion and species associations of a Mexican caverni- cole ricinuleid (Arachnida). Ciencia, Mexico 27: 63-74. Noirot, C. & A. Quennedy. 1974. Fine structure of insect epidermal glands. Annual Review of En- tomology 19:68-80. Noirot, C. & A. Quennedy. 1991. Glands, gland cells, glandular units: some comments on ter- minology and classification. Annales de la So- ciete entomologique de France 27:123-128. Pittard, K. & R.W. Mitchell. 1972. Comparative morphology of the life stages of Cryptocellus pe- laezi (Arachnida, Ricinulei). Graduate Studies of the Texas Tech University 1:3-77. 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 (Arachnida, Ricinulei). American Museum Novitates 2633:1-17. Pollock, J. 1967. Notes on the biology of Ricinulei (Arachnida). Journal of the West African Science Association 12:19-22. Reynolds, E.S. 1963. The use of lead citrate at high pH as an electron-opaque stain in electron mi- croscopy. Journal of Cell Biology 17:208-212. Richardson, H.H., L. Jarret & E.H. Finke. 1960. Embedding in epoxy resins for ultrathin section- ing in electron microscopy. Stain Technology 35: 313-323. Spun; A.R. 1969. A low-viscosity epoxy resin em- bedding medium for electron microscopy. Jour- nal of Ultrastructural Research 26:31-43. Steinbrecht, R.A. 1997. Pore structures in insect ol- factory sensilla: a review of data and concepts. International Journal of Insect Morphology & Embryology 26:229-245. Tichy, H. & EG. Barth. 1992. Fine structure of ol- factory sensilla in myriapods and arachnids. Mi- croscopy Research and Technique 22:372-391. Manuscript received 21 December 2004, revised 20 June 2005. 2005. The Journal of Arachnology 33:613-621 ULTRASTRUCTURE OF MALE GENITAL SYSTEM AND SPERMATOZOA OF A MEXICAN CAMEL-SPIDER OF THE EREMOBATES PALLIPES SPECIES GROUP (ARACHNIDA, SOLIFUGAE) Anja E. Klann\ Alfredo V, PerettF and Gerd Alberti^: ’Zoological Institute & Museum, Emst-Moritz-Amdt-University Greifswald, Johann-Sebastian-Bach-StraBe 11/12, D-17489 Greifswald, Germany. E-mail: anja.klann@uni-greifswald.de; -CONICET — Catedra de Diversidad Animal I, Eacultad de Ciencias Exactas, Fisicas y Naturales, Universidad Nacional de Cordoba, Av. Velez Sarsfield 299, C. P. 5000, Cordoba, Argentina ABSTRACT. The male genital system of Solifugae is divided into three different parts: a) a common genital chamber, b) the paired tubular vasa deferentia and c) the long, thin testes. On each side, the vas deferens splits into two smaller branches resulting in the thin, extremely long testes such that one indi- vidual possesses four tubular testes in total. The epithelium of a testis consists mainly of a glandular part and of a germinal part surrounded by a small layer of muscles. In Eremobates sp., within the germinal part the sperm cells are groups of a few, probably four, mature sperm cells each surrounded by thin extensions of somatic cells. These somatic cells can clearly be distinguished from the cells forming the glandular part which contain large amounts of rough endoplasmic reticulum. Once released into the narrow testicular lumen, the spermatozoa float more or less individually in a proteinaceous secretion. Earlier stages of spermatogenesis could not be detected, suggesting that spermatogenesis may occur in the subadult male (not examined in this study). In general, the sperm is rather simple, representing a round or slightly elongated cell devoid of a flagellum. The relatively small and flat acrosomal vacuole is attached to the disc-like nucleus. The acrosomal filament penetrates the nucleus and is coiled several times around it. In contrast to species of the family Ammotrechidae or Karschiidae, for which sperm cells have already been described, the sperm cells of the Mexican Eremobates sp., which belongs to the family Eremobatidae, show no tendency to form any piles or well ordered groups in the lumen of either the testes or the vasa deferentia. Keywords: Solifugae, genital system, sperm cell, systematics Most camel-spiders (Arachnida, Solifugae), also called sunspiders or wind-scorpions, in- habit tropical, subtropical regions and arid en- vironments in southern Europe, Africa, Asia and the Americas (Punzo 1998). The oldest specimen of Solifugae is known from the Up- per Carboniferous (Pennsylvanian in US ter- minology) of Mazon Creek, Illinois, USA (Selden & Shear 1996). Most of the 1084 re- cent species (Harvey 2002) are nocturnal predators known for their extreme rapidity. The huge chelicerae represent a characteristic feature of their external morphology and they can be easily distinguished from other arach- nids by the presence of racquet organs (mal- leoli). Their position within the Arachnida is not yet fully resolved, since Solifugae express both apomorphic (e. g. highly developed tra- cheal system, two-jointed chelicerae) and ple- siomorphic (e. g. segmentation of the opistho- soma) characteristics (Roewer 1934; Moritz 1993), but they are usually considered to be the sister-group of the Pseudoscorpiones (Weygoldt & Paulus 1979; Shultz 1990; Wey- goldt 1998; Wheeler & Hayashi 1998; Dunlop 2000; Giribet et al. 2002). In any case, Roew- ers classification of the order Solifugae is based on a small set of character systems and therefore lacks a reliable basis for phyloge- netic and subsequent systematic implications (Harvey 2002). So far, only a few electron microscopic studies on this animal group have been com- pleted (see e.g., Brownell & Farley 1974; Al- berti 1979, 1980; Bauchhenss 1983; Ludwig & Alberti 1992; Alberti & Peretti 2002). Ac- 613 614 THE JOURNAL OF ARACHNOLOGY cording to the current literature, the ventrally located male genital system of Solifugae is generally divided into three different parts: a) a common genital chamber, b) the paired tU“ bular vasa deferentia and c) the long, thin tes- tes. Even though there are several studies on this organ system (see e.g. Roewer 1934; War- ren 1939; Junqua 1966), the nomenclature concerning the different parts of the genital system varies considerably between these au- thors. Only the testes and partly the vasa de- ferentia have been fine-structurally investigat- ed (Alberti 1980; Alberti & Peretti 2002). The aim of the present study was to confirm and to substantiate the present knowledge on the male reproductive system and sperm mor- phology and to present the first ultrastructural study of the genital chamber and its accessory glands. METHODS Males of the genus Eremobates Banks 1900, belonging to the Eremobates pallipes (Say 1823) species group according to Brook- hart (pers. comm.), were captured near Pachuca- City, State of Hidalgo, Mexico (20°07'21"N, 98°44'09"W). After dissection of three males in ice-cold cacodylate buffer their genital sys- tems were fixed in 3.5 % glutaraldehyde buff- ered in cacodylate buffer (pH 7.4; 0.1 M). Eixed genital systems were sent to Germany in diluted glutaraldehyde. Postfixation pro- cesses included treatment with OSO4 (2 %) for two hours, rinsing in buffer solutions, dehy- dration in graded ethanols (60-100 %) and embedding in Spurrs medium (Spun* 1969). Ultrathin sections of approximately 70 nm were cut with a Diatome diamond knife using a Leica Ultracut microtome. Sections were stained with saturated uranylacetate (in 70 % methanol) and lead citrate according to Reyn- olds (1963). Eor general orientation semi thin sections (700 nm) were used which were stained according to the methods of Richard- son et al. (1960). Transmission electron mi- croscopy was performed using a Zeiss EM 10 A transmission electron microscope. For scan- ning electron microscopy, the genital system was dehydrated in graded ethanols (60-100 %), then coated with gold-palladium and fi- nally investigated with a LEO DSM 940. A male Eremobates sp. has been deposited as a voucher specimen in the Museo Argentine de Ciencias Naturales “Bernardino Rivadavia” (MACN) in Buenos Aires. RESULTS Scanning electron microscopical obser- vations.— In general, the male genital system consists of a common genital chamber, the vasa deferentia and the testes. Immediately af- ter being removed from the male, the fresh genital system is translucent yellow. The paired tubular vasa deferentia originate from the genital chamber to which small accessory glands are directly attached. Each vas deferens splits into two smaller branches each resulting in extremely long, thin testes which are only partly shown in Fig. 1. Light and transmission electron micro- scopical observations. — Testes: The long, thin tubular testes are surrounded by small muscle cells. The somatic epithelium is com- posed of a larger glandular and a compara- tively small part in which the germinal cells are embedded (so called germinal part). Cells of the glandular part are characterized by many cisternae of rough endoplasmic reticu- lum and Golgi bodies, often located close to the nucleus. Their nuclei are more or less rounded or slightly oval in shape, approxi- mately twice as large as the nuclei of the so- matic cells of the germinal part and located in the basal half of the cells (Fig. 2). Branching somatic cells forming a meshwork constitute the germinal part in which groups of sperm cells are embedded (Fig. 3). In contrast to the cells of the glandular part, the somatic cells of the germinal part are irregularly shaped and contain only a few cell organelles. Apically, in both somatic cell types there is a border of microvilli. Each sperm group consists of a few, probably four, mature sperm cells (Fig. 3). No spermatogenesis could be observed. The sperm cells float more or less distinctively in the narrow testicular lumen containing dif- ferent kinds of proteinaceous secretions most likely produced by the glandular cells (Fig. 4). Towards the vasa deferentia and shortly before the testes open into the vas deferens, the ep- ithelium flattens and no spermatozoa can be observed in the tissue. The sperm cells are rather simple, representing a roundish or slightly elongated cell body devoid of a fla- gellum, but provided with one, rarely two, flat extensions which fold onto the cell body (Fig. 8, 9, 10). In general, the following character- KLANN ET AL.— MALE GENITAL SYSTEM OF A CAMEL-SPIDER 615 Figure 1. — Scanning electron micrograph of the left side of the male genital system of Eremobates sp. (genital chamber and testes are only partly shown; composed picture). Scale bar = 300 p,m. istic cell components can be distinguished in the mature spermatozoa: acrosomal complex, nucleus and cytoplasm including a more or less electron-lucent area. The acrosomal com- plex can be divided into an acrosomal vacu - ole, amorphous subacrosomal material and the acrosomal filament (perforatorium) starting from the amorphous subacrosomal material. 616 THE JOURNAL OF ARACHNOLOGY Figures 2-4. — Testis. 2, Light micrograph of the transversal section through the testis showing germinal and glandular part. Scale bar = 50 pm. 3. Groups of four spermatozoa embedded in somatic cells of the germinal layer. Left, glandular cells. Scale bar = 2 pm. 4. Sperm cell in the lumen of the testis surrounded by globules of secretions. Scale bar = 2 pm. Abbreviations: GC = glandular cell, Lu = lumen of the testis, Mv = microvilli, N = nucleus, SC = somatic cell, Sec = secretion, Sp = sperm cell. The relatively small acrosomal vacuole is at- tached to the electron-dense nucleus. The nu- cleus is penetrated and surrounded by the ac- rosomal filament (Figs. 8, 9). A conspicuous flat extension of the cell contains no organ- elles and slightly inflates towards its posterior end. The sperm cells show no tendency to form well ordered piles or globules either in the lumina of the testes or in the vasa defer- entia. Vas deferens: The epithelium of the vas de- ferens is underlain by a relatively thick outer cross-striated muscle layer interlaced with small tracheae (Figs. 5, 6). The epithelial cells are connected to the basal lamina via hemi- desmosomes. The nuclei of the cells of the epithelium, containing considerable amounts of rough endoplasmic reticulum, are in'egu- larly shaped. The wide lumen is filled with different kinds of secretions forming distinct KLANN ET AL.— MALE GENITAL SYSTEM OF A CAMEL=SPIDER 617 Figures 5-8. — Vas deferens. 5. Light micrograph of the small branch of the vas deferens. Scale bar = 50 jxm. 6. Epithelium, of the smaller branch of the vas deferens underlain by a muscle layer (composed picture). Scale bar = 4 pm. 7. Nerve fibres (indicated by arrows) within the muscle layer. Scale bar = 1 pm. 8. Single sperm cell in the lumen of the vas deferens. Scale bar = 1 pm. Abbreviations: AF = acrosomal filament, AV = acrosomal vacuole, Ax = axon, BL = basal lamina, ELA = electron-lucent area, Ep = epithelium, FP = flat process, Lu = lumen of the vas deferens. Mu = muscle, Mv = microvilli, N = nucleus, Sec = secretion, Sp = sperm cells, Tr = trachea. 618 THE JOURNAL OF ARACHNOLOGY Figures 9, 10. — Schematic drawings of a sperm cell. 9. Longitudinal section. 10. Three-dimensional reconstruction of the sperm body. globules and mature sperm cells (Fig. 5). The muscle layer is innervated as indicated by the number of nerve fibres observed between the cells (Fig. 7). Genital chamber: Several glandular pouch- es extend from the genital chamber and con- stitute the accessory glands. The glands are provided with an epithelium characterized by many rough endoplasmic cisternae, which are often inflated (Fig. 11). Secretory vesicles are only rarely observable. Apically, the cells bear microvilli (Fig. 12). The epithelium is under- lain by thin muscle cells. The genital chamber is directly connected to the genital opening located on the second opisthosomal segment. In certain regions the epithelium forms many finger-like processes extending into the lumen (Fig. 13). The epi- thelium of the genital chamber consists of a monolayer of cells which are characterized by basal membrane infoldings associated with mitochondria, thus forming a typical basal labyrinth (Fig. 14). Apically, the epithelium is provided with small microvilli over which a thin cuticle is located (Fig. 15). The cells sometimes contain extensive areas filled with glycogen (Fig. 16). A thick muscle layer, which is innervated, is located under the epi- thelium. DISCUSSION The two functionally different types of the epithelial cells of the testes in Solifugae have already been described by Alberti (1980) and Alberti & Peretti (2002). Our observations concerning the fine structure of the sperm cells agree with earlier results confirming the relatively simple ground pattern of sperm morphology in Solifugae. Nevertheless there are differences in the arrangement of the sperm. The observed spectrum in Solifugae covers highly ordered sperm cells in piles, both in the epithelium of the testes, in its lu- men and in the lumen of the vas deferens of a karschiid species, groups of sperm that are less ordered and less compact in an ammotre- chid representative and individual cells at least in the lumen of the vas deferens shown in an ammotrechid and the eremobatid species from Mexico studied here. Furthermore the sperm cells differ in shape and structural de- tails. Some types of sperm cells exhibit mem- brane protuberances to various degrees where- as such structures cannot be observed in other representatives at all. However, it is still too early to apply these results to the systematics of Solifugae, since more species from other families need to be examined. The innervated musculature of the vasa deferentia is certainly involved in the transport of the sperm towards the genital opening and perhaps in releasing the sperm fluid. Reports on sperm transfer differ. According to Heymons (1902), Cloudsley-Thompson (1961), Amitai et al. (1962) and Peretti & Wil- lemart (unpub. data) sperm fluid is transferred semi-directly. A spermatophore or a sperm droplet is deposited by a male on the ground and subsequently picked up with his chelic- erae and transferred to the genital orifice of the female. In contrast, Muma (1966, 1967) and Punzo (1998) reported a direct sperm transfer in the eremobatid solpugids Eremo- bates durangonus Roewer 1934, E. paipise- tulosus Fichter 1941 and E. nodularis Muma 1951 from the genital orifice of a male to that one of the female. The function of the accessory glands is speculative. One possibility is that they could take part in the formation of the sperm drop- let. The extrusion of the secretion seems not to happen earlier than mating, since the lu- mina were almost empty in our specimens. A KLANN ET AL.— MALE GENITAL SYSTEM OE A CAMEL-SPIDER 619 Figures 11-16. — Genital chamber. 11. Periphery of an accessory gland (composed picture). Scale bar = 3 |jLm. 12. Cell apices of an accessory gland. Scale bar = 2 p.m. 13. Epithelium overlain by a thin cuticle (composed picture). Scale bar = 5 pm. 14. Basal labyrinth characterized by membrane infoldings associated with mitochondria. Scale bar = 3 pm. 15. Cell apices of the epithelium with border of small microvilli. Scale bar = 2 pm. 16. Glycogen granules. Scale bar = 2 pm. Abbreviations; Cu = cuticle, ER = endoplasmic reticulum, Gly = glycogen granules, Lu = lumen, M = mitochondrion, Mu = muscle, Mv = microvilli, N = nucleus. 620 THE JOURNAL OF ARACHNOLOGY further source of secretion contributing to the formation of the sperm droplet could be the huge vasa deferentia and the glandular part of the testes, A similar function is known from actinotrichid mites (e.g., Alberti & Coons 1999) , Adults, in particular males, live only a short period of time after mating (Heymons 1902; Punzo 1998). Heymons (1902) in particular emphasized that the spermatophore (i.e. the drop containing sperm fluid) is reduced in size after several copulations. Junqua (1966) pro= posed that spermatogenesis occurs in subadult males prior to the adult molt which is sup^ ported by our ultrastructural investigations of adult males in which spermatogenesis was never detected (see also Alberti 1980; Alberti & Peretti 2002). Therefore it is reasonable to suggest that the testes and the vasa deferentia of an adult male serve only as storage sites for sperm cells until they are transferred dur- ing mating. The apomorphic similarities in sperm cells and in the fundamental organization of the tes- ticular tissue between Solifugae and actinotri- chid mites have been pointed out by Alberti (1980) and Alberti & Peretti (2002). Although the Solifugae are commonly regarded as the sister-group of Pseudoscorpiones (together forming the taxon Haplocnemata, e.g. Wey- goldt & Paulus 1979; Dunlop 2000), there are tremendous differences in sperm morphology. Pseudoscorpiones possess complex coiled-fla- gellate spermatozoa (e.g., Werner & Bawa 1988; Dallai & Callaini 1990; Alberti 2000). Thus, comparative spermatology does not support a close relationship between these two animal groups. However, the assumption that the Acari represent a monophylum may be questioned (Alberti 2000; Alberti & Peretti 2002). It may be argued that the differences in the mode of sperm transfer, indirect sper- matophore transfer in Pseudoscorpiones and direct or semi-direct in Solifugae, may con- sequently be reflected in different sperm types. These differences may not necessarily contradict a sister-group relationship between Pseudoscorpiones and Solifugae. However, it can be shown in other arachnid taxa with comparable sperm transfer, e. g., Araneae or Ricinulei, that sperm morphology is not nec- essarily modified in the same manner as in Solifugae or actinotrichid mites (Alberti 2000) . Furthermore, actinotrichid mites show three kinds of sperm transfer: indirect sper- matophore transfer, direct spermatophore transfer using gonopods and direct insemina- tion via a penis, all possessing simple afla- gellate spermatozoa. Evidently there is no simple correlation between sperm structure and mode of sperm transfer (Weygoldt 1990, Alberti & Peretti 2002). The similarity in the testis histology in the Solifugae and actinotri- chid mites is remarkable. If the Solifugae are closely related to the Pseudoscorpiones (as suggested above), the similarity of the testis histology and the aflagellate sperm must be seen as homoplastic. Another interesting as- pect is the occurrence of a transport epitheli- um in the genital chamber, characterized by conspicuous infoldings of membranes associ- ated with numerous mitochondria. Such a dis- tinct epithelium is also present in the genital papillae of actinotrichid mites (Alberti & Coons 1999), but evaluation of this character in terms of phylogenetic systematics requires further investigations on a broader range of taxa. ACKNOWLEDGMENTS A.E.K. received financial support from the Landesgraduiertenforderung Mecklenburg- Vorpommern, and A.V.P. from Consejo Na- cional de Investigaciones Cientfficas y Tec- nicas (CONICET), SECYT UNC Argentina, and UAEH, Mexico. A.E.K. also wishes to ex- press her thanks to Giovanni Talarico and Pe- ter Michalik for comments and suggestions. All authors thank Jack Brookhart and appre- ciate the technical assistance by Christine Put- zar and acknowledge the suggestions of the referees and editors that helped to improve the manuscript. LITERATURE CITED Alberti, G. 1979. Licht- und elektronenmikrosko- pische Untersuchungen an Coxaldriisen von WaL zenspinnen (Arachnida: Solifugae). Zoologischer Anzeiger 203:48-64. Alberti, G. 1980. Zur Feinstruktur des Hodenepi- thels und der Spermien von Eusimonia mirabilis Roewer, 1934 (Solifugae, Arachnida), Zoolo- gischer Anzeiger 204:345-352. Alberti, G. 2000. Chelicerata. Pp, 311-388. In Pro- gress in Male Gamete Ultrastructure and Phylog- eny. (B. G. M. Jamieson, ed.) In Reproductive Biology of Invertebrates. (K. G. Adiyodi & R. G. Adiyodi). VoL 9B. Oxford & IBH Publishing / Wiley, New Dehli & N. Y. Alberti, G. & L.B. Coons. 1999. Acari — mites. Pp. KLANN ET AL.— MALE GENITAL SYSTEM OF A CAMEL-SPIDER 621 515-1265. In Microscopic Anatomy of Inverte- brates. (F. W. Harrison, ed.). Vol. 8C. John Wiley & Sons, Inc., New York. Alberti, G. & A.V. Peretti. 2002. Fine structure of male genital system and sperm in Solifugae does not support a sister-group relationship with Pseu- doscorpiones (Arachnida). Journal of Arachnol- ogy 30:268-274. Amitai, P., G. Levy & A. Shulov. 1962. Observa- tions on mating in a solifugid Galeodes suifuri- pes Roewer. Bulletin of the Research Council of Israel, Section B, Zoology 11:156-159. Bauchheess, E. 1983. Morphology and ultrastruc- ture of sensilla ampullaceae in Solifugae (Chel- icerata: Arachnida). International Journal of In- sect Morphology and Embryology 12:129-138. Brownell, P.H. & R.D. Farley. 1974. The organi- zation of the malleolar sensory system in the sol- pugid, Chanbria sp. Tissue and Cell 6:471-485. Cloudsley-Thompson, J.L. 1961. Observation on the natural history of the camel-spider Galeodes arabs C. L. Koch (Solifugae: Galeodidae) in the Sudan. The Entomologists Monthly Magazine 97:145-152. Dallai, R. & G. Callaini. 1990. Ultrastructure of the Geogarypus nigrimanus spermatozoon (Arach- nida, Pseudoscorpionida). Acta Zoologica 71:37- 43. Dunlop, J.A. 2000. The epistomo-labral plate and lateral lips in solifuges, pseudoscorpions and mites. Ekologia (Bratislava) 19, Supplement 3: 67-78. Giribet, G., G.D. Edgecombe, W.C. Wheeler & C. Babbit. 2002. Phylogeny and systematic position of Opiliones: a combined analysis of chelicerate relationships using morphological and molecular data. Ciadistics 18:5-70. Harvey, M.S. 2002. The neglected cousins: what do we know about the smaller arachnid orders? Journal of Arachnology 30:357-372. Heymons, R. 1902. Biologische Beobachtungen an asiatischen Solifugee nebst Beitragen zur Syste- matik derselben. Abhandlungee der Koniglich Preussischen Akademie der Wissenschaftern 1901:1-65. Junqua, C. 1966. Recherches biologiques et histo- physiologiques sur un solifuge sahariee Othoes saharae Panouse. Memoires du Museum Nation- al dHistoire Naturelle, Series A, 43:1-124 Ludwig, M. & G, Alberti. 1992. Ultrastructure and function of the midgut of camel-spiders (Arach- nida: Solifugae). Zoologischer Anzeiger 228:1- 11. Moritz, M. 1993. Unterstamm Aracheata. Pp. 64- 442. In Lehrbuch der Speziellen Zoologie (begr. von A. Kaestner). 4.ed. Bd.l: Wirbellose Tiere. 4. Teil: Arthropoda (H.-E. Gruner, ed.): G. Fi- scher Verlag, Jena. Muma, M. H. 1966. Mating behaviour in the sol- pugid genus Eremobates Banks. Animal Behav- iour 14:346-350. Muma, M.H. 1967. Basic behavior of North Amer- ican Solpugida. The Florida Entomologist 50: 115-123. Punzo, F. 1998. The biology of camel-spiders (Arachnida, Solifugae). Kluwer Academic PubL, Boston 301pp. Reynolds, E.S. 1963. The use of lead citrate at high pH as an electron-opaque stain in electron mi- croscopy. Journal of Cell Biology 17:208-212. Richardson, K.C., L. Jarett & E.H. Finke. 1960. Embedding in epoxy resins for ultrathin section- ing in electron microscopy. Stain Technology 35: 313-323. Roewer, C.Fr. 1934. Solifugae, Palpigradi. P. 723. In Klassen und Ordnungen des Tierreichs. Vol. 5, 4, 4 (H. G. Bronn, ed.) Akademische Verlags- gesellschaft, Leipzig. Selden, P.A, & W.A. Shear. 1996. The first Meso- zoic Solifugae (Arachnida), from the Cretaceous of Brazil and a redescription of the Palaeozoic solifuge. Palaeontology 39:583-604. Shultz, J.W. 1990. Evolutionary morphology and phylogeny of Arachnida. Ciadistics 6:1-38. Spurr, A.R. 1969. A low-viscosity epoxy resin em- bedding medium for electron microscopy. Jour- nal of Ultrastructure Research 26:31-43. Warren, E. 1939. On the genital system of certain Solifugae. Annals of the Natal Museum 9:139- 172. Werner, G. & S.R. Bawa. 1988. Spermatogenesis in the Pseudoscorpion Diplotemnus sp. with special reference to nuclear changes. Journal of Ultra- structure and Molecular Structure Research 98: 119-136. Weygoldt, P. 1990. Arthropoda — Chelicerata: Sperm Transfer, Pp. 77-1 19. In Reproductive Bi- ology of Invertebrates, Vol 4B. In Fertilization and Development and Parental Care. (K.G. Adi- yodi & R.G. Adiyodi, eds.) Oxford & IBH Pub- lishing / Wiley, New Delhi & N.Y. Weygoldt, P. 1998. Evolution and systematics of the Chelicerata. Experimental and Applied Acarolo- gy 22:63-79. Weygoldt, P. & H.E Paulus. 1979. Untersuchungen zur Morphologic, Taxonomic und Phylogenie der Chelicerata. II. Cladogramme und die Entfaltung der Chelicerata. Zeitschrift fiir zoologische Sys- tematik und Evolutionsforschung 17:177-200. Wheeler, W.C. & C.Y Hayashi. 1998. The phylog- eny of the extant chelicerate orders. Ciadistics 14:173-192. Manuscript received 21 December 2004, revised 23 June 2005. 2005. The Journal of Arachnology 33:622-628 TERGAL AND SEXUAL ANOMALIES IN BOTHRIURID SCORPIONS (SCORPIONES, BOTHRIURIDAE) Camilo I. Mattoni: Division of Invertebrate Zoology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024-5192, USA. E-mail: cmattoni @ amnh.org ABSTRACT. New data concerning developmental anomalies observed among species of the family Bothriiiridae (Scorpiones) are presented. Tergal malformations in Bothriurus coriaceiis, Brachistosternus roigalsinai and Bothriurus noa are described and illustrated. Two new cases of intersexuality in scorpions, in specimens of Brachistosternus pentheri and Bothriurus araguayae, are reported and discussed. Keywords: Developmental anomalies, tergites, intersexuality, Scorpiones, Bothriuridae There are numerous reports of developmen- tal anomalies in scorpions (Table 1). Most re- ports relate to the duplication of posterior body segments (Vachon 1952; Hjelle 1990; Sissom & Shelley 1995, see the latter for overview); reference to other types of devel- opmental anomalies in scorpions is scarce. One work includes information about anom- alies of the legs and pedipalps (Armas 1977) and another describes a tergal and two cara- pacial malformations (Armas 1976). Most re- cently, Teruel (2004) presented a list and brief description of tergal (see below) and pedipal- pal anomalies; however, only Armas (1977) illustrated the anomalies described, a prereq- uisite for understanding the anomalies report- ed. Teruel (2004: 237) references a “cheliceral anomaly’' on one specimen of a buthid, Ly~ chas obsti Kraepelin 1913. This specimen has two teeth on the ventral surface of the fixed finger of one chelicera and one on the other, the latter expression being the typical condi- tion for Lychas (Vachon 1963; Kovafik 1997). Teruel (2004) notes that this is interesting from a taxonomic viewpoint, because, tradi- tionally, the number of ventral teeth on the fixed finger has been used as a strong char- acter in the generic differentiation of the scor- pions of the family Buthidae (Kraepelin 1899; Sissom 1990). Teruel (2004) suggested that using this character to identify buthids from Northwest Africa could present problems, be- cause it could cause erroneous identifications. The two states clearly represent normal vari- ation in morphology, and are not anomalous. Expression of both states in one specimen is clearly an abnormality and the existence of rare abnormalities does not necessitate the need to abandon these character systems. Fur- thermore, many systematists have observed this kind of variation on the chelicerae of sev- eral species, including cases where both che- licerae are different from the usual morphol- ogy of the species (Mattoni 2003; Prendini pers. comm.). They would not consider these represent any obstacle for identifying taxa in question, because these occurrences are rare in scorpion populations, and there are many additional characters that can assist with an identification. The only references to tergal anomalies in scorpions are the works of Armas (1976), who described a specimen of Didymocentrus trin- itarius Franganillo 1930 (Diplocentridae) with fusion of the carapace and the first tergite, and Teruel (2004) who described anomalies in one male of Microtityus jaumei Armas 1974 (Buthidae), that possessed a double anomaly, with tergite V completely divided on the pos- terior half, and tergite VII fused dorsally to metasomal segment 1. Teruel (2004) also re- ported two female diplocentrids {Cazierius parvus Armas 1984 and C gundlachii (Karsch 1880) and one euscorpiid (Euscorpius flavi- caudis (DeGeer 1778)), possessing totally di- vided tergites. The references to sexual malformations are restricted to 5 reports, involving hermaphro- ditism (with male and female genitalia), gyn- andromoiphism (with both sexes discretely combined) and intersexualism (where the en- 622 MATTONI— ANOMALIES IN SCORPIONS 623 Figures 1-3. — Carapace and tergites of malformed scorpions. 1-2. Females of Bothriuriis coriaceus\ 1. specimen from 4 km N Los Vilos; 2. specimen from Cuesta de Chacabuco. 3. female of Brachistosternus roigalsinai, carapace and tergites. Scale = 5 mm. tire body is intermediate between sexes). Mat- thiesen (1968) described a hermaphrodite specimen of the buthid Tityus bahiensis (Perty 1833). Cokendolpher & Sissom (1988) de- scribed two gynandromorphic diplocentrids (a Cazierius gundlachii (Karsch 1880) and Bio- culus comondae Stahnke 1968). Armas (1990) reported a case of one hermaphrodite Alayo- tityus juraguaensis Armas 1973 and a gyn- andromorphic specimen of Tityopsis inae- qualis (Armas 1974) (Buthidae). Maury (1983) described an adult hermaphrodite of Brachistosternus pentheri Mello-Leitao 1931 (Bothriuridae) showing intersexual and gyn- andromorphic characteristics, and with both embryos and hemispermatophores. Another interesting malformation was reported in two males of the bothriurid Bothriurus bonariensis (C.L. Koch 1842), found mating with females in the field, yet presenting only one hemi- spermatophore, the right paraxial organ (that produces the hemispermatophore) being ab- sent in both specimens (Peretti 2000). The last two reports, together with the pedipalps anom- aly reported by Teruel (2004) on Centromach- etes pocockii (Kraepelin 1894) and Urophon- ius granulatus Pocock 1898, are the only references to malformations among Bothriur- idae. The main goal of this contribution is to de- scribe and illustrate the tergal anomalies that were found in specimens of three Bothriuridae species, and to report two more cases of in- tersexuality in scorpions. The specimens studied are preserved in 80 % ethanol and belong to the following collec- tions: AMNH = American Museum of Nat- ural History (New York, USA); MACN-Ar = Museo Argentino de Ciencias Naturales “Ber- nardino Rivadavia” (Buenos Aires, Argenti- na) and CDA = Catedra de Diversidad Ani- mal I, Universidad Nacional de Cordoba (Cordoba, Argentina). Illustrations were pro- duced using a Leica MS5 stereomicroscope equipped with a camera lucida. Photographs were taken with an Olympus Stylus 400 dig- ital camera under long-wave ultraviolet light. Specimens examined. — Bothriurus cori- 624 THE JOURNAL OF ARACHNOLOGY ciceiis Pocock 1893. CHILE: Santiago Region, Chacabuco Province: 1 $, Cuesta de Chaca- buco, S side, elev. 3900 ft, dry mountainside (32°59' S, 70°44' W), 14 I 1985, N. Platnick, O.E Francke, AMNH; Coquimbo Region, Choapa Province: 1 $, 4 km N Los Vilos (31°72' S, 7L3U W), 5 I 1985, N. Platnick, O. F. Francke, AMNH. Bothriurus noa Maury. ARGENTINA: Tucuman Province: 1 $ (par- atype), Tafi del Valle (26°52' S, 65°5U W), 1970 m, 16 I 1981, E. Maury, MACN-Ar 7571. Brachistosternus (Leptosternus) roigal- sinai Ojanguren-Affilastro 2002. CHILE: At- acama Region, Huasco Province: 1 9 , Llanos de Challe National Park, (28°09'39.8" S, 71°03'20.0" W), 205 m, XII 1997, J. Cepeda- Pizarro, CDA. Brachistosternus (L.) pentheri. ARGENTINA: Mendoza Province: 1 d (?), Reserva de la Biosfera Nacunan (34°02' S, 67°54' W), 540 m, 20 XI 2003, C. Mattoni, L. Prendini, J. Ochoa, CDA. Bothriurus ara- guayae Vellard 1934. BRAZIL: Sao Paulo State: 1 S (?), Estagao Ecologica de Itirapina, Municipio de Itirapina (22°15' S, 47°49' W), pitfall, 27 VIII 1999, G. Machado. Two of the females (both B. coriaceiis) were pregnant when preserved. Tergal malformations. — The B. coriaceus specimen from Cuesta de Chacabuco (Fig. 1) shows completely longitudinally divided ter- gites IV and V. Both parts of each tergite rep- resent almost exactly in shape and size the corresponding half of the tergite; only a small central portion is lacking from the posterior edge. The B. noa female shows the same kind of anomaly but only on tergite III. The specimen of B. coriaceus from 4 km N Los Vilos (Fig. 2) also presents some divided tergites, with a different arrangement: tergite IV is completely divided but tergite V is fused sinistraly with the dextral half of tergite VI. The recognition of each part of the tergites is difficult because tergites V and VI are almost the same size. The Brachistosternus (L.) roigalsinai spec- imen (Fig. 3) displays a different kind of mal- formation: the sinistral half of tergite I is free, and the dextral half is joined to the sinistral half of tergite II (one can recognize the tergite because I and II differ in size). The dextral half of tergite II is joined anteriorly to the ter- gite III. All observed specimens with tergal anom- alies do not show any other evident malfor- mation, except for the specimen of B. cori- aceus from Chacabuco that has a slightly abnormal telson vesicle, with the left ventral side a little depressed. The pigmentation pattern of the malformed tergites is normal on the Bothriurus coriaceus and B. noa specimens, and apparently in the Brachistostenus (L.) roigalsinai female as well. However, in the latter case the specimen is not well fixed, and not all the pigmentation has been preserved. Tergal malformations include division of tergites and fusion of tergal parts to one an- other. Armas (1976) described a fusion of ter- gite I to carapace, and Teruel (2004) observed division of tergites and fusion to metasomal segment I. All the specimens with anomalous tergites examined here were adult females. Te- ruel (2004) observed the same pattern but with four females and one male, and the spec- imen referred by Armas (1974) is an adult male. I have examined 226 specimens of B. coriaceus (57 females, 62 males and 107 ju- veniles), and found this kind of anomaly only on the two females referred to here (Mattoni 2003). Despite the few cases reported in scor- pions, the presence of tergal malformations only on adults suggests that they arise during last molt. The causes of these tergal malformations are completely unknown, but they do not seem to affect the life of the scorpion or its mating. The pigmentation pattern in the ter- gites of all the specimens appears not to be altered. Sexual malformations. — Brachistosternus (L.) pentheri: The specimen has intermediate sexual characteristics: the small size and num- ber of pectinal teeth (26/30, left/right) suggest that it is from a female, males posess larger and more numerous teeth (in the B. pentheri population from Nacunan the females usually have 25-32 teeth, and the males 32-41, Roig Alsina & Maury 1984); the telson is also more similar in shape to that of a female; the ped- ipalp chela has almost all the morphometric characteristics of a male, but the internal apophysis near the base of the movable finger (a sexual secondary structure present only on the males) is extremely reduced to approxi- mately half of normal size (Figs. 4-6); the carapace surface is densely covered with blunt granules, as in regular males (females display fewer granules), but the tergites are smooth as MATTONI— ANOMALIES IN SCORPIONS 625 Figures 4-9. — Right pedipalp chelae, ventrointernal view. 4-6. Bmchistostermis pentheri; 4. male; 5. intersexual specimen; 6. female. 7-9. Bothriurus araguayae; 7. male; 8. intersexual specimen; 9. female. Scale = 1 mm. The arrows show the secondary sexual structures. in females (males display a fine granulation); sternites I-III present sparse granulation, and IV and V are smooth (on more typical males, all sternites are granular, whereas those of fe- males are smooth); and the ventral surfaces of metasomal segments I to III are smooth as ob- served in females (these surfaces are granular on males). The specimen presents a paired 626 THE JOURNAL OF ARACHNOLOGY glandular surface on the dorsal side of meta- somal segment V (another sexual secondary structure of males, Cekalovic 1973; Peretti 1997), but these glands are less well devel- oped, being shorter than in regular males. The specimen also has well developed hemisper- matophores, with sperm in the seminal duct, and testis tubules present; female genitalia (ovari-uterus, seminal receptacles and genital atrium) could not be found. The presence of male, seemingly function- al, sexual organs suggests that this specimen is an adult male with intersexual external mor- phology. Some of these characters are similar to those observed by Maury (1983) in a her- maphrodite B. pentheri male: poorly devel- oped apophyses on the pedipalp chela, re- duced dorsal glands on metasomal segment V, intermediate granulation on the carapace and tergites. The specimen described here presents more feminine external characters, like the pectines and telson, than those described by Maury (1983). The main difference between these specimens is the simultaneous presence of well developed embryos and hemisperma- tophores in Maury’s specimen, which identi- fies it as a true hermaphrodite. Bothriurus araguayae: The specimen ex- hibits many external characteristics of a fe- male: smooth carapace and tergites (males typically present a fine and even granulation); absence of a sexual secondary gland on the dorsal side of the telson (present in adult males); metasomal segment V more robust than males, (which have slender segments, Lourengo & Maury 1979); and without an apophysis on the internal surface of the chela, behind the movable finger, that is present in the males, and is replaced in this specimen by a blunt granule (more pronounced than in reg- ular females) (Figs. 7-9). The male character- istics of the specimen are as follows: both hemispermatophores present, pectines with larger and with more pectinal teeth, and gen- ital operculum formed by two triangular isos- celes plates (these are equilateral in females). I could not observe sperm in the seminal ves- icles, because of the poor preservation of the reproductive organs. As in the previous case, 1 regard the B. araguayae specimen as a male, with intersexual external characteristics. The main differences between the intersex- ual specimens of both species are related to secondary sexual structures: the internal apophysis on the chela, which is absent in the B. araguayae specimen, and present, but re- duced, in the B. pentheri specimen; and the metasomal glands, which are absent on the B. araguayae, and present, but reduced, in the B. pentheri. These secondary sexual structures have a clear function during mating: the apophyses on the male chelae help to secure the female chelae during mating (Maury 1975; Peretti 1993), and the metasomal glands produce a secretion that reduces female resistance during mating (Peretti 1997). Further observations are necessary to understand the incidence of such developmental anomalies in scorpion populations, and the influence that they might have on life history (e.g., in reproductive bi- ology), because of possible disadvantages of intersexual males in comparison with normal males. The cause of these mutations among scor- pions is unknown. Among other arthropods, intersexual specimens have been demonstrat- ed to be the result of bacterial infection (Bou- chon et al. 1998; Rigaud & Juchault 1998). We suspect that the intersexual phenomenon is not limited to the species described here but has been largely ignored in other species in which it may occur. Many external characters are widely used for determining the gender of scorpions, but only the dissection of a speci- men can unequivocally confirm its sex, there- by allowing the identification of intersexual and hermaphrodite specimens. Also, one anomalous specimen can lead to a mistake, that was the case with the “male” of Alayo- tityus juraguaensis described by Armas (1984), a specimen with external female char- acteristics and with one paraxial organ, that led to Armas to say that this was the only species on the genus without sexual dimor- phism. But in fact, as later discovered by Ar- mas (1990), the specimen was a hermaphro- dite, with one hemispermatophore and ovari-uterus. I am indebted to Andres Ojanguren-Affilas- tro for the information about the specimen of B. noa, to Glauco Machado for the donation of several specimens of B. araguayae, to Er- ich Volschenk for the help with the language, and to the curators of the collections from which material was loaned for study: Lorenzo Prendini (AMNH), Cristina Scioscia (MACN- Ar) and Luis Acosta (CDA). This note was MATTONI— ANOMALIES IN SCORPIONS 627 Table 1. — Reported cases and kind of anomalies in families of scorpions. The number of registered species showing the anomaly is in parentheses. The taxonomy presented in the table in accordance with Fet et al. (2000). However, see Stockwell (1989), Prendini (2000) and Soleglad & Fet (2003) for alternative hypotheses. Anomaly Family and species Main references Duplication of metasoma Euscorpiidae (2) Buthidae (8) Sissom & Shelley 1995 Vachon 1952; Sisson & Shelley 1995 Tergite division and/or fusion Diplocentridae (3) Bothriuridae (3) Buthidae (1) Euscorpiidae (1) Armas 1976; Teruel 2004 This work Teruel 2004 Teruel 2004 Leg malformation Buthidae (4) Armas 1977 Pedipalp chela compression on females Bothriuridae (2) Buthidae (18) Chactidae (1) Chaerilidae (1) Diplocentridae (3) Euscorpiidae (3) Hemiscorpiidae (1) Liochelidae (1) luridae (1) Scorpionidae (1) Teruel 2004 Pedipalp fusion Buthidae (1) Cao & Solorzano 1991 Males with intersexual characters Bothriuridae (2) Maury 1983; this work Males with one paraxial organ Bothriuridae (1) Peretti 2000 Hermaphrodite Bothriuridae (1) Buthidae (2) Bothriuridae (1) Maury 1983 Matthiesen 1968, Armas 1990 Maury 1983 Gynandromorphy Buthidae (1) Diplocentridae (2) Armas 1990 Cokendolpher & Sissom 1988 highly improved by suggestions by Alfredo Peretti, Lorenzo Prendini and Erich Vol- schenk. I would also like to extend a special thanks to Gail Stratton for her help during the review process and for assistance with the fig- ures. This research was conducted at Catedra de Diversidad Animal I, Facultad de Ciencias Exactas, Efsicas y Naturales, Universidad Na- cional de Cordoba, Argentina; with the finan- cial support of the Consejo Nacional de In- vestigaciones Cientificas y Tecnicas (Argentina). LITERATURE CITED Armas, L.E 1976. Escorpiones del archipielago cu- bano. 6. Familia Diplocentridae (Arachnida: Scorpionida). Poeyana 147:1-35. Armas, L.E 1977. Anomalfas en algunos Buthidae de Cuba y Brazil. Poeyana 176:1-6. Armas, L.E 1984. Escorpiones del Archipielago Cubano. VII. Adiciones y enmiendas (Scorpi- ones: Buthidae, Diplocentridae). Poeyana 275:1- 37. Armas, L.F 1990. Dos casos de anomalia sexual en escorpiones cubanos (Scorpiones: Buthidae). Ciencias Biologicas 21-22:173-175. Bouchon D,, T. Rigaud & P. Juchault. 1998. Evi- dence for widespread Wolbachia infection in iso- pod crustaceans: molecular identification and host feminization. Proceedings of the Royal So- ciety of London, Biological Sciences 265(1401): 081-1090. Cao, J. & L. Solorzano. 1991. Escorpion con pe- dipalpo anomalo. Resumenes II Simposio de Zoologia, La Habana, p. 48. Cekalovic, K.T 1973. Nuevo caracter sexual secun- dario en los machos de Bmchistostennis (Scor- piones, Bothriuridae). Boletln de la Sociedad de Biologia de Concepcion 46:41-51. Cokendolpher, J.C. & W. D. Sissom. 1988. New gynandromorphic Opiliones and Scorpiones. 628 THE JOURNAL OF ARACHNOLOGY Bulletin of the British Arachnological Society 7(9);278-280. Fet, V., W.D. Sissom, G. Lowe & M.E. Braunwald- er. 2000. Catalog of the scorpions of the world (1758-1998). New York Entomological Society, New York. pp. 1-670. Hjelle, J.T. 1990. Anatomy and morphology. Pp. 9- 63. In The Biology of Scorpions (G.A. Polis, ed.). Stanford University Press, Stanford, Cali- fornia. Kovarfk, F. 1997. Revision of the genera Lychas and Hemilychas, with descriptions of six new species (Scorpiones, Buthidae). Acta Societatis Zoologicae Bohemoslovenicae 61:311-371. Kraepelin K. 1899. Scorpiones und Pedipalpi. In Das Tierreich (F. Dahl, ed.). Friedlander und Sohn Verlag, Berlin, 8:1-265. Lourengo, W.R. & E.A. Maury. 1979. Quelques considerations sur la systematique du scorpion bresilien Bothriurus araguayae Vellard, 1934 (Bothriuridae). Bulletin du Museum National d’Historie Naturelle, Paris (Zoologie, Biologie et Ecologie Animales) 2:421-431. Matthiesen, FA. 1968. On the male reproductive organs in some brazilian scorpions. Revista Bras- ileira de Pesquisas Medicas e Biologicas 1(5-6): 273-274. Mattoni, C.I. 2003. Patrones evolutivos en el genero Bothriurus (Scorpiones, Bothriuridae): analisis filogenetico. Doctoral Thesis, Universidad Na- cional de Cordoba, Argentina, i-vii + 249 pp. Maury, E.A. 1975. Sobre el dimorfismo sexual de la pinza de los pedipalpos en los escorpiones Bothriuridae. Bulletin du Museum National d’Histoire Naturelle, Paris, Zoologie 215:765- 771. Maury, E.A. 1983. Singular anomalia sexual en un ejemplar de Brachistosternus pentheri Mello- Leitao 1931 (Scorpiones, Bothriuridae). Revista de la Sociedad Entomologica Argentina 42(1-4): 155-156. Peretti, A.V. 1993. Estudio de la biologia reprod- uctiva en escorpiones argentinos (Arachnida, Scorpiones): un enfoque etologico. Doctoral Thesis, Universidad Nacional de Cordoba, Ar- gentina. i-ix + 307 pp. Peretti, A.V. 1997. Relacion de las glandulas cau- dales de machos de escorpiones Bothriuridae con el comportamiento sexual (Scorpiones). Revue Arachnologique 12(3):31-41. Peretti, A.V. 2000. Existencia de cortejo en el cam- po de machos de Bothriurus bonariensis (Scor- piones: Bothriuridae) que carecen de un organo paraxial. Revista de la Sociedad Entomologica Argentina 59(l-4):96-98. Prendini, L. 2000. Phylogeny and classification of the Superfamily Scorpionoidea Latreille 1802 (Chelicerata, Scorpiones): an exemplar approach. Cladistics 16:1-78. Rigaud T. & P. Juchault. 1998. Sterile intersexuality in an isopod induced by the interaction between a bacterium (Wolbachia) and the environment. Canadian Journal of Zoology 76:493-499. Roig Alsina, A.H. & E.A. Maury. 1984. Sistematica y distribucion geografica de Brachistosternus (L. ) pentheri Mello-Leitao, 1931 (Scorpiones, Bothriuridae). Physis (C) 42(102): 17-21. Sissom, W.D. 1990. Systematics, biogeography and palentology. Pp. 64-160. In The Biology of Scorpions (G.A. Polis ed.). Stanford University Press, Stanford, California. Sissom, W.D. & R.M. Shelley. 1995. Report on a rare developmental anomaly in the scorpion, Centruroides vittatus (Buthidae). Journal of Ar- achnology 23:199-201. Soleglad, M.E. & V. Fet. 2003. High-level system- atics and phylogeny of the extant scorpions (Scorpiones: Orthosterni). Euscorpius 11:1-175. Stockwell, S.A. 1989. Revision of the Phylogeny and Higher Classification of Scorpions (Cheli- cerata). Ph.D. Thesis, University of California, Berkeley, i-iv, 1-413 pp. Teruel, R. 2004. Nuevos casos de anomalias mor- fologicas en escorpiones (Scorpiones: Bothriuri- dae, Buthidae, Chactidae, Chaerilidae, Diplocen- tridae, Euscorpiidae, Hemiscorpiidae, Ischnuridae, luridae, Scorpionidae). Revista Ib- erica de Aracnologfa 7:235-238. Vachon, M. 1952. Etudes sur les scorpions. Ar- chives de ITnstitut Pasteur d’Algerie, Alger. 482 pp. Vachon, M. 1963. De Putilite, en systematique, d’une nomenclature des dents des cheliceres chez les Scorpions. Bulletin du Museum National d’Historie Naturelle, Paris 35(2): 161-166. Manuscript received 31 December 2004, revised 15 August 2005. 2005. The Journal of Arachnology 33:629-639 MODELING OF THE STRESS-STRAIN BEHAVIOR OF EGG SAC SILK OF THE SPIDER ARANEUS DIADEMATUS Els Van Nimmen and Kris Gellyeck: Department of Textiles, Ghent University, Technologiepark 907, B-9052 Zwijnaarde, Belgium. E-mail: Els.VanNimmen@ UGent.be Tom Gheysens: Department of Biology, Ghent University, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium Lieva Van Langenhove: Department of Textiles, Ghent University, Technologiepark 907, B-9052 Zwijnaarde, Belgium Johan Mertens: Department of Biology, Ghent University, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium ABSTRACT. Spider silk has attracted the attention of many scientists because of its desirable physical properties. Most of this attention has been devoted to dragline silk, a thread that has high tensile strength, high strain and ultra-low weight. To help understand structure-property relationships in spider silks, the tensile behavior of egg sac (cylindrical gland) silk of Araneus diadematus Clerck 1757 was compared with dragline (major ampullate gland) and silkworm silks. In addition, stress-strain curves of egg sac silk were simulated by a spring-dashpot model, specifically a Standard Linear Solid (SLS) model. The SLS model consists of a spring in series with a dashpot and in parallel with another spring, resulting in three unknown parameters. The average stress-strain curve of fibers from five different egg sacs could be accurately described by the model. Closer examination of the individual stress-strain curves revealed that in each egg sac two populations of fibers could be distinguished based on the parameters of the SLS model. The stress-strain curves of the two populations clearly differed in their behavior beyond the yield point and were probably derived from two different layers within the egg sac. This indicates that silks in the two layers of A. diadematus egg sacs probably have different tensile behavior. Keywords: Spider silk, tensile behavior, cocoon, cylindrical gland, tubuliform gland, Araneidae Spider silk has attracted considerable atten- tion as a natural fiber in the last 10 years be- cause spider silk, especially dragline silk, shows a unique combination of high strength, high strain and extreme fineness. The silk pro- duced by orb-web-weaving araneid spiders provides ideal material for studying the rela- tionships between molecular structure and me- chanical properties for protein-based structur- al materials. Araneid spiders have seven different gland-spinneret complexes, each of which synthesizes a unique blend of structural polymers and produces a fiber with a unique set of functional properties. An overview of the different spider silks of Araneus diade- matus Clerck 1757, their glands, their function and amino-acid composition is provided in Ta- ble 1. Spiders produce silks that range from Ly- cra-like elastic fibers to Kevlar-like superfi- bers, but it is not known how spiders modulate the mechanical properties of silks. Table 2 gives an overview of the tensile properties of spider silks and some other biological and en- gineering materials. The spider silks that have been most studied are products of the major ampullate (MA) glands. The tensile strength (or a measure of the force needed to break a material) of MA silk is clearly higher than other polymeric bio- materials such as tendon collagen and bone as can be seen in Table 2. Moreover, because of its much higher strain to break value or ex- tensibility, its toughness (as indicated by the work to rupture value in Table 2) or the en- ergy required to break spider silk can be ten 629 630 THE JOURNAL OF ARACHNOLOGY Table 1. — Types and functions of spider silk for Araneus diadematiis (Andersen 1970, Kaplan 1998). Small side chains for amino acids include glycine (Gly) + alanine (Ala) + serine (Ser) — polar = aspartic acid + threonine + serine + glutamic acid + tyrosine + lysine + histidine + arginine. Silk Gland Function Amino-Acids Dragline Major ampullate Orb web frame, radii, dragline Gly (37%), Ala (18%), small side chains (62%), polar (26%) Viscid Flagelliform Prey capture, sticky spiral Gly (44%), Pro (21%), small side chains (56%), polar (17%) Glue-like Aggregate Prey capture, attachment to sticky spiral Gly (14%), Pro (11%), polar glue (49%), small side chains (27%) Minor Minor ampullate Orb web frame, bridging lines Gly (43%), Ala (37%), small side chains (85%), polar (26%) Egg sac Cylindrical (tubuli- form) Reproduction Ser (28%), Ala (24%), small side chains (61%), polar (50%) Wrapping Aciniform Wrapping captured prey Ser (15%), Gly (13%), Ala (11%), small side chains (40%), polar (47%) Attachment Piriform Attachment to environ- mental substrates Ser (15%), small side chains (32%), polar (58%) times greater than that of other biological ma- terials. Since initial modulus (for definition see Table 2) is a measure of stiffness, it is fair to say that spider MA silk is amongst the stiff- est and strongest polymeric biomaterials known. However, the initial modulus or stiff- ness of MA silk is well below that of Kevlar, carbon fiber and high-tensile steel, engineer- ing materials that are commonly employed to transmit and support tensile forces. Note also that the strength of MA silk is somewhat less than that of these engineering materials. Nev- ertheless, MA silk is still tougher than these engineering materials because of its large ex- tensibility. See Table 2 for a summary of def- initions concerning mechanical properties. The viscid silk (Gosline et al. 1994) that forms the glue-covered catching spiral, is an- other truly remarkable spider silk material. Its initial modulus or stiffness is three orders of magnitude lower than that of MA silk and is comparable with that of a lightly cross-linked rubber. With a maximum strain of approxi- mately 270%, viscid silk is not exceptionally stretchy compared to other rubbery materials, but its strength, at approximately 0.5 GPa, makes viscid silk roughly ten times stronger than any other natural or synthetic rubber. Of all the silks, MA silk has been an object of desire for materials engineers because of its extreme performance properties, particu- larly its strength. Investigators have already been searching for more than 15 years to pro- duce “synthetic” dragline silk in quantities sufficient for applications such as bullet-proof vests, parachute cords, surgical sutures and substitutes for ligaments. However, commer- cial production of “synthetic” MA silk is still not possible. We have focused on the me- chanical and structural properties of spider silk of the egg sac, which to this point, is not well studied. We believe that it is precisely through correlating chemical, microstructural and consequent property differences between silks that knowledge of how the spider con- trols the fiber function will be acquired. Egg sac silk is secreted by the cylindrical (= tubuliform) glands. At any point along its length, the egg sac fiber must be able to bend easily in one plane but otherwise resist bend- ing and stretching. As reported by Barghout et al. (2001), these mechanical properties are imparted by a multiaxial anisotropic micro- structure that is not observed for MA silk. Barghout et al. (1999) also observed the pres- ence of non-periodic lattice crystals identified previously in the MA silk of Nephila clavipes Linnaeus 1767 (Thiel et al. 1997). Moreover, they found that these crystals in A. diadema- tus Clerck 1757 egg sac silk are twisted par- allel to the chain direction in contrast to what is found for MA silk. This is suggested to be the reason for the lower stiffness that is found for A. diadematus egg sac silk compared to MA silk (Stauffer et al. 1994). Stauffer et al. (1994) compared the physical VAN NIMMEN ET AL.— STRESS-STRAIN BEHAVIOR OF EGG SAC SILK 631 Figure 1. — The standard linear solid model that was used to simulate the stress-strain behavior of egg sac silk of Araneus diadematus. properties of three silks (secreted by the major ampullate, minor ampullate and cylindrical glands) from N. clavipes and Araneus gem- moides Chamberlin & Ivie 1935. Comparing silks within each species, they concluded that major ampullate silk is substantially stronger than either of the other two silks. Egg sac silk is next, followed closely by minor ampullate silk. The strain of these different silks seemed comparable. The dominant, repeated crystallizable mo- tifs in egg sac silk of A. diadematus are sim- ilar to the motifs that form p-sheet crystals in MA silk spun by N. clavipes (Guerette et al. 1996; Thiel et al. 1997). The number of times these motifs are repeated for Araneus egg sac silk are however somewhat smaller than the corresponding values for Nephila MA silk. From the materials science viewpoint it is ex- pected that similar primary structures at the molecular level will lead to similar ordering schemes at microstructural scales. This view- point is axiomatic in our use of egg sac silk to obtain further insights into the structure of MA silk. Working with Araneus egg sac silk offers a significant advantage relative to work- ing with MA silk: useful amounts are pro- duced in a convenient (compact) form. In a previous study (Van Nimmen et al. 2003), the effects of UV-light and humidity on the stress-strain properties of egg sac silk of A. diadematus were demonstrated. Another study (Van Nimmen et al. 2004) considered the effect of strain-rate on the tensile proper- ties of egg sac silk of A. diadematus. The aim of the present study was to inves- tigate how the stress-strain behavior of egg sac silk compared with the behavior of drag- line silk and cocoon silk obtained from silk- worms. We expected that spider egg sac and silkworm cocoon silks would have similar tensile properties because they serve similar functions (providing shelter and protection). Attention was focused on the shape of the stress-strain curves. Mechanical properties are often character- ized only by breaking force, breaking strain Table 2. — Tensile mechanical properties of spider silks and other materials as derived from the literature (Gosline et al. 1999; Denny 1976). Initial modulus is defined as the modulus in the elastic range of the diagram in which strain changes are still reversible, it is usually calculated from the slope of the initial elastic region of the force-strain curve, also the term stiffness is used; strength (or tensile strength) is a measure for the breaking force or the force required to break the material; strain to break is the increase in length of a specimen produced by the breaking force, usually expressed as a percentage of the original length; toughness is a measure of the required energy to break a material and is calculated as the area contained by the force-strain curve up to the breaking point, often indicated as the work to rupturevalue. Material Initial modulus (GPa) Strength (GPa) Strain to break (%) Work to rupture (MJ m-3) Araneus MA silk 10 1.1 27 160 Araneus viscid silk 0.003 0.5 270 150 Bombyx mori silk 7 0.6 18 70 Tendon collagen 1.5 0.15 12 7.5 Bone 20 0.16 3 4 Elastin 0.001 0.002 150 2 Resilin 0.002 0.003 190 4 Synthetic rubber 0.001 0.05 850 100 Kevlar 49 130 3.6 2.7 50 Carbon 300 4 1.3 25 High-tensile steel 200 1.5 0.8 6 632 THE JOURNAL OF ARACHNOLOGY Dragline — Egg sac B. mori — A. pernyi Strain (%) Figure 2. — The average stress-strain curves of different silks as measured by a single-strength tester (gauge length 20 mm, testing speed 20 mm/min) based on 169 tests of Araneiis diadematiis dragline silk, 403 tests of A. diadematiis egg sac silk, and 49 tests each of Bombyx mori and Antheraea pernyi silk. and initial modulus. However, we are also in- terested in the time-dependent behavior that is also partly included in the stress-strain curves. In this study, visco-elastic models, based on spring-dashpots, are used to simulate the stress-strain behavior for spider egg sac silk. This will help to relate the mechanical and visco-elastic characteristics to the structural properties that will be investigated in further research. Finally, because of the high vari- ability that was noted for the tensile properties within each egg sac, a cluster analysis was performed in order to find out if different fiber populations or layers could exist within an egg sac. METHODS General methods. — Five egg sacs of Ara- neiis diadematiis Clerck 1757 were collected in a bower in Belgium (Merelbeke, 51° north latitude and 3° east longitude) in autumn. One of these A. diadematiis spiders with her egg sac is deposited as a voucher specimen in the “Zoology Museum” (UGMD 104091), Ghent University in Belgium. Since the egg sacs were collected in their natural habitat, we expected that the measured mechanical behavior would better represent the real characteristics than if they were pro- duced by lab-reared spiders. The egg sacs were removed shortly after oviposition. After removing the clearly visible outer cover, one hundred fibers were gently removed at ran- dom from the inside of each egg sac, with care taken to stress the fibers as little as possible. For the dragline samples, some A. diade- matiis were reared in the laboratory and from thirty spiders a sample of dragline thread was manually reeled off as spiders hung freely sus- pended in space. From every sample, ten fi- bers were prepared and tested. Fibers were also tested from cocoons of the silkworms Bombyx mori and Antheraea pernyi (Tussah silk), grown at the Silk Museum of Meliskerke (The Netherlands). Since the sam- ples we obtained were already a thorough blend of fibers of different cocoons, we decid- ed to reduce the number of tests to 50 for both silks. All samples were kept in a conditioned laboratory of 20 °C ± 2 °C and relative hu- midity of 65 ± 2% for at least 24 hours before testing. The FAVIMAT-ROBOT (Textechno) was used to analyze the tensile properties of the egg sac, cocoon and dragline fibers. It is a semi-automatic single fiber strength tester, working according to the principle of constant rate of extension (standards: DIN 51221, DIN 53816, ISO 5079). The instrument is equipped VAN NIMMEN ET AL.— STRESS-STRAIN BEHAVIOR OF EGG SAC SILK 633 with a balance allowing the mass to be mea- sured at a high resolution of 0.1 mg. More- over, this instrument includes an integrated measuring unit for linear density i.e.; mass per unit length, expressed in dtex, which equals decigrams per kilometer. This measure has the considerable advantage that the linear density, a measure for fineness, is determined simul- taneously with the tensile properties. This is particularly advantageous for natural fibers. The linear density is measured according to the vibroscopic method (ASTM D 1577— BIS- FA 1985/1989 chapter F). Because of the extreme fineness of dragline thread, it was unfortunately not possible to si- multaneously determine the linear density of the dragline fibers. Instead, diameters of these fibers (in pm) were measured on a large num- ber of samples with image analysis on a light microscope and the conversion was made to dtex taking into account a specific density of 1.3 g/cm^ as reported in the literature (Vollrath & Knight 2001). The tensile properties were tested in stan- dardized conditions of 20 ± 2 °C and relative humidity of 65 ± 2 % with a gauge length of 20 mm, a test speed of 20 mm/min, and a pre- tension of 0.05 cN/dtex. For the linear density, a test speed of 5 mm/min and a pre-tension of 0.08 cN/dtex were applied. Visco-elastic models. — The Maxwell mod- el: The stress-strain curve of polymers is often mathematically described by models indicat- ing the visco-elastic behavior of these poly- mers. When a material is extended by an ap- plied force, there is, besides the elastic component, a further component whose action opposes the applied force but whose magni- tude depends on the speed of extension. This second component decays relatively slowly with time. When the applied force is subse- quently removed, the same component also acts to resist the internal elastic forces that bring about contraction. This time dependency of polymers is also indicated as visco-elastic- ity (Saville 1999). Their behavior is fitted by a visco-elastic model as the relationship be- tween the applied stress and resultant strain contains a time-dependent element. Most visco-elastic models consist of a com- bination of springs and dashpots. The spring represents the elastic solid-like behavior where Hooke’s law is valid (F = Ee where F is load or force, E is elastic modulus and e is strain), whereas the dashpot represents the time-dependent, viscous liquid-like behavior where Newton’s law is valid (F = Ti(de/dt) where t| is the viscosity or damping constant). In the simplest Maxwell-model (Tobolsky et al. 1951), the visco-elastic behavior of a fiber (or yarn) is described by a spring (with elastic modulus E) and a dashpot (with damp- ing constant or viscosity iq) in series. This be- havior obeys the following equation (with the strain and F the force): d£ 1 dF F dt ~ E dt ^ Ti This model is often used to describe stress- relaxation, a phenomenon that is observed when a polymer is extended by a given amount and then held at that extended length. If the force required to do this is monitored, it is found to rise immediately to a maximum value and then slowly decrease with time. To use this model to describe stress-strain curves in tensile testing, we take into account a constant increase of strain with time, so that we can pose that e = r t, with r a constant. Equation (1) then becomes: F E dt T| (2) with as starting condition F(0) = Fv, where F^ is the preload, from which the following so- lution is obtained: F(e) = F^ + Tir 1 — exp Equation (3) can be written as: F(e) = F^ + A(1 — e“^®) with A = Tir and (3) (4) This equation allows parameters A and B to be estimated by means of a non-linear regres- sion. The standard linear solid model.- An exten- sion of this Maxwell model is the so-called standard linear solid (SLS) model, where a linear spring in parallel is added (Fig. 1). Taking into account this spring in equation (2) and by differentiating, equation (4) can then be written as follows: F(e) = F^ + A(1 — e“^®) + C-e with A = iqr and B = — and C = E^ (5) Tir 634 THE JOURNAL OF ARACHNOLOGY Egg sac 1 Egg sac 3 Strain (%) Egg sac 5 Strain (%) Figure 3. — Simulation by means of the standard linear solid model for two statistically different fiber populations found within an egg sac by means of a cluster analysis. The parameters A, B and C can then be esti- mated by means of non-linear regression. The Voigt model: Another time-dependent phenomenon is creep. If instead of a fixed ex- tension, a fixed force is applied to the mate- rial, an initial extension of a magnitude is found that is expected from the force-strain curve followed by a further slow extension with time. For the description of creep or ten- sile testing under constant increase of load. VAN NIMMEN ET AL.— STRESS-STRAIN BEHAVIOR OF EGG SAC SILK 635 Table 3. — The average values (Mean) and the standard deviations (SD) of the parameters A, B and C of the SLS model for the 5 egg sacs of Araneus diademcitus for 2 statistically different fiber populations (“1” and “2”) as found by means of a cluster analysis {n — number of fibers within each population). A B c Mean SD Mean SD Mean SD n Egg sac 1 1 1.84 0.14 0.46 0.05 0.015 0.010 52 2 2.39 0.20 0.36 0.06 -0.006 0.021 14 Combined 1.95 0.28 0.44 0.06 0.010 0.015 66 Egg sac 2 1 1.75 0.05 0.50 0.05 0.013 0.003 33 2 1.58 0.04 0.51 0.07 0.012 0.001 27 Combined 1.67 0.09 0.50 0.06 0.012 0.002 60 Egg sac 3 1 1.70 0.07 0.42 0.04 0.011 0.002 31 2 1.72 0.15 0.54 0.03 0.012 0.003 24 Combined 1.71 0.11 0.47 0.07 0.011 0.003 55 Egg sac 4 1 1.72 0.11 0.45 0.07 0.013 0.002 43 2 1.48 0.11 0.56 0.09 0.016 0.004 29 Combined 1.62 0.16 0.50 0.10 0.014 0.004 72 Egg sac 5 1 1.28 0.11 0.58 0.06 0.021 0.007 19 2 1.50 0.13 0.47 0.05 0.010 0.003 48 Combined 1.44 0.16 0.50 0.07 0.013 0.007 67 the simplest model used is the Voigt model. This model consists of a spring (elastic con- stant E) in parallel with a dashpot (with damp- ing constant t]). The visco-elastic behavior is then described by the following differential equation (with e the strain and F the force): de F = Ee + Ti— (6) dt Using the correct starting conditions for creep or tensile testing under constant increase of load, solutions for this equation can be found. Since these are not valuable for this study, the reader is referred to the literature (Saville 1999). Other visco-elastic models: The models de- scribed above can be extended to more ele- ments, such as the “four-elements modeT’ consisting of a Maxwell-element in series with a Voigt element or more generalized Maxwell and Voigt models considering a fi- nite or infinite number of Maxwell or Voigt elements connected in parallel or in series. Since it is beyond the scope of this study, the reader is again referred to the literature for further description (Saville 1999). RESULTS The tensile behavior of silks. — First, it should be remarked that although 500 egg sac fibers and 300 dragline fibers were tested, not all were successful mostly due to the fineness of the fiber. For the calculation of the average stress-strain curves, for which the shape is the most important, only curves with strain to break values higher than 10% were consid- ered. The curves were stopped at the average strain to break values of all available tests. It can be expected that the measurements show a small error since probably the weakest fibers could not be tested. However, from the his- togram of the strength values, the contribution of stronger fibers is not higher than that of the weaker fibers In addition, the high variability in the stress-strain curves among the different egg sacs should be noted, which can also be found in the literature on dragline silks (Mad- sen et al. 1999; Garrido et al. 2002). 636 THE JOURNAL OF ARACHNOLOGY Fig. 2 shows the average stress-strain curves of the different silks of A. diadematus (dragline, egg sac), B. mori and A. pernyi. It is clear that egg sac silk shows a completely different stress-strain behavior from dragline silk and even the functionally comparable silkworm cocoon silks. All stress-strain curves start with a small elastic region. For the dragline, B. mori and A. pernyi fibers, this region is followed by a plastic region and fi- nally by strain hardening where the stress again linearly increases with strain. However spider egg sac silk shows a plastic-hardening region that is extremely flat. Since in this re- gion the stress increases again linearly with strain, we will simply use the term “hardening region” to indicate this region. Although egg sac silk shows about the same strain to break as dragline silk, the tensile strength of dragline silk is three to four times higher. The initial modulus (calculated from the slope of the initial straight line portion), which is a measure of stiffness of the fiber, is significantly higher for egg sac silk than for dragline thread (67 cN/dtex versus ±47 cN/ dtex) {P < 0.001). Simulation of tensile behavior of egg sac silk. — For this research, the stress-strain data of the five egg sacs were used, from which the average stress-strain curve shown in Fig. 2 was produced. Since we were working with tensile testing with constant increase of exten- sion, the Maxwell-model as described earlier was used to describe the stress-strain behavior. Starting from equation (4), the parameters A and B were estimated by means of a non-lin- ear regression. We concluded that the Max- well-model does not completely satisfy the simulation of the stress-strain curve for the egg sac silk fibers. We then applied the SLS model, in which the 3 parameters A, B and C of equation (5) were estimated by means of a non-linear re- gression. With the average data of the stress- strain curves, for each egg sac a correlation of higher than 99% with a relative error (de- fined as (Fg^pgj.j,^^g,^mi f^piedicted)^f^expen mental) Small- er than 0.1% was observed, except in the ini- tial elastic region where the maximum relative error at about 0.4-0. 5% strain exceeds 0.4% to 1%. To get an indication of the variability within the egg sac, the non-linear regression was re- peated for each of the individual stress-strain curves of the 100 fibers that were tested for each of the five egg sacs. Because of the ob- served high variability, we performed a clus- ter analysis (with the statistical software SPSS) on the estimated parameters A, B and C in order to identify statistically different clusters or fiber populations. The result of this cluster analysis is given in Table 3. Within the different egg sacs, two clusters (indicated as “1” and “2”) of statis- tically different fiber populations could be de- tected. In this analysis, clusters of less than 10 fiber data were removed. The clusters or fiber populations for egg sac 1, egg sac 4 and egg sac 5 show completely different A, B and C values. In other words, the level of the more horizontal hardening region (indicated by A), the shape of the yield (or transition) region (indicated by B) and the slope of the harden- ing region (indicated by C) of their stress- strain curves are significantly different. For egg sac 2, only the A-values of the clusters are significantly different, while the confi- dence regions of the parameters B and C are overlapping. With respect to egg sac 3, the B- values of the clusters are significantly differ- ent, while the confidence regions of the pa- rameters A and C are overlapping. Based on the cluster analysis, the stress- strain curves of the individual fibers from each egg sac were split into 2 groups and the average curve of each group was calculated. These average stress-strain curves based on the two different fiber populations for each egg sac are shown in Fig. 3. It can be con- cluded that the fiber populations seem to differ mostly in the level of the relatively flat so- called hardening region and thus the breaking stress value. The initial modulus and the mod- ulus of the hardening region, i.e. the tangent modulus at the yield point, seem to be quite equal for both fiber populations. DISCUSSION The tensile behavior of silks. — The shapes of the stress-strain curves that we found and that were also seen by Van Nimmen et al. (2004) are similar to those that were found by Stauffer et al. (1994). However, Stauffer et al. 1994 determined different absolute values for strength and strain. As their testing procedures were different from our, it is difficult to eval- uate the discrepancies. They found for Ara- neus gemmoides MA silk final breaking points VAN NIMMEN ET AL.— STRESS-STRAIN BEHAVIOR OF EGG SAC SILK 637 at extensions of about 15 ± 2% (n — 10) with a final stress of 4.7 ± 0.5 GPa and for egg sac silk breaking strains at 19 ± 2% {n = 10) with tensile strengths of 2.3 ± 0.2 GPa. They obtained much higher stress values than found elsewhere for MA silk (see Table 2) because, for diameter measurements, they took into ac- count the ten smallest diameter points in sev- eral sections of the silks. With respect to strain, we found much higher values (30 ± 9%, n = 183) for MA silk and 32% ± 16%, ' n = 398 for egg sac silk), with a much higher variability, probably due to the greater number of tests performed. It is not clear if this dif- ference is due to the difference in testing pro- cedure or to the spider species. However, other published data of MA Araneus silk mention a strain to break value of 27% (Denny 1976) which agrees better with our strain data. In order to make further comparisons possible with the tensile properties presented in Table 2, our breaking stress and stiffness values were converted to the GPa unit, taking into account a specific density of 1.3 gicvn? (VolL rath & Knight 2001). The breaking stress val- ues thus obtained were 0.94 ± 0.36 GPa {n = 183) for MA silk and 0.27 ± 0.05 GPa {n = 398) for egg sac silk. The stiffness values, calculated from the slope of the initial elastic region, resulted in values for MA silk of 6.1 ± 2.4 GPa (n = 167) and for egg sac silk of 8.7 ± 0.9 GPa (n = 434). The stiffness value for MA silk seems low compared to the value of 10 GPa that is given in Table 2. Probably the testing condi- tions play a role in this difference (forced or unforced silking, single or multifilament, cli- mate, strain rate, gauge length, etc). Denny’s (1976) analysis of the strain-rate dependence of MA silk demonstrated that the initial stiff- ness increases from 9.8-20.5 GPa when the strain rate is increased from 0.0005 s“‘ to 0.024 s“f Also the spinning conditions (e.g. drawing speed, body temperature) have been reported to affect the tensile properties (Voll- rath et al. 2001). We believe the different stress-strain behav- ior of dragline and egg sac silk is partly due to different amino acid compositions. Glycine (Gly) and alanine (Ala) are most abundant in draglines, while serine (Ser) and Ala are most abundant in egg sac silk (Table 1). Moreover, the proline rich motif Gly-Pro-Gly-X-X oc- curs in dragline silk but not in egg sac silk (Guerette et al. 1996; Gosline et al. 1999). Thiel et al. (1997) believe that the structure of the proline residue forces a severe kink in an extended backbone chain. On the other hand, the total content of the small amino acids Gly, Ala and Ser, which is usually taken as an in- dication of crystal forming potential (Gosline et al. 1986), is almost the same for dragline and egg sac silk (Table 1). Thus, we would expect the crystallinity of both fibers to be similar. However, in tensile testing, the weak- est regions, i.e. the more amorphous regions, most affect the stress-strain behavior. Conse- quently, two silks with similar crystallinity may exhibit dissimilar tensile properties. Thus, the different stress-strain curves of MA and egg sac silk are probably more a reflection of differences in the arrangement (chain lengths, number of coils, etc.) of the structural elements of the amorphous regions than of the crystalline domains. Since glycine is the simplest amino-acid (side group H), while serine is an amino-acid with a much more voluminous side group (CH2OH), the difference in strength between dragline and egg sac silks may be mainly at- tributed to the more compact structure which can be built with glycine, resulting in a struc- ture that is more resistant to stress. Although the structure of the glycine-rich regions of MA silk is imperfectly understood, there is consensus that these regions are part of a more oriented amorphous phase (Jelinski et al. 1999; van Beek et al. 2002). Moreover, the proline-rich regions in MA silk are expected to include more turns, resulting in a higher number of hydrogen bonds and thus in a more stress resistant structure. A more intensive study of the spinning process, structure and morphology of spider silk, especially egg sac silk, is required to further explain the differ- ence in tensile behavior. We also note that the shapes of the stress- strain curves obtained for the silkworm silks are more similar to dragline silk than to egg sac silk, even though the silkworm and spider use the former two silks for completely dif- ferent functions. Since the main constituents of B. mori and A. pernyi silks are also glycine and alanine (44% Gly, 29% Ala, 12% Ser in B. mori and 27% Gly, 43% Ala, 11% Ser in A. pernyi (Kishore et al. 2002)), the higher similarity in behavior to dragline silk could be expected. 638 THE JOURNAL OF ARACHNOLOGY Simulation of tensile behavior of egg sac silk. — The different fiber populations vary mostly in the hardening region, that is, the re- gion beyond the yield point. The initial elastic region, and the modulus of this region, that is usually used to define the stiffness, appears not to differ for the two fiber populations. As mentioned before, the spring in the SLS mod- el represents the solid character whereas the dashpot indicates the liquid character. By add- ing a (elastic) spring to the Maxwell model, an element is added that results in a linear relation between stress and strain beyond the yield point. The significance of the coefficient C indicates that there is indeed a significant, although small, increase in stress as a function of strain beyond the yield point. During post- yield extension, the long molecules tend to be- come oriented along the stress axis and, as a result, a structure may be obtained which ap- proaches that of a crystalline material. This is, in fact called “strain-induced crystallization” (Wainwright et al. 1976) and leads to a nota- ble increase in the value of the instantaneous elastic modulus. A link with the twisted non- periodic lattice (NPL) crystals demonstrated by Barghout et al. (1999) can be made. The twist of these regions may result in the flat- tened behavior beyond the yield point, i.e. the lower tangent modulus at the yield point, for egg sac silk compared to dragline silk (in which the twist of the NPL crystals is not ob- served). Since the fibers were randomly selected from each egg sac, the two fiber populations can probably be attributed to different layers that constitute the egg sac. Our own prelimi- nary structural research of the egg sac of Ar- aneus diadematus indeed confirms the exis- tence of different layers, especially observed as a slight difference in color and in the stack- ing of the fibres above and below the eggs. Different layers in the egg sac structure are also found for the spider Zygiella x-notata (Gheysens et al. in press). A more detailed study in which an attempt is made to divide the different layers will be required to confirm this. This study has shown that egg sac silk of Araneus diadematus has a completely differ- ent tensile behavior from dragline of the same spider. In contrast to what was expected given the functions of the different silks, more sim- ilarities were found between spider dragline silk and cocoon silks of Bombyx mori and An- theraea pernyi than between the latter and the spider egg sac silk. We suggest that the dif- ference in stress-strain behavior is partly due to the different amino acid composition, and especially the structure of the amorphous do- mains. A further structural and morphological study of egg sac silk is required to further ex- plain its special stress-strain behavior. The stress-strain curve of spider egg sac silk can be accurately simulated by the stan- dard linear solid model with 3 parameters to be estimated. A more detailed analysis of the estimated parameters A, B and C revealed that for each egg sac two clusters or populations of fibers could be found, mostly differing in the stress level of the region beyond the yield point. Since the fibers were taken randomly from each egg sac, it is suggested that the dif- ferent behavior of the two fiber populations is due to the different tensile behavior of two layers constituting an egg sac. A further study will be required to relate the mechanical prop- erties to the functions of these different layers. LITERATURE CITED Andersen, S.O. 1970. Amino Acid Composition of Spider Silks. Comparative Biochemistry and Physiology 35:705-711 Barghout, J.Y.J., B.L. Thiel, C. Viney. 1999. Spider {Araneus diadematus) cocoon silk: a case of non- periodic lattice crystals with a twist? Internation- al Journal of Biological Macromolecules 24: 211-217. Barghout, J.Y.J., J.T Czernuszka, C. Viney. 2001. Multiaxial anisotropy of spider (Araneus diade- matus) cocoon silk fibres, Polymer 42:5797- 5800. Denny, M. 1976. The physical properties of spider’s silk and their role in the design of orb-webs. Journal of Experimental Biology 65:483-506. Garrido, M.A., M. Elices, C. Viney, J. Perez-Ri- gueiro. 2002. The variability and interdepen- dence of spider drag line tensile properties. Poly- mer 43:4495-4502. Gheysens T, L. Beladjal, K. Gellynck, E. Van Nim- men, L. Van Langenhove, J. Mertens. In press. Cocoon construction of the Zygiella x-notata (Arachnida, Araneidae). Journal of Arachnology. Gosline, J.M., M.E. DeMont, M.W. Denny. 1986. The structure and properties of spider silk. En- deavour, New Series 10:37-43. Gosline, J.M., C.C. Poliak, P.A. Guerette, A. Cheng, M.E. Demont, M.W. Denny. 1994. Elastomeric Network Models for the Frame and Viscid Silks from the Orb Web of the Spider Araneus diade- matus. Pp. 328-341. In Silk Polymers: Material VAN NIMMEN ET AL.— STRESS-STRAIN BEHAVIOR OF EGG SAC SILK 639 Science and Biotechnology. (D. Kaplan, W.W. Wade, B. Farmer. & C. Viney, eds) ACS Sym- posium Series 544, Washington DC. Gosline, J.M., RA. Guerette, C.S. Ortlepp, K.N. Savage. 1999. The mechanical design of spider silks: From fibroin sequence to mechanical func- tion. Journal of Experimental Biology 22:3295- 3303. Guerette, P.A., D.G. Ginzinger, B.H.E Weber, J.M. Gosline. 1996. Silk properties determined by gland specific expression of spider fibroin gene family. Science 272:112-114. Jelinski, L.W., A. Blye, O. Liivak, C. Michal, G. LaVerde, A. Seidel, N. Shah, Z. Yang. 1999. Ori- entation, structure, wet-spinning, and molecular basis for supercontraction of spider dragline silk. International Journal of Biological Macromole- cules 24:197-201. Kaplan, D.L. 1998. Fibrous proteins — silk as a model system. Polymer Degradation and Stabil- ity 59:25-32. Kishore, A.L, M.E. Herberstein, C.L. Craig, E Se- parovic. 2002. Solid-state NMR Relaxation Stud- ies of Australian spider silks. Biopolymers 61: 287-297. Madsen, B., Z.Z. Shao, F. Vollrath. 1999. Variabil- ity in the mechanical properties of spider silks on three levels: interspecific, intraspecific and in- traindividual. International Journal of Biological Macromolecules 24:301-306. Saville, B.P. 1999. Physical testing of textiles; Wood- head Publishing Ltd., ISBN 1 85573 367 6. Stauffer, S.L., S.L. Coguill, R.V Lewis. 1994. Comparison of physical properties of three silks from Nephila Clavipes and Araneus Gemmoides. Journal of Arachnology 22:5-11. Thiel, B.L., K.B. Guess, C. Viney. 1997. Non-pe- riodic lattice crystals in the hierarchical micro- structure of spider (major ampullate) silk. Bio- polymers 41:703-719. Tobolsky, A.V., B.A. Dunell, R.D, Andrews. 1951. Stress relaxation and Dynamic properties of polymers. Textile Research Journal 21:404-41 1. van Beek, J.D., S. Hess, F. Vollrath, B.H. Meier. 2002. The molecular structure of spider dragline silk: Folding and orientation of the protein back- bone. Proceedings of the National Academy of Sciences 99(16): 10266-10271. Van Nimmen, E., P. Kiekens, J. Mertens. 2003. Some material characteristics of spider silk; In- ternational Journal of Materials & Product Tech- nology 18:344-355. Van Nimmen, E., K. Gellynck, L. Van Langenhove, J. Mertens. 2004. The difference in tensile be- havior of different silks of the spider A. diade- matus. Pp. 503-512. In Design and Nature II: Comparing design in nature with science and en- gineering. (M.W Collins & C.A. Brebbia, eds.). WIT Press, United Kingdom (ISBN 1-85312- 721-3). Vollrath, F. and D. Knight. 2001. Liquid crystalline spinning of spider silk. Nature 410:541-48. Vollrath E, B. Madsen, Z. Shao. 2001. The effect of spinning conditions on the mechanics of a spi- der’s dragline silk. Proceedings of the Royal So- ciety of London Series B Biological Sciences 268:2339-2346. Wainwright S.A., W.D. Biggs, J.D. Currey, J.M. Gosline. 1976. Pp. 33-41. In Mechanical Design in Organisms. Edward Arnold (Publishers) Lim- ited. London. ISBN 07131 2502 0. Manuscript received 24 May 2005, revised 8 Au- gust 2005. INSTRUCTIONS TO AUTHORS (revised October 2003) General: Manuscripts are accepted in English only. Authors whose primary language is not English may consult the editors for assistance in obtaining help with manuscript preparation. All manuscripts should be prepared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to consult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than three printed journal pages should be prepared as Feature Articles, shorter papers as Short Communications. One invited Review Article per year will be solicited by the editors and published in the third issue at the discretion of the editors. Suggestions for review articles may be sent to the Managing Editor. Submission: Send one electronic version of the entire man- uscript (in PDF or Microsoft Word format) or send four iden- tical copies of the typed material together with copies of illus- trations to the Managing Editor of the Journal of Arachnology. Paula E. Cushing, Managing Editor, Denver Museum of Nature and Science, Zoology Department, 2001 Colorado Blvd., Denver, CO 80205-5798 USA [Telephone: (303) 370- 6442; FAX: (303) 331-6492; E-mail: PCushing(^dmns.org or PECushing(^juno.com]. The Managing Editor will forward your manuscript to one of the Subject Editors for the review process. You will receive correspondence acknowledging the receipt of your manuscript from the Managing Editor, with the manuscript number of your manuscript. Please use this num.ber in all correspondence regarding your manuscript. Correspondence relating to manu- scripts should be directed to the appropriate Subject Editor. After the manuscript has been accepted, the author will be asked to submit the manuscript on a PC computer disc in a widely-used word processing program. The file also should be saved as a text file. Indicate clearly on the computer disc the word processing program used. Voucher Specimens: Voucher specimens of species used in scientific research should be deposited in a recognized scien- tific institution. All type material must be deposited in a rec- ognized collection/institution. FEATURE ARTICLES Title page. — The title page will include the complete name, address, and telephone number of the author with whom proofs and correspondence should be exchanged, a FAX num- ber and electronic mail address if available, the title in capital letters, and each author’s name and address, and the running head (see below). Abstract. — The heading in bold and capital letters should be placed at the the beginning of the first paragraph set off by a period. A second abstract, in a language pertinent to the nationality of the author(s) or geographic region(s) empha- sized, may be included. Keywords. — Give 3-5 appropriate keywords following the abstract. Text. — Double-space text, tables, legends, etc. throughout. Three levels of heads are used. • The first level (METHODS, RESULTS, etc.) is typed in capitals and on a separate line. • The second level is bold, begins a paragraph with an indent and is separated from the text by a period and a dash. • The third level may or may not begin a paragraph but is italicized and separated from the text by a colon. Use only the metric system unless quoting text or referencing collection data. All decimal fractions are indicated by the peri- od (e.g., -0.123). Citation of references in the text: Cite only papers already published or in press. Include within parentheses the surname of the author followed by the date of publication. A comma separates multiple citations by the same author(s) and a semi- colon separates citations by different authors, e.g., (Smith 1970), (Jones 1988; Smith 1993), (Smith 1986, 1987; Smith & Jones 1989; Jones et al. 1990). Include a letter of permission from any person who is cited as providing unpublished data in the form of a personal communication. Citation of taxa in text: Please include the complete taxonom- ic citation for each arachnid taxon when it appears first in the paper. For Araneae, this taxonomic information can be found on-line at http://research.amnh.org/entomology/spiders/cata- log81-87/INTto2.html. For example, Araneus diadematus Clerck 1757. Literature cited section. — Use the following style and include the full unabbreviated journal title. Opell, B.D. 2002. How spider anatomy and thread configura- tion shape the stickiness of cribellar prey capture threads. Journal of Arachnology 30:10-19. Krafft, B. 1982. The significance and complexity of communi- cation in spiders. Pp. 15-66. In Spider Communications: Mechanisms and Ecological Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Footnotes. — Footnotes are permitted only on the first printed page to indicate current address or other information concerning the author. All footnotes are placed together on a separate manuscript page. Tables and figures may not have footnotes. Running head. — The author’s surname(s) and an abbrevi- ated title should be typed all in capital letters and must not exceed 60 characters and spaces. The running head should be placed near the top of the title page. Taxonomic articles. — Consult a recent taxonomic article in the Journal of Arachnology for style or contact the Subject Editor for Systematics. Papers containing the original taxo- nomic description of the focal arachnid taxon should be given in the Literature Cited section. Tables. — Each table, with the legend above, should be placed on a separate manuscript page. Only horizontal lines (usually three) should be included. Tables may not have footnotes; instead, include all information in the legend. Make notations in the text margins (if possible) to indicate the preferred location of tables in the printed text. Must be double spaced. Illustrations. — Original illustrations should not be sent until the article is accepted for publication. Electronic submis- sions of illustrations is acceptable for review of the manuscript. However, final versions of illustrations of accepted manu- scripts must still be submitted in hard copy (camera-ready, instructions below). Address all questions concerning illustra- tions to the Editor of the Journal of Arachnology: Dan Mott, Editor-In-Chief, Department of Biology & Chemistry, Texas A&M International University, 5201 University Blvd., Laredo, TX 78041-1900 USA [Telephone (956) 326- 2583; FAX: (956) 326-2439; E-mail: dmott(^tamiu.edu]. All art work must be camera-ready — i.e., mounted and labeled — for reproduction. Figures should be arranged so that they fit (vertically and horizontally) the printed journal page, either one column or two columns, with a minimum of wasted space. When reductions are to be made by the printer, pay particular attention to width of lines and size of lettering in line drawings. Multiple photos assembled on a single plate should be mount- ed with only a minimum of space separating them. In the case of multiple illustrations mounted together, each illustration must be numbered sequentially rather than given an alphabetic sequence. Written on the back should be the name(s) of author(s) and an indication of top edge. Indicate whether the illustration should be one column or two columns in width. The overall dimensions should be no more than 1 1 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater difficul- ty in shipping and greater risks of damage for which the Journal of Arachnology! assumes no responsibility. In manu- scripts for review, photocopies should be included, and should be reduced to the exact measurements that the author wants to appear in the final publication. Make notations in the text mar- gins to indicate the preferred position of illustrations in the printed text. Color plates can be printed, but the author must assume the full cost, currently about $600 per color plate. Legends for illustrations should be placed together on the same page(s) and separate from the illustrations. Each plate must have only one legend, as indicated below: Figures 1-4. — A-us x-iis, male from Timbuktu: 1. Left leg; 2. Right chelicera; 3. Dorsal aspect of genitalia; 4. Ventral aspect of abdomen. Figures 27-34. — Right chelicerae of species of A-us from Timbuktu: 27, 29, 31, 33. Dorsal views; 28, 30, 32, 34. Prolateral views of moveable finger; 27, 28. A-us x-us, holo- type male; 33, 34. A-usy-us, male. Scale = 1.0 mm. Assemble manuscript for mailing. — Assemble the sepa- rate sections or pages in the following sequence; title page, abstract, text, footnotes, tables with legends, figure legends, figures. Page charges, proofs and reprints. — Page charges are vol- untary, but non-members of AAS are strongly encouraged to pay in full or in part for their article ($75/joumal page). The author will be charged for changes made m the proof pages. Reprints are available only from the Allen Press and should be ordered when the author receives the proof pages. Allen Press will not accept reprint orders after the paper is published. The Journal of Arachnology also is publishecf by BioOne. Therefore, you can download the PDF version of your article from the BioOne site or the AAS site if you are a member of AAS or if your institute is a member of BioOne. PDF’s of arti- cles older than one year will be freely available from the AAS website. SHORT COMMUNICATIONS Short Communications are usually limited in length to three journal pages, including tables and figures. They will be print- ed in a smaller (10 point) typeface. The format for these is less constrained than for feature articles: the text must still have a logical flow, but formal headings are omitted and other deviations from standard article format can be permitted when warranted by the material being covered. A fossil harvestman (Arachnida, Opiliones) from the Mississippian of East Kirkton, Scotland by Jason A. Dunlop & Lyall I. Anderson 482 A revision of the spider genus Taurongia (Araneae, Stiphidioidea) from southeastern Australia by Michael R. Gray 490 Revision of spider taxa described by Kyukichi Kishida: Part 1 . Personal history and a list of his works on spiders by Hirotsugu Ono 501 Ethology Foraging strategies of Eriophom edax (Araneae, Araneidae): A nocturnal orbweaving spider by Leonor Ceballos,Yann Henaut & Luc Legal 509 The wasp Argiope bruennichi (Arachnida, Araneidae): Ballooning is not an obligate life history phase by Andre Walter, Peter Bliss & Robin F.A. Moritz 516 Can simple experimental electronics simulate the dispersal phase of spider ballooners? by Janies A. Bell, David A. Bohan, Richard Le Fevre & Gabriel S. Weyman 523 Nocturnal navigation in Leucorchestris arenicola (Araneae, Sparassidae) by Thomas Norgaard 533 Use of Anopheles-STpQcifiQ, prey-capture behavior by the small juveniles of Evarcha culicivom, a mosquito-eating jumping spider by Ximena J. Nelson, Robert R. Jackson & Godfrey Sune 541 Egg sac structure of Zygiella x-notata (Arachnida, Araneidae) by T. Gheysens, L. Beladjal, K. Gellynck, E. Van Nimmen, L. Van Langenhove & J. Mertens 549 Notes on the natural history of a trapdoor Ancylotrypta Simon (Araneae, Cyrtaucheniidae) that constructs a spherical burrow plug by Astri Leroy & John Leroy 558 Morphology & Physiology The spermatozoa of the one-palped spider Tidarren argo (Araneae, Theridiidae) by Peter Michalik, Barbara Knoflach, Konrad Thaler & Gerd Alberti 562 On the occurrence of the 9 + 0 axonemal pattern in the spermatozoa of sheetweb spiders (Araneae, Linyphiidae) by Peter Michalik & Gerd Alberti 569 Evidence for the directional selection on male abdomen size in Mecolaesthus longissimus Simon (Araneae, Pholcidae) by Bernhard A. Huber 573 Effects of prey quality on the life history of a harvestman by Aino Hvam & Soren Toft 582 Chromosomal data of two pholcids (Araneae, Haplogynae): A new diploid number and the first cytoge- netical record for the New World clade by Douglas de Araujo, Antonio Domingos Brescovit, Cristina Anne Rheims & Doralice Maria Celia 591 Six stridulating organs on one spider (Araneae, Zodariidae): Is this the limit? By Rudy Joque 597 First ultrastructural observations on the tarsal pore organ of Pseudocellus pearsei and P. boneti (Arachnida, Ricinulei) by Giovanni Talarico, Jose G. Palacios- Vargas, Mariano Fuentes Silva & Gerd Alberti 604 Ultrastructure of male genital system and spermatozoa of a Mexican camel-spider of the Eremobates pallipes species group (Arachnida, Solifugae) by Anja E. Klann, Alfredo Peretti & GerdAlberti 613 Tergal and sexual anomalies in bothriurid scorpions (Scorpiones, Bothriuridae) by Camilo 1. Mattoni 622 Modeling of the stress-strain behavior of egg sac silk of the spider Araneus diadematus by Els Van Nimmen, Kris Gellynck, Tom Gheysens, Lieva Van Langenhove & Johan Mertens 629 The International Society of Arachnology expresses its gratitude to the Congress Organizing Committee: Jean-Pierre Maelfait, President (fN, TEREC Ugent) Leon Baert, Secretary (RBINS) Mark Aldreweireldt Lynda Beladjal Dries Bonte Domir De Bakker Frederick Hendrickx Danny Vanacker (TEREC Ugent) Rudy Jocque (RMCA) Robert Bosmans (AMfNAL) SMtTHSONIAN INSTITUTION LIBRARIES A CONTENTS The Journal of Arachnology 3 9088 01180 9522 Volume 33 Featured Articles Number 2 Ecology Horizontal and vertical distribution of spiders (Araneae) in sunflowers by Stano Pekar 197 Laboratory methods for maintaining and studying web-building spiders by Samuel Zschokke & Marie E. Herberstein 205 The life history of Yllenus arenarius (Araneae, Salticidae) — evidence for sympatric populations isolated by the year of maturation by Maciej Bartos 214 Spatial association between a spider wasp and its host in fragmented dune habitats by Dries Bonte & Jean-Pierre Maelfait 222 Early succession of a boreal spider community after forest fire by Seppo Koponen 230 Are salt marsh invasions by the grass Elymus athericus a threat for two dominant halophilic wolf spiders? by Julien Petillon, Frederic Ysnel, Jean-Claude Lefeuvre & Alain Canard 236 The diet of the cave spider Meta menardi (Latreille 1 804) (Araneae, Tetragnathidae) by Peter Smithers 243 The spider fauna of the irrigated rice ecosystem in central Kerala, India across different elevational ranges by P.A. Sebastian, M.J. Mathew, S. Pathummal Beevi, John Joseph & C.R. Biju 247 Ecological profiles of harvestmen (Arachnida, Opiliones) from Vitosha Mountain (Bulgaria): a mixed modelling approach using GAMS by Plamen G. Mitov & Ivailo L. Stoyanov 256 Influence of grazing by large mammals on the spider community of a Kenyan Savanna biome by Charles M. Warui, Martin H. Villet, Truman P. Young & Rudy Joque 269 Biodiversity & Biogeography Spider (Araneae) communities of scree slopes in the Czech Republic by Vlastimil Ruzicka & Leos Klimes 280 Faunistic similarity and historic biogeography of the harvestmen of southern and southeastern Atlantic rain forest of Brazil by Ricardo Pinto-da-Rocha, Marcio Bernardino da Silva & Cibele Bragagnolo 290 A survey of spiders (Araneae) with holarctic distribution by Yuri M. Marusik & Seppo Koponen . . 300 Fauna and zoogeography of Spiders (Araneae) in Bulgaria by Christo Deltshev 306 Geographical context of speciation in a radiation of Hawaiian Tetragnatha spiders (Araneae, Tetragnathidae) by Rosemary G. Gillespie 313 Diversity of arboreal spiders in primary and disturbed tropical forests by Andreas Floren & Christa Deeleman-Reinhold 323 Lycosidae Gender specific differences in activity and home range reflect morphological dimorphism in wolf spiders (Araneae, Lycosidae) by Volker W. Framenau 334 Evolution of ornamentation and courtship behavior in Schizocosa: insights from a phylogeny based on morphology (Araneae, Lycosidae) by Gail E. Stratton 347 Factors affecting cannibalism among newly hatched wolf spiders (Lycosidae, Pardosa amentata) by Aino Hvam, David Mayntz & Rikke Kruse Nielsen 'ill Data on the biology of Alopecosa psammophila Buchar 200 1 (Araneae, Lycosidae) by Csaba Szinetar, Janos Eichardt & Roland Horvath 384 Size dependent intraguild predation and cannibalism in coexisting wolf spiders (Araneae, Lycosidae) by Ann L. Rypstra & Ferenc Samu 390 Review of the oriental wolf spider genus Passiena (Lycosidae, Pardosinae) by Pekka T. Lehtinen ... 398 The function of long copulation in the wolf spider Pardosa agrestis (Araneae, Lycosidae) investigated in a controlled copulation duration experiment by Andras Sziranyi, Balazs Kiss, Ferenc Samu & Wolfgang Harand 408 Larval chaetotaxy in wolf spiders (Araneae, Lycosidae): systematic insights at the subfamily level by Beata Tomasiewicz & Volker W. Framenau 415 Taxonomy, Systematics & Paleontology A redescription of Porrhomma cavernicola Keyserling (Araneae, Linyphiidae) with notes on Appalachian troglobites by Jeremy A. Miller 426 The fossil spider family Lagonomegopidae in Cretaceous ambers with descriptions of a new genus and species from Myanmar by David Penney 439 The generic relationships of the new endemic Australian ant spider genus Notasteron (Araneae, Zodariidae) by B.C. Baehr 445 Tarsal scopula significance in Ischnocolinae phylogenetics (Araneae, Mygalomorphae, Theraphosidae) by Jose Paulo Leite Guadanucci 456 A preliminary study of the relationships of taxa included in the tribe Poltyini (Araneae, Araneidae) by Helen M. Smith 468 Contents continued on inside back cover